In-Service Performance of Aluminum Structural Details

NTIS #
SSC-447
IN-SERVICE PERFORMANCE OF
ALUMINUM STRUCTURAL DETAILS
This document has been approved

For public release and sale; its
Distribution is unlimited
SHIP STRUCTURE COMMITTEE

2007
Ship Structure Committee
RADM Craig E. Bone

U. S. Coast Guard Assistant Commandant,
Marine Safety and Environmental Protection
Chairman, Ship Structure Committee
Mr. W. Thomas Packard Dr. Roger Basu
Director, Senior Vice President
Survivability and Structural Integrity Group American Bureau of Shipping
Naval Sea Systems Command
Mr. Joseph Byrne Mr. Nash
Director, Office of Ship Construction Director General, Marine Safety,
Maritime Administration Safety & Security
Transport Canada
Mr. Kevin Batesen Dr. Neil Pegg
Director of Engineering Group Leader - Structural Mechanics
Military Sealift Command Defence Research & Development Canada - Atlantic
CONTRACTING OFFICER TECHNICAL REP. EXECUTIVE DIRECTOR

Mr. Chao Lin / MARAD Lieutenant Benjamin A. Gates
Glenn Ashe / ABS U. S. Coast Guard
Mr. Robert Sedat / USCG
SHIP STRUCTURE SUB-COMMITTEE

AMERICAN BUREAU OF SHIPPING DEFENCE RESEARCH & DEVELOPMENT ATLANTIC
Mr. Glenn Ashe Dr. David Stredulinsky
Mr. Phil Rynn Mr. John Porter
Mr. Balji Menon
Mr. Derek Novak
MARITIME ADMINISTRATION MILITARY SEALIFT COMMAND

Mr. Chao Lin Mr. Michael W. Touma
Mr. Carl Setterstom Mr. Paul Handler
Mr. Richard Sonnenschein Mr. James Kent
Mr. David Heller
ONR / NAVSEA TRANSPORT CANADA
Dr. Paul Hess Mr. Richard Stillwell
Dr. Jeffery E. Beach
Mr. Yapa Rajapakse
Mr. Allen H. Engle
UNITED STATES COAST GUARD SNAME
Capt. Patrick E. Little Mr. Jaideep Sirkar
Mr. H. Paul Cojeen Mr. Alan Rowen
Mr. Rubin Sheinberg Mr. Norman Hammer
Mr. Robert Sedat Mr. John Conlon
Technical Report Documentation Page
1. Report No. 2. Government Accession No. 3. Recipient’s Catalog No.
SSC - 447
4. Title and Subtitle 5. Report Date
In-Service Performance of Aluminum Structural Details Nov. 1, 2006
6. Performing Organization
Code
5813C.FR
7. Author(s) 8. Performing Organization
R. Kramer Report No.
SR-1444
9. Performing Organization Name and Address 10. Work Unit No. (TRAIS)
Alionscience – JJMA
311 Legget Drive 11. Contract or Grant No.
Kanata, ON (Canada) K2K 1Z8
12. Sponsoring Agency Name and Address 13. Type of Report
Ship Structure Committee Final Report
C/O Commandant (G-MSE/SSC)
United States Coast Guard
2100 Second Street, SW 14. Sponsoring Agency Code
Washington, DC 20593-0001 G-P
15. Supplementary Notes
Sponsored by the Ship Structure Committee and its member agencies
16. Abstract
The primary objective of the project was to compare and evaluate the design criteria and standards
currently used in naval and commercial ships for the hull and structural members. This report reviewed
the basic concepts in several current ship and structures regulations. The design of bottom structure, as
both local structure and as part of the hull girder was the specific focus. We expected to identify factors of
safety in either the load or strength formulations or both. As well as identify the best practices that
incorporate latest models of structural behavior that are adequately validated by theory and
experimentation. This will then be applied to new unified structural design ships.
17. Key Words 18. Distribution Statement

Distribution Available From:
National Technical Information Service
U.S. Department of Commerce
Springfield, VA 22151
Ph. (703) 605-6000
19. Security Classif. (of this 20. Security Classif. (of this 21. No. of 22. Price
report) page) Pages
Unclassified Unclassified 104
Table of Contents
1. INTRODUCTION ...............................................................................................................1
1.1 Determination of Details & Classes of Details for Survey ........................................1
1.2 Fatigue Behavior of Aluminum Structural Details ....................................................2
1.3 Conclusions and Recommendations...........................................................................2
2. DETERMINATION OF DETAILS & CLASSES OF DETAILS FOR SURVEY.............3
2.1 List of Ship Details & Background Information ........................................................3
2.2 Review of Previous Ship Structure Committee Reports ............................................4
2.2.1 SSC 266 “Review of Ship Structural Details” 1977 ......................................5
2.2.2 SSC 272 “In-Service Performance of Structural Details” 1978.....................6
2.2.3 SSC 294 “Further Survey of In-Service Performance of Structural Details”
1980................................................................................................................8
2.2.4 Other SSC Reports Reviewed ........................................................................8
3. FATIGUE AND FRACTURE BEHAVIOR OF ALUMINUM STRUCTURAL
DETAILS.............................................................................................................................9
3.1 Damage Tolerant Design and Fracture Analysis........................................................9
3.2 History and Development of the European Standards ...............................................9
3.3 Friction Stir Welding................................................................................................11
4. EUROPEAN RECOMMENDATIONS FOR ALUMINUM ALLOY STRUCTURES
– FATIGUE DESIGN (ERAAS) .......................................................................................13
4.1.1 Design S-N Curves in the ERAAS...............................................................13
4.1.2 Data Base of the Recommendations.............................................................16
4.1.3 Base Metal 5000/6000 Series.......................................................................18
4.1.3.1 Butt Weld, Transverse ...................................................................19
4.1.3.2 Butt Weld, Longitudinal ................................................................21
4.1.3.3 Longitudinal Fillet Welds ..............................................................22
4.1.3.4 Fillet Weld, Transverse, Non-Load-Carrying................................23
4.1.3.5 Web Attachments...........................................................................25
4.1.3.6 Attachment at Edge of Flange........................................................25
4.1.3.7 Vertical Attachment on Flange ......................................................26
4.1.3.8 Cruciform Joint ..............................................................................28
4.1.3.9 Cover Plate.....................................................................................29
5. COMPARISON BETWEEN ERAAS AND OTHER CODES .........................................31
5.1 Comparison to Aluminum Codes .............................................................................31
5.2 Comparison of ERAAS Fatigue 1992 & Aluminum Association Design
Manual......................................................................................................................33
5.3 Comparison of ERAAS to Various Steel Codes ......................................................34
6. TABLES FOR DETAIL CATEGORIES AFTER ERAAS FATIGUE 1992
DECISIONS LEADING TO FATIGUE STRENGTH DESIGN VALUES .....................37
7. COMPARISON OF FATIGUE STRENGTH VALUES FOR STRUCTURAL
DETAILS...........................................................................................................................45
8. FRICTION STIR WELDING............................................................................................57
8.1 General Information on Friction Stir Welding .........................................................57
8.1.1 The Process ..................................................................................................57
8.1.2 The Principle of Operation ...........................................................................58
ii
8.1.3 Friction Stir Welding Process Advantages...................................................62
8.1.4 Materials and Thickness...............................................................................62
8.1.5 Weld Properties and Characteristics ............................................................64
8.1.6 Welding Parameters .....................................................................................68
8.1.7 Joint Geometries...........................................................................................68
8.1.8 Applications in Shipbuilding and Marine Industries....................................69
8.2 Fatigue Behavior of Friction Stir Welds in 6000 Series Aluminum Alloys ............70
8.2.1 Fatigue Behavior of Friction Stir Welds in Aluminum Alloy 6082.............70
8.2.1.1 First experimental data on aluminum alloy 6082-T6.....................70
8.2.1.2 Data on Friction Stir Welds in Aluminum 6082 in the T4 and
T6 Tempers ....................................................................................78
8.2.1.3 Additional Wöhler Diagram for Aluminum 6082-T4....................84
8.2.2 Fatigue Behavior of Friction Stir Welds on Other 6XXX Aluminum Alloys87
8.2.2.1 Experimental Data on Aluminum Alloy A6N01-T5 .....................87
8.2.3 Experimental Data from Fatigue Tests on Transverse FSW Butt Welds on
Aluminum Alloy 6013-T6............................................................................90
8.3 Conclusions on Friction Stir Welding......................................................................96
9. CONCLUSIONS AND RECOMMENDATIONS ............................................................99
10. REFERENCES ................................................................................................................101
10.1 European Standards................................................................................................101
10.2 Friction Stir Welding..............................................................................................102
iii
List of Figures
Figure 1 Equi-distant Design Curve Mesh ............................................................................... 14
Figure 2 Effect of Classification Considering Equi-Distant and Parallel Design Line
Concepts on Attainable Fatigue Strengths ....................................................................... 14
Figure 3 Example of the effect of parallel and equi-distant design lines on fatigue strength
according to the different structural details after ERAAS Fatigue 1992 ...................... 15
Figure 4 Full-Scale Background Data for ERAAS-Fatigue Document ................................. 18
Figure 5 Base Metal 6005A (AlMgSi0) Small Specimen Data ................................................ 18
Figure 6 Base Metal 5083 (AlMg4,5Mn) Full Scale Component Data................................... 19
Figure 7 Transverse Butt Weld, Overfill Ground Flush, Extruded & Built-Up Components
............................................................................................................................................... 20
Figure 8 Proposed Design Line for Transverse Butt Weld, Overfill Ground Flush,
Extruded & Built-Up Components.................................................................................... 20
Figure 9 One-sided Transverse Butt Weld on Built-up Components .................................... 21
Figure 10 Two-sided Transverse Butt Weld on Built-up Components.................................. 21
Figure 11 Longitudinal Butt Weld, Overfill Ground Flush.................................................... 22
Figure 12 Longitudinal Butt Weld, As Welded........................................................................ 22
Figure 13 Longitudinal Fillet Weld with/without Stop-Starts................................................ 23
Figure 14 Intermittent Longitudinal Fillet Weld..................................................................... 23
Figure 15 Fillet Weld, Transverse, Non-Load-Carrying (Web Stiffener) ............................. 24
Figure 16 Transverse Fillet Weld, Non-Load-Carrying (Web Stiffener), ATLLS-Lehigh
Test Results.......................................................................................................................... 24
Figure 17 Web Attachments ...................................................................................................... 25
Figure 18 Attachment at Flange Edge, with transition radius, r>50 mm.............................. 26
Figure 19 Attachment at Flange Edge, no transition radius................................................... 26
Figure 20 Vertical Attachment (Long’l) - Flange of Built-up Beam, no transition radius .. 27
Figure 21 Vertical Attachment (Long’l) - Flange of Built-up/Extruded Beam, no transition
radius.................................................................................................................................... 27
Figure 22 Transverse Attachment Vertical on Flange, Built-up Beams................................ 28
Figure 23 Cruciform Joint, Built-up Beams............................................................................. 28
Figure 24 Cover plate, Built-up Beam, TUM Test Results ..................................................... 29
Figure 25 Cover plate, Built-up Beam, Lehigh Test Results .................................................. 29
Figure 26 Cover plate, Built-up Beam, Lehigh and TUM Test Results................................. 30
Figure 27 Comparison of ERAAS Fatigue to Various Aluminum Codes.............................. 32
Figure 28 Comparison of ERAAS Fatigue to BS 8118 ............................................................ 32
Figure 29 Comparison of ERAAS & Aluminum Association Fatigue @ 2x106 Cycles........ 33
Figure 30 Comparison of ERAAS & Aluminum Association S/N Curves ............................ 33
Figure 31 Comparison of ERAAS Fatigue to Various Steel Codes........................................ 35
Figure 32 Ratio of Fatigue Design Values (2×106 Cycles)-ERAAS (Al) & Eurocode 3 (Steel)
............................................................................................................................................... 36
Figure 33 Schematic Illustration of Friction Stir Welding ..................................................... 58
Figure 34 FSW – Principles of the Friction Stir Weld Process............................................... 59
Figure 35 Schematic for the Tip of the FSW Tool ................................................................... 60
Figure 36 Samples of FSW Welds ............................................................................................. 60
Figure 37 Definitions of FSW terminology............................................................................... 61
iv
Figure 38 FSW Setup.................................................................................................................. 61
Figure 39 Transverse Section of 6mm Wrought Aluminum Welded to Cast Aluminum .... 63
Figure 40 Double Sided Friction Stir Weld in 75 mm Thick Aluminum Extrusion............. 64
Figure 41 Illustration of the Regions Associated with a Friction Stir Weld.......................... 65
Figure 42 Nugget at Center of Friction Stir Weld ................................................................... 66
Figure 43 Typical Joint Configurations for Friction Stir Welds............................................ 68
Figure 44 Samples of FSW Components .................................................................................. 69
Figure 45 Deck panels made from FSW profiles ..................................................................... 70
Figure 46 Specimen Geometry................................................................................................... 71
Figure 47 Wöhler Diagram for friction stir butt weld in aluminum alloy 6082-T6 ............. 73
Figure 48 Wöhler Diagram for AA6082-T6 (failure occurred in parent material).............. 76
Figure 49 Comparison between the two Wöhler Diagrams .................................................... 77
Figure 50 Wöhler Diagram for AA6082-T4 Friction Stir Welds R=0.5 ................................ 80
Figure 51 Wöhler Diagram for AA6082-T6 Friction Stir Welds (R=0.5).............................. 83
Figure 53 Wöhler Diagram for friction stir welds in aluminum alloy 6082-T4 .................... 86
Figure 54 Beam-type Specimen ................................................................................................. 88
Figure 55 Comparison between Wöhler Diagram for friction stir welds in A6N01-T5
(Japanese Industrial Standard) and design curve for longitudinal butt welds............. 89
Figure 57 Wöhler Diagram for FSW butt welds in the as welded conditions on aluminum
alloy 6013–T6....................................................................................................................... 92
Figure 58 Wöhler Diagram for FSW butt welds in the as milled conditions on aluminum
alloy 6013 – T6..................................................................................................................... 95
Figure 59 Transverse friction stir welds no surface post-treatment compared to fatigue
strength of transverse butt welds (Aluminum Data Bank – TUM)................................ 97
Figure 60 Transverse friction stir welds without surface post-treatment on aluminum alloys
6XXX compared with Design Standards .......................................................................... 98
v
List of Tables
Table 1 Detail Classification and Number of Observed Details in SSC 272............................ 7
Table 2 Detail Classifications and Number of Observed Details in SSC 294 .......................... 8
Table 3 Correlation between ERAAS and Aluminum Association Details ........................... 34
Table 4 Butt Weld Transverse - Simple Specimen & Extruded Component (ERAAS B1-B8)
............................................................................................................................................... 37
Table 5 Butt Weld Transverse – Built up Components (ERAAS B9 to B11)........................ 38
Table 6 Butt Weld – Longitudinal (ERAAS C1 & C2) ........................................................... 39
Table 7 Fillet Weld – Longitudinal (ERAAS D1, D2 & D3) ................................................... 39
Table 8 Fillet Weld – Transverse (non-load carrying), Web Stiffeners & Attachments
(ERAAS Design Curves E1 & E2)..................................................................................... 40
Table 9 Attachment at Flange Edge (ERAAS E3 &E4) .......................................................... 41
Table 10 Vertical Attachment on Flange – Longitudinal (ERAAS E5, E6 &E7) ................. 42
Table 11 Vertical Attachment on Flange – Transverse (ERAAS E8) .................................... 43
Table 12 Fillet Weld–Transverse (load carrying) Cruciform, Cover plate (ERAAS F2, F3)
............................................................................................................................................... 44
Table 13 Comparison of Design Curve Fatigue Strength Values for Transverse Butt Welds
to Experimental Data (at 2*106 cycles) ............................................................................. 45
Table 14 Comparison of Design Curve Fatigue Strength Values for Parent Material and
Experimental Data (at 2*106 cycles).................................................................................. 48
Table 15 Comparison of Design Curve Fatigue Strength Values for Longitudinal Butt and
Longitudinal Fillet Welds and Experimental Data (at 2*106 cycles) ............................. 48
Table 16 Comparison of Design Curve Fatigue Strength Values for Transverse Fillet Welds
and Experimental Data (at 2*106 cycles) .......................................................................... 50
Table 17 Comparison of Design Curve Fatigue Strength Values for Welded Transverse or
Longitudinal Attachments on Load-Carrying Structural Components and
Experimental Data (at 2*106 cycles).................................................................................. 52
Table 18 Typical Mechanical Properties of Friction Stir Welded Aluminum Specimens ... 67
Table 19 Chemical Composition of AA6082-T6 in Initial FSW Fatigue Tests ..................... 71
Table 20 Mechanical Properties of AA6082-T6 ....................................................................... 71
Table 21 Experimental points from FSW-joints fatigue tests................................................. 72
Table 22 Results of Linear Regression Analysis ...................................................................... 74
Table 26 Friction Stir Welding Process Parameters ............................................................... 81
Table 29 Mechanical Properties of AA6082 ............................................................................. 84
Table 30 Friction Stir Welding Process parameters................................................................ 85
Table 31 Details of mean life S-N curve obtained from regression analysis of test results.. 85
Table 32 Chemical Composition of A6N01S-T5 ...................................................................... 87
Table 33 Tab. 15 Chemical Composition of Aluminum Alloy 6008 ....................................... 87
Table 34 Welding Data ............................................................................................................... 90
Table 35 Experimental points from fatigue tests – FSW as welded ....................................... 91
vi
Table 37 Experimental points from fatigue tests – FSW as milled ........................................ 93
vii
1. INTRODUCTION
In accordance with the Statement of Work included in the Solicitation for this proposal, the
Objectives of this project are:
1. Develop a survey of in-service performance of aluminum structural details for use

by designers and fabricators of aluminum vessels. Start to fill some of the gaps
identified in SSC-410, “Fatigue of Aluminum Weldments.” Consideration shall
be given to several types of vessels (catamarans, SES’s, ACV’s, SWATH’s,
monohulls, etc.) using aluminum details.
2. Construct a database of the in-service information and develop a rating system to

assist designers and fabricators with the selection of fatigue resistant details and
fabrication procedures. Coordinate the survey data with fabrication and QA
procedures employed during the initial construction recommending minimum
criteria that should be used to enhance the behavior of aluminum details subjected
to the fatigue environment.
3. Introduce fatigue to the design stage of a vessel and address the fatigue issue up-
front instead of the current practice which relegates this to a maintenance problem
causing undue expense throughout the life of the vessel.
The execution of this project was unable to perform the surveys initially planned although some
of the ground work was developed and is presented herein. The preliminary work included
interviews with various small boat yards who indicated good success with the welding and
performance of aluminum details. As anticipated, these interviews were also conducted with
yards and owners that were more reluctant to provide information on relevant experience, which
is understandably attributed to the competitive nature of this industry. Regardless, there is
significant information regarding the performance of aluminum structural details in small and
high speed aluminum craft. Gathering this information would develop a good database of
information and provide designers good insight to the selection of details with improved
performance in the fatigue environment.
To help compensate for the lack of survey work there is more emphasis on the existing fatigue
data and design standards for aluminum structural details. This focuses on the work that has
been developed in Europe, where significant effort has been devoted to the testing of small and
large specimens in support of addressing the fatigue of aluminum during the design process.
1.1 Determination of Details & Classes of Details for Survey
Section 2 of this report presents the preliminary plans that had been intended for the surveys.
The efforts for Section 2 included a review of all previous SSC reports to determine survey
procedures successfully employed for similar efforts in steel. This effort was beneficial and
helped to establish the categories of structural details that would have originally been sought
during the surveys. It is worth noting that all previous SSC efforts addressed steel, large
displacement vessels while the current project is focused on aluminum and aluminum high speed
1
craft. Therefore, the classes of details defined in Section 2 would have been modified to address
the differences between these types of vessels and would have evolved during the execution of
the surveys.
1.2 Fatigue Behavior of Aluminum Structural Details
The information gathered from the European community regarding the fatigue of aluminum is
presented in Section 3 though Section 7 of this report. This information includes some specific
comparisons between different design standards, including the Aluminum Association, regarding
the use of fatigue in the design of aluminum structural systems. The fatigue data presented
throughout these sections are for specific structural details with their geometry provided for easy
reference. Many of these details and their structural functions can be related to those used in the
marine industry for estimation of design performance in the current application of aluminum
craft. Additional information on environmental effects can also be obtained from design
standards such as Eurocode 9, which includes knockdown factors for the Marine and Immersed
environments.
Section 8 of this report presents information for Friction Stir Welding, FSW, procedures and is
taken from the body of work gathered for the fatigue behavior of aluminum from European
sources. It includes information on the fatigue performance of aluminum details fabricated with
FSW as well as comparison to similar details fabricated with traditional fusion welding
procedures.
1.3 Conclusions and Recommendations
Section 9 presents conclusions and recommendations for additional work in this field and
recognizes the work that has also been done for fracture mechanics and crack propagation in
aluminum details, again with reference to European sources.
2
2. DETERMINATION OF DETAILS & CLASSES OF DETAILS FOR SURVEY
This section of the report presents the preliminary list of details to be investigated during the
vessel surveys. This list of details may evolve as the task progresses and additional “trouble
spot” areas become identified. The list of details presented for this project is assumed to be
readily accessible, i.e., it is assumed that there will not be any survey of details contained within
tanks or other inaccessible areas such as voids. All surveys will take place while the vessel is
either pier side or underway during routine operations. The Survey Team will not cause any
form of damage to anything on the vessels being surveyed. There will not be any Non-
destructive evaluation of structure using any techniques such as dye penetrant, magnetic particle
or ultrasonic sound. All survey data shall be limited to visual recording with light tamping of a
small hammer, if beneficial, to detect flaws in the metal. Photographic records of surveyed
details will be developed.
2.1 List of Ship Details & Background Information
The following list of details will be specifically sought for investigation during the vessel
surveys for this project:
• Bow, bottom and cross structure in way of areas subjected to repeated slam loads.
(Ability to survey cross structure will depend on access.)
• Stiffener end connections, i.e., vertical bulkhead to longitudinal deck stiffening and
vertical bulkhead to bottom shell stiffening.
• Typical longitudinal stiffener/transverse web frame intersection on strength deck, bottom
and side shell.
• Transverse bulkhead structure in way of haunch.
• Deck beam to side frame connection in transversely framed ships.
• Details in way of hard chines.
• Machinery Foundations and structural details in way of machinery spaces.
• Waterjet and bow thruster foundations.
• Structural detailing in way of free-standing tanks.
• Structural details in way of windows, hatches and other openings/penetrations through
light superstructure scantlings.
Other information sought prior to and during the surveys will include:
• Operating environment of vessel.

• Classification notation for craft certified in accordance with regulatory body requirements
– is the vessel designed in accordance with regulatory requirements but unclassed?
• Aluminum alloy/temper and weld metals used in fabrication.
• In-service histories of details throughout the fleet including repair histories.
• Approximate costs, or ranking of costs, associated with maintenance of details with poor
histories.
3
The vessel surveys will also include photographic documentation to the greatest extent practical.
This will, of course, require the permission of the vessel owner/operator who will be assured of
complete anonymity for all photographs used in the survey and final report, if desired.
One of the primary objectives of this project is to perform vessel surveys that will allow for the
investigation of structural details and evaluate their in-service performance. This can include
details and classes of details that have good in-service records as well as those that perform
poorly, although the emphasis will be on the latter, i.e., structural details with poor in-service
performance histories.
The preliminary list of details was based on experience with these vessels, input from the Project
Technical Committee, PTC, and review of the SSC reports shown below. While it was not
expected to gather specific, relevant information regarding details from the SSC reports it was
expected that there might be some good insight into the survey procedures and the manner in
which details are classified. The review of the SSC reports confirmed most of the expected
procedures and helped to validate the efforts that are expected. The SSC reports reviewed
include:
• SSC-266 “Review of Ship Structural Details”

• SSC-272 “In-Service Performance of Structural Details”
• SSC-294 “Further Survey of In-Service Performance of Structural Details”
• SSC-318 “Fatigue Characterization of Fabricated Ship Details for Design”
• SSC-367 “Fatigue Technology Assessment and Avoidance Strategies in Marine
Structures”
• SSC-379 “Improved Ship Hull Structural Details Relative to Fatigue”
• SSC-400 “Weld Detail Fatigue Life Improvement Techniques”
As noted by their titles, not all of the reports address the stage of investigation associated with
vessel survey. The lessons learned from the latter reports will be incorporated into the current
project as applicable. Hopefully this will help to reduce the learning curve for aluminum by
taking advantage of the lessons learned in treating the fatigue problem in steel.
While it is not the intention of this project to limit the details sought for investigation during
survey, there are some practical limitations. Since the objective of this project is the survey of
details with poor in-service performance histories, it is expected that those details included in the
list above with better histories will receive less attention than those more prone to problems. It is
fully expected that the relevance of such histories will become apparent prior to and during the
actual surveys. The list of details provided above represents an optimistic checklist that could
potentially include details with poor in-service histories. It will be difficult to completely survey
all these classes of details within the scope of this project.
2.2 Review of Previous Ship Structure Committee Reports
To date, the Ship Structure Committee has published over 400 reports. Dozens of these reports,
starting in 1946, deal with one or more aspects of fatigue. All but one of these reports deals with
steel and virtually all information reflect displacement type vessels that operate at relatively low
4
speed in an open ocean, unrestricted environment. The current project specifically addresses the
in-service performance of structural details in aluminum vessels, with a tendency towards high-
speed, although the latter is not a necessary condition. Regardless, it was well recognized that
the great bulk of data and research currently contained within the SSC reports addresses steel
displacement vessels and that the direct relevance to the current project may be limited.
It is interesting to note the progression of SSC reports relative to steel fatigue. There was a
natural progression in the topics they covered that reflects the evolution of a continuing research
project. The initial reports are similar in nature to the current project for aluminum, data
gathering, survey type reports that attempt to start quantifying the in-service performance of the
steel details subjected to fatigue. These are followed by reports that reflect the environments and
loading histories that cause fatigue and proposals to start quantifying and predicting fatigue
damage during the design stage. Subsequent reports propose improved structural details for
resistance to cracking in the fatigue environment along with improved welding and fabrication
procedures to minimize the crack initiation mechanism associated with all welding procedures.
The progression of SSC reports addressing steel fatigue reflects what is expected to be a similar,
although abbreviated series of SSC reports for aluminum. It is expected to be abbreviated
because of the potential learning curves that can be applied from steel to aluminum even though
the operational and loading profiles of the respective vessel types can be significantly different.
Regardless, the review of the SSC reports conducted for the current project did provide some
insight into format and procedure for the surveys, if not so much for the specific types of details.
It should be noted that SSC 266, SSC 272 and SSC 294 present comprehensive classes of details
to be included in vessel surveys with significant survey data included in SSC 272 and SSC 294.
These are all generic details, relevant to displacement type vessels, and certainly applicable to
the surveys conducted at the time for the steel ships involved.
2.2.1 SSC 266 “Review of Ship Structural Details” 1977
As mentioned above, this report contains a fairly comprehensive presentation of the typical
structural details required to complete the design of any steel displacement vessel. They are
global categories and suggested the manner of classification for the current project. There are 15
classes of details presented in SSC 266 as follows:
1. Clearance Cuts
2. Snipes
3. Tight Collars
4. Reeving Slots
5. Structural Intersections
6. Miscellaneous Cutouts
7. Patches
8. Stanchion end Connections
9. Tripping Brackets
10. Face Plates
11. Stiffener Ends
5
12. Clip Connection
13. Chock
14. Panel Stiffener
15. Beam Bracket
The details are generic and include nominal information regarding clearances, radii, weld types,
gusset and chock locations, etc. A simple sketch is provided for each detail within SSC 266. It
is anticipated that the current project will include sketches/photographs for all relevant details.
There are no specific results for SSC 266 that will benefit this project. In addition to defining the
15 classes of details noted above, SSC 266 summarizes much of the classification society
requirements relative to structural detail development. It provides some background on damage
histories found from previous work and also includes introductory ideas regarding strength and
fatigue criteria for detail design.
One of the conclusions from SSC 266 is that there is very little feedback of the performance of
structural details back to the designer. Since many of the aluminum vessels fabricated do not
require certification of the details in the design it is expected that many of the larger fabricators
may track performance for their own, internal use, but not necessarily make the data generally
available to classification societies or other designers. This closely tracked performance could
be considered an advantage in the competitive world of high-speed ferry design. With no
incentive or requirement to provide such performance history to the class societies, it would not
be unexpected if no feedback is provided.
2.2.2 SSC 272 “In-Service Performance of Structural Details” 1978
This report summarizes actual survey work performed for the project. It includes a tremendous
amount of data collected on fifty (50) displacement type, steel vessels. Newport News
Shipbuilding performed the task. Thirty three (33) of the fifty (50) vessels surveyed were
inspected at the NNS facility while they were in for scheduled maintenance, inspections,
overhauls or unscheduled emergency repairs. The average vessel had a displacement of 34,980
long tons and an LBP of 622 feet. This availability and access to such large vessels for extended
periods represents a significant difference to the availability and access for the current project.
Similar to SSC 266, this project also created various classes of details for grouping the in-service
performance data. SSC 272 used 12 classes of details, which along with the observed number of
details, is presented in Table 1. SSC 272 also included the results of the survey indicating the
number of failed details.
The volume of details surveyed is tremendous compared to the expectations for the current
project.
6
Table 1 Detail Classification and Number of Observed Details in SSC 272
Detail Classification Number of Observed Details
1. Beam Bracket 50,750
2. Tripping Bracket 20,640
3. Non-Tight Collars 16,250
4. Tight Collars 18,000
5. Gunwale Connection 100
6. Knife Edge Crossing None found
7. Miscellaneous Cutouts 252,870
8. Clearance Cutouts 48,510
9. Structural Deck Cuts 6030
10. Stanchion Ends 6270
11. Stiffener Ends 30,760
12. Panel Stiffeners 40,030
The vessels surveyed for this project will be significantly smaller than those surveyed for SSC
272, probably in the range of 1% to 5% of the average displacement quoted above. Also, they
will be fully operational and in-service, i.e., performing their daily, commercial operations, not
laid-up for repair, maintenance, inspection, etc. The surveyors will be traveling to the ships, not
have the convenience of ships that are laid-up for repairs or other services over relatively
protracted periods of time. The level of effort for the current project will not allow for more than
a few days of actual survey.
Regardless, the procedures for conducting the surveys in SSC 272 are similar to those anticipated
for the current project. All surveys will be performed on readily accessible structure without
causing any damage to any of the surrounding ship systems. i.e., insulation, paint, drop ceilings,
etc. The surveyors on the current project may also employ small hammers or other devices to
lightly tap structure suspected of containing a crack. The surveyors for the current project will
not use any of the more advanced non-destructive evaluation procedures associated with typical
classification society/shipyard QA procedures. None of the advanced NDE procedures were
used during SSC 272.
One of the differences between the details classifications of SSC 266 and SSC 272 suggests the
similar direction anticipated for the current project, i.e., the inclusion of localized, special areas.
These are noted by such classes as “Gunwale Connection” and “Knife Edge Crossing” in SSC
272 whereas SSC 266 did not contain any details that approached this level of specific definition.
As noted above, the current project anticipates developing classes that relate to specific areas of
repeated failures on aluminum high-speed vessels.
Although its exact form is yet to be determined, the results of the surveys for the current project
will also develop a database presentation for ease of access to the data. Both SSC 272 and SSC
294 also present numerous photographs helping to describe the details. Both of these reports
also maintain the anonymity of all vessels included in their surveys. The reports do include the
number of each type of vessel included in the survey but do not track in-service performance
history as a function of the type of ship. All of these practices have been anticipated for the
current project since its proposal development stage.
7
2.2.3 SSC 294 “Further Survey of In-Service Performance of Structural Details” 1980
This report is a continuation of the efforts in SSC 272. It was also performed by Newport News
Shipbuilding and used the same approach as presented in SSC 272. The surveys completed for
SSC 294 involved thirty-six (36) ships, most of which were surveyed at NNS. As a matter of
completeness, Table 2 presents the same data for SSC 294 that Table 1 presented for SSC 272.
Table 2 Detail Classifications and Number of Observed Details in SSC 294

Detail Classification Number of Observed Details
1. Beam Bracket 17,836
2. Tripping Bracket 13,372
3. Non-Tight Collars 4724
4. Tight Collars 2654
5. Gunwale Connection 72
6. Knife Edge Crossing None found
7. Miscellaneous Cutouts 43,819
8. Clearance Cutouts 8797
9. Structural Deck Cuts 1504
10. Stanchion Ends 820
11. Stiffener Ends 9969
12. Panel Stiffeners 13,807
Again, it can be seen that there was a tremendous amount of data collected for SSC 294. This
project had the same advantage as SSC 272, i.e., the vessels had extended availabilities as a
result of being at NNS for scheduled maintenance, overhauls, etc. This advantage will not
present itself for the current project.
2.2.4 Other SSC Reports Reviewed
As the SSC reports continue beyond SSC 294 the data and information relates to further stages of
the fatigue evaluation/prevention cycle. Most of that information is not relevant to the objective
of this section of the report, i.e., definition of the list of details to be surveyed. Limited amounts
of that information will be relevant to future development of this project and a larger amount
may become relevant to future projects associated with fatigue of aluminum weldments. As
discussed above, this information will help to shorten the learning curve for fatigue of aluminum
weldments by borrowing from the lessons learned with fatigue of steel weldments.
The preliminary list of details to be included in the vessel surveys for this project provides a
good starting point that reflects the known areas of concern of poor in-service candidates. More
vessel specific areas may reveal themselves as a result of the surveys. Trends between such
vessel specific areas will be investigated?
It is expected that the list of details will evolve as the project continues. This will present no
problem as the information will be gathered during the surveys and the survey team will have
good knowledge of the areas with poor histories as a result of the preliminary discussions held
with all operators.
8
3. FATIGUE AND FRACTURE BEHAVIOR OF ALUMINUM STRUCTURAL
DETAILS
The main reason to include discussion on the European standards and Friction Stir Welding,
FSW, in this report regarding in-service performance is to introduce the large volume of work
that has been developed in Europe to support the use of aluminum in structural applications. The
information presented in this report will help the reader understand the history and development
of much of the design standards, U.S. and European, concerning aluminum. There has been
extensive testing of small and large specimens developed in support of the European standards
regarding fatigue of aluminum. Also, the work presented herein provides a blueprint for the
progression of the development of fatigue in aluminum structural weldments that was anticipated
for this and follow-up SSC reports. The same procedures used to refine the understanding of
fatigue in general applications could be applied to structural details specific to aluminum and
aluminum high speed craft.
3.1 Damage Tolerant Design and Fracture Analysis
This report does not include any of the specific work or reference information that is available
through some of the European standards on this subject, i.e., Eurocode 9. The topic is introduced
through these headings only to alert the reader that there is extensive information available. This
includes design curves for da/dN, crack extension per cycle of load as a function of crack tip
stress intensity. These curves can be used to estimate the remaining life in a cracked detail once
crack initiation has been identified and critical crack length defined. The life is defined in
number of load cycles which can then be translated into calendar time to help evaluate the
criticality of joint repair. Regardless, there is significant information for crack propagation,
fracture analysis and crack growth rate data through the reference provided above.
3.2 History and Development of the European Standards
The development of European specifications for the design of aluminum structures has been
supported by activities that have taken place from the 1980’s up to the present day when they are
reaching their final stage with the drafting of the Eurocode 9 or EN 1999-1:2004.
Following the initial contacts through the establishment of the INALCO International
Conference on Aluminum Weldments, the Committee for Aluminum Fatigue Data Exchange and
Evaluation was formed. Its task was to unite the two databases on aluminum fatigue data started
a few years earlier at Iowa State University (Prof. Dr. W. W. Sanders, Jr.) and Technical
University of Munich (Prof. Dr. D. Kosteas) and expand these to include any available data. This
joint project continued for several years, the data bank was later based and maintained at the
Technical University of Munich (TUM), but had to be discontinued in the 1990’s for lack of
funds. Nevertheless, the initial statistical-regression analyses on small specimen data and data
from larger specimen testing started in the 1980’s at the TUM formed the basis for the first
European common document.
The second phase of development is characterized by the analysis of comprehensive new data on
small specimens but also encompasses a considerable amount of component tests (aluminum
9
beams) supplied by Alusuisse, Switzerland. The project was carried out by the Section of Light
Metal Structures and Fatigue at the Technical University of Munich (Kosteas) with the support
of a statistician from the Pechiney Research Center in Voreppe/France, and Dr. R. Jaccard,
Alusuisse, Zürich/Switzerland. This project was also discussed with researchers and company
representatives from different European countries in two workshops in Zürich and Munich. A
representative compilation of results was published: Jaccard, R., Kosteas, D., Ondra, R.:
Background Document to Fatigue Design Curves for Welded Aluminum Components. IIW Doc.
No. XIII-1588-95. The enhanced data led to the European Recommendations for Aluminum
Alloys Structures (ERAAS) Fatigue Design Recommendations in 1992.
These evaluations were the baseline for the first drafts of the Eurocode 9 sections on aluminum
fatigue design in a third development phase, enhanced by further material (including welded
aluminum beam tests from other laboratories like TNO Delft/The Netherlands, The Swiss
Federal University in Lausanne, and ATLSS/Lehigh University, U.S.A.) and a number of
comparative analyses with other simultaneously emerging codes (International Institute of
Welding, IIW, on Fatigue Design, Aluminum Association Design Manual), as well as the
introduction of a number of new issues often following re-evaluation of data. During the final
three years when the completed codes were being compiled, a completely new approach and
change of format had to be followed. The European document EN 1990 states general provisions
for the quality management, defining consequence and reliability classes for structures,
definitions which ultimately lead to the adoption of respective execution classes for structural
components. This also led to the split of the former single document for fatigue into a document
for “design” [now EN 1999-1-3] and a document for “execution” (or manufacturing quality and
control) [now EN 1090-3]. This development was undertaken for steel as well.
The classification of weld quality through allowable imperfections was undertaken after the
document EN ISO 10042. Certain inconsistencies may arise in this procedure, as comparative
studies at TUM with respective national specifications have shown. The issue of relevance to
fatigue behavior is not yet completed for specific imperfections in this code. The quantification
of the quality classes and the harmonization of imperfection limit sizes will be one of the main
challenges in the coming years.
This report offers a compilation of important fatigue data and its evaluation as presented in the
European specifications. Much of the fatigue data used to develop the European standards is
proprietary data sponsored by private industry and is not available for general review. Instead,
comparative analyses and data in the form of S-N diagrams was developed and used in the codes.
The information for the European Standards is presented in four parts in this report:
1. Section 4 presents the data described above as it was evaluated for the European
Recommendations for Aluminum Alloy Structures, ERAAS, and later Eurocode 9.
2. Section 5 is a short comparison between the fatigue strengths calculated for the European
Recommendations (based on the comprehensive experimental data of over 25,000 data
points from small specimens and 2,500 data points from component tests) and various
other aluminum design codes including The Aluminum Association, Inc., Washington,
DC.
10
3. Section 6 summarizes, in tabular form, the background information for the definition of
design fatigue strength values from the available data.
4. Section 7 presents the same experimental results as above but in a comparison to the
International Institute of Welding, IIW, fatigue design rules for aluminum structures – it
should be noted that it represents the status of development at the end of 1999.
3.3 Friction Stir Welding
In addition to the discussion on the European Standards, Section 8 of this report also presents
general information and test results on Friction Stir Welding with some comparative test results
to similar welds developed using traditional fusion processes.
The report presents a general introduction to the processes, procedures and hardware associated
with FSW. It provides a comprehensive description of the mechanics underlying the FSW
process and an understanding of why the basic FSW process is better for fatigue sensitive
materials like aluminum compared to the traditional fusion welding procedures.
The report also presents fatigue data of 6XXX aluminum alloys using FSW with comparative
data to typical fusion welding procedures, as available.
11
4. EUROPEAN RECOMMENDATIONS FOR ALUMINUM ALLOY STRUCTURES
– FATIGUE DESIGN (ERAAS)
Early in the development of the European codes three groups of structural detail S-N curves were
established with respective slope values of 7.0, 4.3 and 3.4. As a remnant of the analysis
procedure for varying load spectra (especially from older fatigue design standards for steel
bridges with standardized loading and factors to accommodate other specific loading sequences
through the linear damage accumulation Miner-rule) the concept of parallel and equi-distant S-N
curve band can still be recognized in the Eurocode document. The following discussion and
figures help explain these assumptions.
4.1.1 Design S-N Curves in the ERAAS
All design S-N curves in ERAAS Fatigue are based primarily on experimental, full-scale
component data. Most of this data was generated in the mid life range up to 2×106 cycles.
Therefore, the assessment lines did not have to be conservatively extrapolated.
Design codes in the past were based almost exclusively on small specimen data in the lower
fatigue life ranges. In order to obtain a safe design line, especially in the fatigue-relevant region
around 2×106 cycles, this data had to be extrapolated and was conservative at longer lives. The
procedure to formulate such design lines was to fix it at the center of gravity of existing
experimental data - most often in an area around 104 and up to 105 cycles - and rotate it
downward at longer lives, maintaining a conservative philosophy. Due to the new, full-scale
component data at 2×106 cycles and above it was possible to formulate a “best-fit” design curve,
fixing it at the appropriate strength value as a lower data boundary, practically corresponding to
the mean minus two standard deviations strength value. A certain degree of uniformity for
practical reasons or concurrence with other existing design codes was attained through an
assumed slope, common for a group of structural details. The slopes are m=7.00 for parent
material and m=3.37 or m=4.32 for the welded details. These values have been calculated at two
characteristic stress-life pairs from the respective S-N data plots of individual datasets. This
concept was checked against some other options, common in other specifications. In general, it
can be demonstrated that neither an equi-distant parallel design line concept, Figure 1, nor other
slopes than those mentioned above would better represent the experimental data. Figure 2
demonstrates that because of the best-fit concept at 2×106 cycles a classification due to an equi-
distant line concept would lead to lower characteristic fatigue strength. A variation of other slope
values, for instance m=3.4, 3.8, 4.3 would “punish” the material for most of the structural details
under consideration in relation to the introduced best-fit design line concept.
By introducing these simplifications it was still possible to obtain characteristic fatigue classes,
as engineers are used to with other codes, with the advantage of a best-fit concept at the most
relevant life range. No problems in practice should be expected for the different slopes during the
design procedure even when calculating the equivalent damage of different spectrum loadings. In
today’s practice, computers are used for such calculations, all necessary equations are given
within the Recommendations and software for PCs is available to perform such fatigue
assessments.
13
60 N/mm²
55
50
53.2 Detail
47.2 Classification
1.128 = 3.33
characteristic 45
design values 41.8
10
ERAAS 40
35
37.1
equi-distant
32.9 design curve
30
25
29.1
25.8
mesh
23 22.9
20 20.3
JAKOON07.PRS 18 N/mm²
Figure 1 Equi-distant Design Curve Mesh
3.37
Stress Range
3.40 ERAAS
60.0 N/mm² design value = best-fit
3.80
GAIN i.e. upper possible limit
4.32
LOSS
4.30 next equi-distant
reference point
equi-distant 18.0 N/mm²

classification range
10E5 Cycles to Failure 2*10E6 5*10E6
JAKOON05.PRS
Figure 2 Effect of Classification Considering Equi-Distant and Parallel Design Line

Concepts on Attainable Fatigue Strengths
14
EQUI-DISTANT ERAAS Fatigue
PARALLEL compared at
m = 3.4 to 2*10E6 cycles
20
GAIN in %
10
0
-10
-20
LOSS in %
-30
E1/E3/E5-32.9
E2/E6/E8-22.9
B10-32.9
B11-29.1
E4/E7-18.0
B1-53.2
B1-47.2
B2-47.2
B2-41.8
B3-41.8
B3-37.1
B4-37.1
B4-32.9
B5-41.8
B6-37.1
B7-32.9
B8-29.1
B9-37.1
C1-60.0
C1-53.2
C1-47.2
C2-41.8
C2-37.1
D1-41.8
D1-37.2
D2-37.1
D2-32.9
D3-32.9
D3-29.1
F1-29.1
F1-25.8
F2-25.9
F2-22.9
F2-20.3
F3-20.3
F3-18.0
Ib
JAKOON09.PRS
Figure 3 Example of the effect of parallel and equi-distant design lines on fatigue strength
according to the different structural details after ERAAS Fatigue 1992
Using an equi-distant parallel S/N curve band result in the following:
At 2*106 cycles
• a “Loss”, observed in all cases, due to the new classification at the next respective
reference strength value of the equi-distant mesh, as this is, in general a lower value than
the ERAAS design value resulting from the “best-fit” concept.
At 105cycles and a parallel band slope of 3.40

• if the original ERAAS curve had a slope of 3.37 (which is practically identical to the
assumed value parallel band slope of 3.40) then the “Loss” at 105 cycles has the same
value as above at 2*106 cycles due to the new reference class value,
• if the original ERAAS curve had a slope of 4.32 (shallower than the assumed parallel
band slope of 3.40) then, depending on the classification value of the detail, the “Loss”
will be reduced compared to the value at 2*106 cycles or may even be turned into a
“Gain”
At 105cycles and a parallel band slope of 4.30

• if the original ERAAS curve had a slope of 3.37 (steeper than the assumed parallel band
slope of 4.30) then the “LOSS” will be enhanced, compared to its value at 2*106 cycles
• if the original ERAAS curve had a slope of 4.32 (which is practically identical to the
assumed parallel band slope of 4.30) then the “LOSS” will maintain its value
It is evident from these comparisons that there is no “GAIN” from a classification following an
equi-distant parallel band of S/N curves and the ERAAS design curve classification on the basis
of a best-fit curve value at 2x106 cycles serves the material far better in this area critical for
applications susceptible to fatigue. The only compromise was made in adopting two sets of
structural welded details with different slopes each, but parallel curves within the set. Whatever
“LOSS” appears compared to effective strength values - that is the design is too conservative for
short lives, where design is not generally affected by fatigue criteria.
15
4.1.2 Data Base of the Recommendations
The development of the ERAAS Fatigue document was based mainly on experimental fatigue
data of full-scale components. Results on small specimens were only used to investigate
tendencies of notch or R-ratio or plate-thickness influences. At the time, data was only provided
by Alusuisse-Lonza Services [3], Austria Metal (AMAG) [4] and TUM [5, 6] on various
aluminum alloys and welded structural details on extruded or built-up beams. A summary of this
database is given in Figure 4.
All data were stored in individual datasets describing only one structural detail with typical
manufacturing and loading characteristics, as part of the Aluminum Data Bank. These data sets
were processed either individually or grouped together in “families” and then analyzed
statistically. Approximately 160 individual data sets and 120 families have been analyzed for the
fatigue data of full-scale specimen. Detailed results are covered in [3].
Further decisions and the final ERAAS document design curves were based on the systematic
documentation, regression analysis and evaluation of all data, including small specimen data
from different institutions as well as the small specimen data generated at TUM studying
manufacturing variations in butt and fillet welded details. Results were summarized during the
two workshops in Munich and Zurich and were included, along with an international
comparative study on structural detail classification and fatigue strength values, in [7]. It is
mainly these results that form the background for the information in this report.
Further experimental data generated at other institutions will be similarly documented and
evaluated for the purpose of the new design standard Eurocode 9: Aluminum Design. Respective
analyses already allow the general statement of the validity of the ERAAS Fatigue Design
curves. In a few specific cases a modified or simplified structural detail classification may have
to be adopted. New data on welded beams, as well as small specimens, has been produced in the
last two years at:
• EPF-Lausanne by Hirt et al [18],

• ATLSS Laboratories at Lehigh University by Fisher and Menzemer [8], and
• TNO-Delft by Soetens et al. [19, 20, 21].
Only the experimental data produced at Lehigh is included in the following diagrams, and
compared to existing results by TUM.
16
Structural Detail ERAAS Alusuisse-Lonza AMAG TUM
Ref. No Number of Number of Number of
Data-points Data-points Data-points
A2 33
A4 72
A5 20
B5 44
B6
B7 30
B8 85
B9 28 17
B10 3
B11 40
C1
C2 17
D1 15 5
D2 25 88
D3 57
E1 116 7 159
E2 124
E3
E4 118
E5 58
17
E6
E7 53
E8 22
F1 22
F2
F3 27 30 69
Figure 4 Full-Scale Background Data for ERAAS-Fatigue Document
4.1.3 Base Metal 5000/6000 Series
ERAAS Detail Class A3 (simple extrusions) and A4 (Components)
Results for the 5000 alloys give values of 128 to 140 N/mm² for 2×106 cycles and 180 to 185
N/mm² for 1×105 cycles at R=0, with results approximated by 106 N/mm² at 2×106 cycles and
150 N/mm² at 1×105 cycles for R=+0.5. The general decision for simple extrusions/machined
parts for both 6000/5000 alloys is to use a design curve at 95 N/mm² at 2×106 cycles at R=+0.5
for no environmental effects, Figure 5. With a slope of m=7.00 the line shows a value of 146
N/mm² at 1×105 cycles. In the case of environmental effects (corrosion) a value of 67 N/mm² for
the 5000 series and 55 N/mm² for the 6000 series at 2×106 cycles and R=+0.5 and small
specimens are indicated.
Figure 5 Base Metal 6005A (AlMgSi0) Small Specimen Data
18
For base metal components an uppermost strength value of 87 N/mm² at 2×106 cycles and R=0
was indicated for both the 5000 and 6000 alloy beams. At lower lives the 5000 series exhibited
higher values. Some Alusuisse data show values of 95 N/mm² for extrusion profiles in the 6000
series. Taking into account the available information and considering respective factors in
transforming strength values from R=0 and R=-1 to R=+0.5, and dropping the former proposal of
different design curves for the two alloy groups the final decision was made for a design curve at
R=+0.5 with 70 N/mm² at 2×106 cycles and a slope of m=7.00, Figure 6.
Figure 6 Base Metal 5083 (AlMg4,5Mn) Full Scale Component Data
4.1.3.1 Butt Weld, Transverse
Simple Elements – ERAAS detail classes B1 (55 N/mm²), B2 (50 N/mm²), B3 (45 N/mm²) and
B4 (40 N/mm²) at 2×106 cycles cover various manufacturing qualities for transverse butt welds,
welded from one or both sides, with overfill dressed flush or intact and are based on small
specimen data [7]. The proposed design curves are maintained.
Extruded components – ERAAS detail classes B5 (45 N/mm²), B6 (40 N/mm²), B7 (35
N/mm²) and B8 (30 N/mm²) at 2×106 cycles cover various manufacturing qualities for transverse
butt welds, welded from one or both sides, with overfill dressed flush or intact and are based on
extruded shapes data [7]. The design curves of B6, B7 and B8 are maintained. Considering a re-
classification of detail B5 see the following information about detail B9 for built-up components.
Built-up Components – ERAAS detail classes B9 (40 N/mm²), B10 (35 N/mm²) and B11 (30
N/mm²) at 2×106 cycles cover various manufacturing qualities for transverse butt welds, welded
from one or both sides, with overfill dressed flush or intact and are based on built-up components
data [7] (beams with longitudinal welds connecting web to flange).
Comparing the data for cases B5 and B9, extruded and built-up components with butt weld
overfill ground flush, it is observed, that there respective scatter bands cannot be distinguished
for all practical purposes. It is also obvious from Figure 7 that even design curve B9 does not
19
cover the lower life range in a satisfactory way. Here a common design line with a shallower
slope is proposed as shown in Figure 8.
Figure 7 Transverse Butt Weld, Overfill Ground Flush, Extruded & Built-Up Components
Figure 8 Proposed Design Line for Transverse Butt Weld, Overfill Ground Flush,
Extruded & Built-Up Components
The respective diagrams for detail class B10 (35 N/mm²) and B11 (30 N/mm²), for welds from
both sides and one side only, are given in Figure 9 and Figure 10. The respective ERAAS
design curves are maintained.
20
Figure 9 One-sided Transverse Butt Weld on Built-up Components
Figure 10 Two-sided Transverse Butt Weld on Built-up Components
4.1.3.2 Butt Weld, Longitudinal
ERAAS detail class C1 (60 N/mm²) (ground flush) and C2 (45 N/mm²) (as welded)
Taking into account that a minimum design value of 60 N/mm² has been adopted for base
material, a design curve for detail class C1 at R=+0.5 with 60 N/mm² at 2×106 and a slope of
m=4.32 was adopted, leading to 120 N/mm² at 1×105 cycles, Figure 11. Care must be taken to
ensure a satisfactory weld root, so that fatigue cracks will not emanate from it. Appropriate
backing bars may be required.
The design curve for detail C2, Figure 12, is defined for R=+0.5 with a strength of 45 N/mm² at
2×106 cycles and a slope of m=3.37
21
Figure 11 Longitudinal Butt Weld, Overfill Ground Flush
Figure 12 Longitudinal Butt Weld, As Welded
4.1.3.3 Longitudinal Fillet Welds
Detail class D1 (45 N/mm²) for longitudinal fillet weld with no stop-starts, had very few data
points available. They lie in the scatter band of data points for detail class D2 (40 N/mm² -
longitudinal fillet weld with stop-starts) with a tendency to higher fatigue strength values. Figure
13 includes these data points for D1 together with data for D2. Some obvious outliers should be
checked through analysis of fractured surfaces for an explanation of possible imperfections
leading to their behavior. Figure 14 shows the test results for detail class D3 (35 N/mm²) for the
longitudinal intermittent fillet weld.
22
Figure 13 Longitudinal Fillet Weld with/without Stop-Starts
Figure 14 Intermittent Longitudinal Fillet Weld
4.1.3.4 Fillet Weld, Transverse, Non-Load-Carrying
ERAAS detail class E1 (35 N/mm²) (full and half stiffener), Figure 15.
Test results from Alusuisse [3] support values up to 45 N/mm². TUM beam test results of the
first program [4] lie at somewhat lower fatigue strengths, but the fact that they represent failures
at web stiffeners under loading points should be considered by calculating the principal stresses.
TUM beam test results of the second program [6] lie in the scatter band of the previous program,
especially at life ranges of 105 and 6×105 cycles. Sufficient data at high stress ranges had not
been tested at that time.
23
Figure 15 Fillet Weld, Transverse, Non-Load-Carrying (Web Stiffener)
Similarly, a comparison to TUM beams with a transverse vertical flange attachment from the
second program [6] was undertaken. Taking into account the principal stresses and that these
attachments are mounted at the outer-most part of the beam, the results fit into similar values for
web stiffeners.
For TUM results an R-ratio effect appears on high stress levels (Δσ ≈ 123 N/mm²). Longer lives
could be seen for R=-1 compared to R=+0.1. A different behavior compared between full
stiffeners welded on both flanges and half web stiffeners welded only on the compression flange
cannot be seen. Due to the lack of information for lives above 106 cycles a design curve with 35
N/mm² at 2×106 cycles and a slope of m=3.37 was proposed, Figure 15.
Figure 16 Transverse Fillet Weld, Non-Load-Carrying (Web Stiffener), ATLLS-Lehigh

Test Results
24
Figure 16 shows the results from [8] which complete and verify existing TUM test data and
establish the long life range fatigue behavior of the detail.
4.1.3.5 Web Attachments
ERAAS detail class E2-23 N/mm² (rectangular, circular, hollow shapes)
Figure 17 Web Attachments
Proposed fatigue strength values are based on ALS beam test results. These support the values at
2×106 cycles for both rectangular shapes and round tubes welded to the beam web. A design line
with 23 N/mm² at 2×106 cycles and a slope of m=3.37 was adopted, Figure 17.
4.1.3.6 Attachment at Edge of Flange
ERAAS detail class E3 (35 N/mm²) (with transition radius, r). The design line is based on
Alusuisse beam data with higher actual values especially for the higher life range, Figure 18.
Two data points have not been taken into account in fixing the design line as they seem to be
outliers.
25
Figure 18 Attachment at Flange Edge, with transition radius, r>50 mm
Figure 19 Attachment at Flange Edge, no transition radius
The values of the design proposal of the European Aluminum Association, EAA, have been
considered and a value of 18 N/mm² at 2×106 cycles with a slope of m=3.37 have been adopted
for E4, without any transition radius to the flange attachment, Figure 19.
4.1.3.7 Vertical Attachment on Flange
ERAAS detail class E5 (35 N/mm²), longitudinal on extruded beam with transition radius. No
actual data available, design line assumed in accordance to attachment at flange edge, detail E4.
Design curve maintained but experimental verification is desirable for extruded or built-up
beams.
26
ERAAS detail class E6 (23 N/mm²), longitudinal on extruded beam without transition radius and
ERAAS detail class E7 (18 N/mm²), longitudinal on built-up beam without transition radius. S-N
diagrams indicate somewhat higher fatigue strength values for Alusuisse test results on extruded
beams in comparison to TUM test results on built-up beams, Figure 20. Observing all data
points, especially those in the region around 2×106 cycles resulting from Alusuisse tests with
extruded beams, the proposal to treat both extruded and built-up beams with a single design
curve appears to be a logical solution, Figure 21.
ERAAS detail class E8 (23 N/mm²), transverse on built-up beam. The design curve is based on
the evaluation of a total of 22 data points distributed on three stress levels. It has been considered
to propose a design line at 25 N/mm² with a slope of 4.32, Figure 22.
Figure 20 Vertical Attachment (Long’l) - Flange of Built-up Beam, no transition radius
Figure 21 Vertical Attachment (Long’l) - Flange of Built-up/Extruded Beam, no transition

radius
27
Figure 22 Transverse Attachment Vertical on Flange, Built-up Beams
4.1.3.8 Cruciform Joint
ERAAS detail class F1, transverse load-carrying fillet weld/cruciform joint (toe crack failure).
The design curve is based on TUM beam tests from 1986 indicating values of 35 to 40 N/mm²
for R=-1 at 2×106 cycles for the alloy 5083. Grouping all alloys for different R values at 30
N/mm² was proposed. This value is true for full penetration, butt-weld-like joints or for double
fillet welds with failures at toe cracks. This pattern of behavior is supported in a satisfactory
manner also by the TUM small specimen tests of 1991 [7]. ERAAS detail class F2, transverse
load-carrying fillet weld/cruciform joint (throat crack failure). The design curve is based on
Alusuisse data with small specimens supports 35 N/mm² at 2×106 cycles for R=0. There are also
similar TUM beam test results from 1986. Other test results from TUM small specimens on
cruciform joints from 1991 give a value of 28 N/mm² for R=+0.1. For design purposes a value of
25 N/mm² at 2×106 cycles with a slope of 4.32 was defined, Figure 23.
Figure 23 Cruciform Joint, Built-up Beams
28
4.1.3.9 Cover Plate
ERAAS detail class F3, transverse load-carrying fillet weld/cover plate. Analysis of both TUM
beam test results on the 7020 alloy, taking into account reduction factors for fatigue strength
values at R=+0.6, indicate a value of 20 N/mm² at 2×106 cycles, being adopted for design
purposes with a slope of m=4.32, Figure 24.
Figure 24 Cover plate, Built-up Beam, TUM Test Results
Further data were provided by Fisher/Menzemer, Lehigh University [8], Figure 25. Data points
were generated up to 1×107 cycles, verifying and completing the existing TUM results and
establishing the fatigue behavior of this detail, especially for the long life range, Figure 26.
Figure 25 Cover plate, Built-up Beam, Lehigh Test Results
29
Figure 26 Cover plate, Built-up Beam, Lehigh and TUM Test Results
30
5. COMPARISON BETWEEN ERAAS AND OTHER CODES
This section of the report presents a brief comparison of the data in ERAAS and various other
aluminum codes. Of particular note are:
• Aluminum Association Specification for Aluminum Structures

• British Standard BS 8118 Structural use of Aluminum
• Association of American Railroads
• Ontario Highway Bridge Code
Discussion is also presented for various steel codes to help compare the behavior of steel and
aluminum in the fatigue environment.
5.1 Comparison to Aluminum Codes
Several Aluminum Codes [9, 10, 11, 12, 13] were compared to the ERAAS-Fatigue document.
The following diagrams in Figure 27 show the design values at 2×106 cycles. This comparison
indicates the historical development of respective recommendations, but, except for the recent
edition of the British document BS 8118-1992, it is pointed out that ERAAS has been based on
the evaluation of well documented data on full-size components after a homogeneous statistical-
regression evaluation.
The comparison between the design values of ERAAS and BS 8118, for instance at 2×106
cycles, for the different structural details in both codes is demonstrated in Figure 28.
31
100
140
120 80
100
60
80
60 40
40
20
20
0 0
A1
A2
A3
A4
B1 B2 B3 B4 B5 B6 B7 B8 B9 B10 B11
ERAAS BS 8118 Al. Assoc. ERAAS BS 8118 Al. Assoc.
Ass. American RR. Ontario HB Code LdV-DVS 1608 Ass. American RR Ontario HB Code LdV DVS 1680
100
80
60
40
20
0
E1 E2 E3 E4 E5 E6 E7 E8
ERAAS BS 8118 Al. Assoc.
Ass. American RR Ontario HB Code LdV DVS 1608
Figure 27 Comparison of ERAAS Fatigue to Various Aluminum Codes
N/mm²
70
65
60
55
50
45
40
35
30
25
20
15
10
5
0
-B9
-E5
3.7-F1
3.7-F1
2.7-F3
1.1-A1
1.1-A2
1.1-A3
1.1-A4
1.2-A5
1.3-A5
1.4-A5
3.1-B1
3.2-B2
3.3-B3
3.4-B4
3.6-B5
3.5-B6
3.5-B7
3.5-B8
2.1-C1
2.2-C2
2.2-D1
2.3-D2
2.4-D3
2.9-E2
2.6-E6
2.6-E7
-E3
-B10
-B11
3.11-F2
2.10-E1
2.11-E4
2.10-E8
1.5-
1.6-
1.7-
3.9-
2.8-
3.10-
2.5-
Detail Classification BS 8118 / ERAAS FAT

JAKOON12.PRS
Figure 28 Comparison of ERAAS Fatigue to BS 8118
32
5.2 Comparison of ERAAS Fatigue 1992 & Aluminum Association Design Manual
The information in Figure 29 and Figure 30 focuses on the comparison between ERAAS and
the Aluminum Association for fatigue life of detail classes as 2 x 106 cycles. Table 3 provides
the description of the various detail classes between the two standards.
Comparison betw een ERAAS FAT 1992 and

AA Manual Draft Fatigue Design Stresses
N/mm² 130 95 117
at 2E6 Cycles
90 butt we ld
85
80 extrusion
75 2x106 cycles
70 simple
65 built-up
60 element component
55
50
45
40
35
30
25
20
15
10
5
0
- E8
A3 - A!,A2
B4 - C9 to 12
A5 - C7,D7,E7
A1 - A1,A2
A2 - A1,A2
A4 - A1,A2
C2 - B3/4/5
D1 - B3/4/5
D2 - B3/4/5
B1 - B9
B2 - B11,B12
B3 - B11,B12
E3 - BCD13
E5 - BCD16
E1 - C6,C21
D3 - E?
F3 - E5,E15
B5 -
B6 -
B7 -
B8 -
B9 -
C1 -
F1 - E17
B10 -
B11 -
E8 - C19
E4 - DE14
E7 - D14,E19
E2 -
E6 -
F2 - F5,15,17
A5 -
A5 -
A5 -
R-dependency
Figure 29 Comparison of ERAAS & Aluminum Association Fatigue @ 2x106 Cycles

Spann.-Doppelampl. [N/mm²]
N/mm²
100 80
z
45 A
z
36 z
z 22 B
17 z
C
10
D
F
E?
1
1E+04 1E+05 1E+06 1E+07 1E+08
Lastspiele N
Figure 30 Comparison of ERAAS & Aluminum Association S/N Curves
33
Table 3 Correlation between ERAAS and Aluminum Association Details
Detail ERAAS AA
Base Metal Elements A3 A1,A2
5000/6000 Components A4 A3,A4
Notches, Holes A5 CDE7,E8
Butt Weld flush B1 B9
Transverse > 150° B2 B11/12
Simple Element > 130° B3 B11/12 ?
backing B4 C9 to 12
Butt Weld flush B5
Transverse > 150° B6
Extruded > 130° B7
B8
Butt Weld flush B9
Transverse > 150° B10
Built-Up B11
Butt Weld flush C1
Longitudinal > 150° C2 B3/4/5
Fillet Weld no interruptions D1 B3/4/5
Longitudinal stop-starts D2 B3/4/5 ?
intermittent D3 E?
Web Stiffener fillet transverse E1 C6/21
Web Attachment E2
Attachment at transition ∅ E3 BCD13
Flange Edge no transition ∅ E4 DE14
Attachment on extruded E5 BCD16
Flange - Vertical transition ∅
& Longitudinal extruded E6
no transition ∅
built-up E7 E19
no transition ∅
Attachment on built-up E8 C19
Flange - Vertical
Transverse
Fillet-Transverse toe crack F1 E17
Cruciform throat crack F2 F5/15/18
Cover plate F3 E5/15/20
5.3 Comparison of ERAAS to Various Steel Codes
Several Steel Codes [14, 15, 16, 17] were also compared to the ERAAS Fatigue document. The
diagrams in Figure 31 show the comparison of design values at 2×106 cycles.
34
Steel codes have been based entirely on small specimen data. The Eurocode 3: Steel Design
presents a design concept based on stress range Δσ with no R-ratio dependency. All other cited
steel codes give maximum stress amplitude maxσa value with an R-ratio dependency. These
values had to be transformed accordingly for an R-ratio of R=+0.5 (as given in ERAAS Fatigue
for the basic design curve).
200 140
180
160 120
140 100
120 80
100
80 60
60 40
40 20
20
0
0
B1
B2
B3
B4
B5
B6
B7
B8
B9
B10
B11
A1
A2
A3
A4
ERAAS Eurocode EC3 DIN 15018 DS 804 DASt Ri 011 ERAAS Eurocode EC3 DIN 15018 DS 804 DASt Ri 011
140 140
120 120
100 100
80 80
60 60
40 40
20 20
0 0
C1
C2
D1
D2
D3
F1
F2
F3
E1
E2
E3
E4
E5
E6
E7
E8
ERAAS Eurocode EC3 DIN 15018 DS 804 DASt Ri 011 ERAAS Eurocode EC3 DIN 15018 DS 804 DASt Ri 011
Figure 31 Comparison of ERAAS Fatigue to Various Steel Codes
It is interesting to observe the ratio between design values for aluminum and steel as given in
different documents. A ratio of 3:1 (based on the ratio of the modulus of elasticity of the two
materials) between steel and aluminum has been previously stated. A direct comparison of
fatigue design values at 2×106 cycles between the Eurocode 3: Steel Design and ERAAS Fatigue
(Aluminum), Figure 32, shows that for a majority of structural details a value at or below 2.3:1
appears. Only in a single case is a value above 3:1 observed.
35
JAKOON03.PRS
Figure 32 Ratio of Fatigue Design Values (2×106 Cycles)-ERAAS (Al) & Eurocode 3 (Steel)
36
6. TABLES FOR DETAIL CATEGORIES AFTER ERAAS FATIGUE 1992
DECISIONS LEADING TO FATIGUE STRENGTH DESIGN VALUES
The tables in this section of the report present a brief summary, discussion and justification for
using the fatigue strength design values relative to ERAAS.
Table 4 Butt Weld Transverse - Simple Specimen & Extruded Component (ERAAS B1-B8)
Detail ERAAS Further Data Final Remarks
B1 - 55 N/mm² TUM small specimens ERAAS design curve is
Butt weld, transverse 1991 & ALS small maintained.
overfill dressed flush specimens. See TUM /
from both sides A-L study 1991 Part 6,
simple element Doc. A1 and Doc. Z61
to Z92
B2 - 50 N/mm² Conceived as an ERAAS design curve is
Butt weld, transverse interpolation between maintained.
overfill >150° B1 and B3, based
from both sides mainly on ALS small
simple element specimens.
B3 - 45 N/mm² Based on ALS small ERAAS design curve is
Butt weld, transverse specimens. See TUM / maintained.
overfill > 130° A-L study 1991 Parts 2,
from one or both sides 3 & 6, Doc. Z61 to Z92
simple element and Doc. M34 to M40.
B4 - 40 N/mm² TUM small specimens. TNO small specimens ERAAS design curve is
Butt weld, transverse See TUM / A-L study V-weld from one side, 6 maintained although
from one side only on 1991 Part 6, Doc. A2- + 12 mm. data on 24 mm falls on
permanent backing bar A3-A4. EU 269 1994, AlDaBa or slightly below line.
simple element data set no. T0004.0 and
T0005.0. All TNO 6 & 12 mm
welds as well all INEGI
INEGI small specimen 12 mm welds may be
12 + 24 mm. pooled & belong to the
EU 269 1994, AlDaBa same scatter band which
data set no. T0006.0 and allows sufficient safety
T0007.0. margin to design curve.
INEGI 24 mm welds are
significantly lower than
all other. This may
compare to jakoonI- indicate a different type
08.prs of imperfection.
B5 - 45 N/mm² ALS beams, Doc. M44 New ERAAS design
Butt weld, transverse to M46 or Doc. A7-A8. curve proposed - with
overfill dressed flush TUM / A-L study 1991, designation B12 - 45
from both sides Parts 2 and 6. N/mm² & m = 7.00.
extruded components
B6 - 40 N/mm² ALS beams (Cosandey ERAAS design curve is
Butt weld, transverse short beams). maintained.
overfill > 150° TUM / A-L study 1991,
37
from one or both sides Parts 2 and 6, Doc. M31
extruded components to M33.
B7 - 35 N/mm² ALS beams. ERAAS design curve is
Butt weld, transverse TUM / A-L study 1991, maintained.
overfill > 130° Parts 2 and 6, Doc.
from one or both sides M29-M30.
extruded components
B8 - 30 N/mm² ALS data sets no. 7184, ERAAS design curve is
from one side only 7186, 7188 in TUM maintained.
without perman. analysis of EAA-COST
backing 506 project 1989.
extruded components
Table 5 Butt Weld Transverse – Built up Components (ERAAS B9 to B11)

B9 - 40 N/mm² TUM beams 1991. New ERAAS design
Butt weld, transverse TUM / A-L study 1991, curve proposed - with
overfill dressed flush Parts 2 and 6, Doc. designation B12 - 45
from both sides M41-M42 or A5-A6. N/mm² & m = 7.00.
built-up components
B10 - 35 N/mm² TUM beams 1986. Data TNO beam data. ERAAS design curve is
Butt weld, transverse sets no. B8001, B8002, EU 269 1994 program, maintained.
overfill > 150° B8003 and B8028, AlDaBa data set no.
from both sides B8029, B8030. T0008.0. Only 3 data
built-up components TUM / A-L Study 1991, points generated fitting
Parts 2 and 6, Doc. M24 into the general pattern
+ M27. of existing data. Exact
description of weld
quality not available,
detail classification
taken as mentioned in
report.
compare to jakoonI-
07.prs
B11 - 30 N/mm² TUM beams. Data sets ERAAS design curve is
Butt weld, transverse no. B8031, B8032, maintained.
from one side only B8033, B8034, B8035,
without perman. B8036, B8037.
backing TUM / A-L study 1991,
built-up components Parts 2 and 6, Doc. M25
+ M26.
38
Table 6 Butt Weld – Longitudinal (ERAAS C1 & C2)
C1 - 60 N/mm² Design curve based on ERAAS design curve is
Butt weld, longitudinal ALS beam data maintained.
overfill ground flush (Cosandey) taking into
account as an upper
limit the fact that a
minimum design value
of 60 N/mm² has been
fixed for base material.
TUM / A-L study 1991,
Parts 2 and 5, Doc.
M47b.
The above are also in

accordance to values
proposed in the EAA-
COST 506, March 1989
document.
C2 - 45 N/mm² Design curve values ERAAS design curve is
Butt weld, longitudinal based on ALS beam maintained.
overfill > 130° data (Cosandey) as
proposed in TUM / A-L
study 1991, Parts 2 and
5, Doc. M47b. Also in
accordance with values
proposed in the EAA-
COST 506, March 1989
document.
Table 7 Fillet Weld – Longitudinal (ERAAS D1, D2 & D3)

D1 - 45 N/mm² Design curve based on ERAAS design curve is
Fillet weld, longitudinal TUM beams 1991. maintained
no interruptions, no
stop-start positions, no
tack welds
Fillet weld, longitudinal TUM beams 1986 1991. maintained, although a
with stop-start positions few data points fall
or tack welds below it. These should
be checked as outliers.
Fillet weld, longitudinal TUM beams 1991. maintained.
intermittent
39
Table 8 Fillet Weld – Transverse (non-load carrying), Web Stiffeners & Attachments
(ERAAS Design Curves E1 & E2)
Detail ERAAS Further Data Final remark
E1 - 35 N/mm² ALS Beams Fisher/Menzemer Due to scant data at
Fillet weld, transverse,
TUM/A-L Doc M.48 and Lehigh beams 1993 longer lives a value of
web stiffener M.49 support values up to 35 N/mm² was adopted.
extruded / built-up 45 N/mm² for 2*106, esp. if Results complete No significant
beam the raise of 10% for stresses and verify existing difference in behavior
at inner flange side and a TUM test data and between full stiffeners
levelling-off of the S-N establish especially welded on both flanges
curve is assumed. for the long life and half stiffeners
ALS small specimens range the fatigue welded only on the
TUM/A-L Doc Z.105 and behavior of the compression (due to
Z.114 give min value of 63 detail. external loading) flange
N/mm² at R=0 resulting in has been observed.
50 N/mm² at R=+0.5. Lehigh data verifies
TUM Beams 1986 behavior in the long life
7020 & 5083 lie in the range.
scatter band at tested life
range between 105 to 5*105 Two data points have
TUM beams 1991 not been taken into
somewhat lower stresses account in fixing the
than ALS beams, but fact design line as they seem
should be considered that to be outliers (also in
web stiffeners were loaded. the sense of regression
Beam data with transverse evaluation) and their
flange attachment fits in at fracture surfaces should
105 (accounting for nominal be analyzed for any
stress at crack site), but no imperfections before
more at longer lives. final classification.
Data from both TUM
programs show R- The ERAAS design
dependency, i.e. longer curve is maintained.
lives at R=-1, for higher
stress levels.
E2 - 23 N/mm² Based on ALS beam tests ERAAS design curve is
Web attachments, round supporting values both for maintained.
or rectangular shapes rectangular shapes and
round tubes.
40
Table 9 Attachment at Flange Edge (ERAAS E3 &E4)
E3 - 35 N/mm² Based on ALS beam ERAAS design curve is
Attachment at flange data. A best-fit proposal maintained.
edge with transition according to Design Two data points have not
radius R>50 mm Proposal EAA as of been taken into account in
26.07.89 gave values up fixing the design line as
to 41 N/mm² at 2*106. they seem to be outliers
A final curve near to the (also in the sense of
COST 506-EAA study regression evaluation) and
was adopted. their fracture surfaces
should be analyzed for
any imperfections before
final classification.
E4 - 18 N/mm² Based on ALS beam There seems to be a
Attachment at flange data. certain problem at the
edge Best-fit proposal after lowest tested level with
no transition radius Design Proposal EAA data points on the
as of 26.07.89 was assumed design curve E4,
reconsidered and values one point being also lower
nearer the COST 506- but another point on the
EAA analysis adopted. same level and at same
lives appearing as a run-
out. A fracture surface
analysis could provide an
answer for unusual
imperfections.
Either careful design and

manufacturing of this
detail should be required
or the design curve has to
be lowered to approx. 16
N/mm².
Such a detail should not

be allowed. It is bad
design and the significant
enhancement in strength
is demonstrated by detail
E3.
41
Table 10 Vertical Attachment on Flange – Longitudinal (ERAAS E5, E6 &E7)
E5 - 35 N/mm² No data available. ERAAS design curve
Attachment on flange, Design curve in maintained.
vertical, longitudinal accordance to Experimental
transition radius ≥ 50 attachment at flange verification desirable,
mm edge, at first for for extruded and/or
extruded beam extruded beam only. built-up beam.
E6 - 23 N/mm² ALS beam test results EPFLausanne/ Beam Data ERAAS design curve
Attachment on flange, with slope m=3.00 INALCO´92 E6 has to be corrected
vertical, longitudinal (Doc Z-104 A- all constant amplitude data by assuming (3 options)
no transition radius L/TUM) with run-outs at 108 Either:
extruded beam (including variable ampl. new E6 - 20.0 N/mm²
data up to 5*106) falls with slope m=3.37
within scatter band of data or
of Doc Z-104 - variable E6=F3 - 20.0 N/mm²
amplitude data from 5*106 with slope m=4.32
to 5*107 though, falling on or
or slightly below ERAAS E6=E7 - 18.0 N/mm²
E6 line. and slope m=3.37
The latter being
Maddox / INALCO ´82 proposed, abandoning
small specimen data the original distinction
Values slightly below between extruded and
ERAAS curve at 2*106 built-up beams.
Kosteas / VA H5 1971 Classification on the

6 mm 7020 small basis of nominal
specimen data at PS=90% stresses calculated at
(slope 3.60) approx. 3-fold crack site, i.e. on outer
safety margin to ERAAS flange side, but without
curve at lower and taking into account
medium life range geometrical effect of
attachment on cross-
section values (moment
of inertia, relocation of
center of gravity). The
respective stress
reduction would be in
the order of 50% in the
case of TUM beams, or
80% in the case of
EPFL beams.
E7 - 18 N/mm² TUM beam test ERAAS design curve
Attachment on flange, results with slope maintained
vertical, longitudinal m=3.27.
no transition radius
built-up beam
42
Table 11 Vertical Attachment on Flange – Transverse (ERAAS E8)
E8 - 23 N/mm² TUM Beams 1991 TNO/INEGI 1994 ERAAS design curve
Attachment on flange, small specs. 12 mm, may be maintained.
vertical, transverse R=+0.1 significantly One could even think of
built-up beam above ERAAS curve proposing a design line
beams, only 3 data at 25 N/mm² with a
points, falling in scatter slope of 4.32
band of small specs
INEGI results higher
than TNO
results verify former
TUM beam tests, with
somewhat shallower
slope though (3 points
only!)
Maddox INALCO ´82

small specimens
Maddox &Webber 1987

effect of high residual
stresses even in small
specimens
Kosteas 1971
small specimens, all
data significantly high
strength values, espec.
for 6 mm
compare to jakoonI-
01.prs
43
Table 12 Fillet Weld–Transverse (load carrying) Cruciform, Cover plate (ERAAS F2, F3)
F1 - 30 N/mm² TUM beams 1986 - 35 ERAAS design curve is
Fillet weld, transverse to 40 N/mm² for R= -1 maintained.
load-carrying, cruciform (5083) at 2*106. Group
toe-crack failure alloys for various R-
ratios indicate values at
30 N/mm², true for full
penetration butt-weld-
joints or double fillet
welds with failures at
toe cracks. These results
supported by the TUM
small specimen, 1991.
TUM / A-L study 1991,
Parts 5 and 6.
F2 - 25 N/mm² ALS data on small TNO small specimens ERAAS design curve is
Fillet weld, transverse specimens support 6+12 mm, constant maintained.
load-carrying, cruciform values of 35 N/mm² for ampl., AlDaBa data set
throat-crack failure R=0 at 2*106. no. T0012.0 & Z0013.0. According to the report
TUM beams 1986. See TNO/Dutch Rail small the characteristic
INALCO ´95 paper by specimens classification is under
Jaccard/Kosteas/Ondra. 6+12+24 mm, variable detail F2 there seem to
TUM small specimen ampl., AlDaBa data set be no problems in
data supports values of nos. T0016.0, T0017.0 relation to the existing
28 N/mm² for R=+0.1 at and T0018.0. ERAAS design curve
2*106. TUM / A-L INEGI small specimens values.
study 1991, Parts 5 and 12+24 mm, constant
6. ampl., AlDaBa data set
nos. T0014.0 and
T0015.0
TNO beams
6+12+24 mm, constant
ampl., AlDaBa data set
nos. T0019.0, T0020.0
and T0021.0.
All EU 269 1994
reports.
Kosteas small specs
1971 significantly
higher than results for
beam details.
F3 - 20 N/mm² TUM beams 1986 and Fisher/Menzemer, ERAAS design curve is
Fillet weld, transverse TUM beams 1991 Lehigh 1993 maintained.
load-carrying Results complete and
cover plate verify TUM results and
establish the fatigue
behavior for long life.
44
7. COMPARISON OF FATIGUE STRENGTH VALUES FOR STRUCTURAL
DETAILS
The information presented below addresses comparative data from ERAAS FAT, Eurocode 9
ENV 1999-2 (predecessor of current document prEN 1999-1-3) and assumptions for the
International Institute of Welding, IIW, fatigue design rules.
Table 13 Comparison of Design Curve Fatigue Strength Values for Transverse Butt Welds
to Experimental Data (at 2*106 cycles)
Detail ΔσC-m1 Detail ΔσC-m1 ΔσC-m1 Detail ΔσC-m1 ΔσC-m1 See Fig.*
ERAAS 1992 Eurocode 9 IIW Recomm. 1996
Bead ground off, welded from both sides
B1 55-7 3.1 55-6 44-5 211 50-3
B5/12 45-7 3.1 44-5 28-4 a)
B9/12 45-7
Bead angle >150°, welded from both sides
B2 50-4.32 3.2 39-4 35-4 212 40-3
B6* 35-3.37 3.2 35-4 28-4 b)
B10 35-3.37
Bead >130°, welded from (one or) both sides
B3 45-4.32 3.2 29-3.2 18-3.2 212 40-3
213 32-3 c)
B7 35-3.37
Welded from one side only, with or without permanent backing
B4 40-4.32 3.3 35-4 25-3.2 215 25-3
225 22-3
216 28-3 d)
B8 30-3.37 3.4 29-3. 18-3.2
B11 30-3.37
Welded from one side only, lack of penetration (LOP), root defect
3.5 14-3.2 (216) 18-3
ERAAS: bold lines / EC9: fine lines / IIW: dash-dot lines
* Please note that all Figures in this section of the report are referenced by the letters “a” through
“e”, with various letters used more than once. Each figure is related to the specific table that
cites the reference and the letters can be found in the lower left hand corner of each figure.
45
ERAAS / IIW Doc. XIII-1588-95 Transverse Butt, Overfill Dressed Flush Both Sides
B1 (55-7) Simple Element or B5/12 (45-7) Extrusion and B9/12 (45-7) Built-Up Component
EC 9 Welded Butt Joint Ground Flush Cat. 3.1 (55-6) or (44-5) for flats and solids
(44-5) or (28-4) for open shapes
not applicable for hollow sections
IIW Recom 96 Det. No. 211 (50-3,0) Transverse Butt (X or V) Ground Flush
stress range [N/mm²] NC ND N L
300
at 2E6 cycles
% %
data components (beams) ERAAS 100 ERAAS 100
EC 9 80 EC 9 62
100 IIW 91 IIW 111 ?
extrusions sm all
50 specim ens
10 a)
5
1E+04 1E+05 1E+06 1E+07 1E+08 1E+09
ms.dos_6/(eigene)
cycles to failure N FINAL DRAFT prENV 1999-2 April 1997
daten/praes/inalco98/inalco.pr4-no11
ERAAS / IIW Doc. XIII-1588-95 Transverse Butt Overfill>150°, Both Sides

Simple Element B2 (50-4,32);Extruded Component B6 (35-3,37);
Built-Up Component B10 (35-3,37)
EC 9 Welded Butt Joint Double Sided Cat. 3.2 (39-4) or (35-4) for flats and solids
IIW Recom 96 Transverse Butt, Both Sides Det. No. 212 (40-3,0) >150°
stress range [N/mm²] NC ND N L at 2x1E06
300
data components 50-4,32

45-4,32
100 (beams) 35-3,37
50 (39-4)
35-4,0
28-4,0
40-3,0
10 b)
5
1E+04 1E+05 1E+06 1E+07 1E+08 1E+09
ms.dos_6/(eigene)
46
ERAAS / IIW Doc. XIII-1588-95 Transverse Butt Overfill >130°, Welded from ( One) or Both
Sides Simple Element B3 (45-4,32); Extruded Component B7 (35-3,37)
EC 9 Welded Butt Joint Double Sided Cat. 3.2 (39-4) or (35-4) for flats and solids
IIW Recom 96 Transverse Butt Both Sides Det. No. 212 (40-3,0) >150° and No. 213 (32-3,0) >130°
300
data components
100 (beams)
45-4,32
35-3,37
data small specimens
50 (39-4)
35-4,0
28-4,0
40-3,0
c)
32-3,0
10
5
1E+04 1E+05 1E+06 1E+07 1E+08 1E+09
ms.dos_6/(eigene)
ERAAS / IIW Doc. XIII-1588-95 Transverse Butt / One Side Only

Simple Element B4 (40-4,32) On Permanent Backing Bar
Extruded Component B8 (30-3,37) Without Permanent Backing Bar
Built-Up Component B11 (30-3,37) Without Permanent Backing Bar
EC 9 Welded Butt Joint Single Sided Cat. 3.3 (35-4) or (25-3,2) Backed
or Cat. 3.4 (29-3,2) or (18-3,2) Unbacked
IIW Recom 96 Transverse Butt Det. No. 215 (25-3,0) Perm. Backing and No. 225 (22-3,0)
300
data small
specimens
40-4,32
100 30-3,37
35-4,0
50 data components 29-3,2
(beams) 25-3,2
18-3,2
25-3,0
10
d) 22-3,0
5
1E+04 1E+05 1E+06 1E+07 1E+08 1E+09
ms.dos_6/(eigene)
47
Table 14 Comparison of Design Curve Fatigue Strength Values for Parent Material and
Experimental Data (at 2*106 cycles)
Detail ΔσC-m1 Detail ΔσC-m1 ΔσC-m1 Detail ΔσC-m1 ΔσC-m1

Parent material
7020 – simple extrusions, mechanically formed parts, components
A1 130-7 1.1 121-7
A2 85-7 1.3 96-7 111 80-5
122 ? 40-3
5000/6000 - simple extrusions, mechanically formed parts, components
A3 95-7 1.2 86-7
A4 70-7 1.4 69-7 111 71-5
122 ? 40-3
Table 15 Comparison of Design Curve Fatigue Strength Values for Longitudinal Butt and
Longitudinal Fillet Welds and Experimental Data (at 2*106 cycles)
Detail ΔσC-m1 Detail ΔσC-m1 ΔσC-m1 Detail ΔσC-m1 ΔσC-m1 See Fig.
Longitudinal butt welds
Continuous, no stop-starts, bead ground off
C1 60-4.32 2.14 60-4.5 55-4.5 311 50-3
312
Continuous, no stop-p-starts
C2 45-4.32 2.15 44-4.5 313 45-3 a)
With stop-starts
2.16 35-4 311 36-3
313
Longitudinal fillet welds
On one or both sides, continuous, no stop-starts
D1 45-4.32 2.15 322 40-3
With stop-starts
D2 40-4.32 2.16 323 36-3 b)
Intermittent weld
D3 35-4.32 2.17 324 32 ? -3
Weld toe at hole (or notch)
2.18 325 28 ? -3
48
ERAAS / IIW Doc. XIII-1588-95 Longitudinal Butt WeldOne / Side Only / Continuous,No Stop-Start
Built-Up Component C1 (60-4,32) Ground Flush or C2 (45-4,32) Overfill >130°
EC 9 Longitudinal WeldSingle Sided Cat. 2.14 (60-4,5) or (55-4,5) Ground Flash, Continuous
or Cat. 2.15 (44-4,5) Continuous
Cat. 2.16 (35-4) Stop-Start
IIW Recom 96 Longitudinal Butt Weld Det. No. 311/312/313 (50/45-3,0) Continuous
No. 311/313 (36-3) With Stop-Start
300
ground off
60-4,32
100 45-4,32
data components (beams)
60-4,5
50 55-4,5
44-4,5
35-4,0
50-3,0
a)
45-3,0
10 35-3,0
5
1E+04 1E+05 1E+06 1E+07 1E+08 1E+09
ms.dos_6/(eigene)
ERAAS / IIW Doc. XIII-1588-95 Fillet Longitudinal / One or Both Sides

D1 (45-4,32) Continuous, No Stop-Start / or D2 (40-4,32) With Stop-Start
D3 (35-4,32) Intermittent
EC 9 Attachment With Fillet Longitudinal Cat. 2.15 (44-4,5) No Stop-Start; Cat. 2.16 (35-4) With
Stop-Start; (Weld Toe) Cat. 2.17 (31-3,5) Intermittent or Cat. 2.18 (28-3,5) Cope Hole
IIW Recom 96 Fillet Longitudinal Det. No. 322/323 (40/36-3) Continuous
No. 324 (32 ? -3) With Stop-Start / No. 325 (28 ? -3) Cope Hole
300
data components (beams) w ith stops-starts 45-4,37

40-3,37
100 intermittent 35-3,37
44-4,5
50 35-4,0
31-3,5
28-3,5
40-3,0
b)
36-3,0
10 32-3,0
28-3,0
5
1E+04 1E+05 1E+06 1E+07 1E+08 1E+09
ms.dos_6/(eigene)
49
Table 16 Comparison of Design Curve Fatigue Strength Values for Transverse Fillet Welds
and Experimental Data (at 2*106 cycles)
Transverse Fillet Weld
Cruciform, load-carrying fillet, crack at weld toe
F1 30-4.32 (3.6) 412 25-3
(3.7)
2.3 25-3.2
(2.9) a)
2.4 22-3.2
Cruciform, load-carrying fillet, crack through weld (root)
F2 25-4.32 3.8 18-3.2 414 16-3
Cover plate, transverse fillet load-carrying
F3 20-4.32 2.8 22-3.2 711 20-3 b)
2.9 20-3.2
50
ERAAS / IIW Doc. XIII-1588-95 Transverse Fillet Load-Carrying F1 (30-4.32) Toe
"Cruciform" or F2 (25-4.32) Throat/Root
EC 9 Member Joint T ranseverse Fillet T oe Cat. 3.6/3.7 2.3 (25-3.2) for L=30 mm, T =15 mm
away from edge
or Cat. 2.9 2.4 minus 1 cat. (22-3.2) at corner
T hroat/Root Cat. 3.8 (18-3.2)
IIW Recom Det. No. 412 (25-3.0) Toe or Det. 414 (16-3.0) Throat
at 2x1E06
300
at 2E6 cycles thoat crack toe
% %
throat crack toe
100 ERAAS 100 30-4,32
small specs 64 IIW 83 25-4.32
100 72 EC 9 73
ex per. data
25-3,2
50 components (beams)
22-3,2
18-3,2
cruciform 25-3,0
a)
16-3,0
10
5
1E+04 1E+05 1E+06 1E+07 1E+08 1E+09
ms.dos_6/(eigene)
ERAAS / IIW Doc. XIII-1588-95 Transverse Fillet Load-Carrying F3 (20-4.32) "Coverplate"

EC 9 Transverse Fillet, Toe, Cat. 2.8 (22-3.2) for L>200 mm, T=15 mm
away from edge
Cat. 2.9 2.8 minus 1 cat. (20-3,2) if on edge
IIW recom Det. 711 (20-3,0) Transverse Fillet Toe, ReinforcementD t< 0,8t

300
range of data components

100 (beams)
50
20-4,32
coverplate 22-3,2
b)
20-3,2
10
20-3,0
5
1E+04 1E+05 1E+06 1E+07 1E+08 1E+09
ms.dos_6/(eigene)
51
Table 17 Comparison of Design Curve Fatigue Strength Values for Welded Transverse or
Longitudinal Attachments on Load-Carrying Structural Components and Experimental
Data (at 2*106 cycles)
Attachments
Weld toe transverse to stress (longitudinal or transverse butt/fillet weld)
Web stiffener
E1 35-3.37 2.1 31-3.2 511 28-3 a)
2.9 28-3.2 512 25-3
Web attachment
E2 23-3.37 2.6 20-3.2 512 36-3 b)
2.17 31-3.5 513 28-3
Longitudinal attachment at flange edge, with/without transition radius
E3 35-3.37 2.11 25-3.2 526 36-3
2.12 28-3.2 28-3
2.13 31-3.2 22-3 c)
E4 18-3.37 2.10 18-3.2 525 18-3
E4* (16-3.37) 16-3
14-3
Vertical longitudinal attachment on flange, with/without transition radius
E5 35-3.37 2.13 31-3.2 approx. 522 32-3
(523) 25/20-3
(524) 18/16-3 d)
E6* 18-3.37 2.8 22-3.2 No corresp. 521 (28/25/20) / 18-3
category
E7 18-3.37 ? No corresp. 521 (28/25/20) / 18-3
category
Vertical transverse attachment on flange
E8 23-3.37 2.2 28-3.2 No corresp. 511 36/28/25-3 approx. e)
category
52
ERAAS / IIW Doc. XIII-1588-95 Attachment With Transverse Fillet, E1 (35-3.37) Web Stiffener,
Extruded/Built-Up Beam
EC 9 Welded Attachment, Transverse Weld Toe Cat. 2.1 (31-3.2) or actually Cat. 2.9 (28-3.2)
IIW Recom 96 Det. No. 511 or 512 (28/25-3) Transverse Fillet, As Welded

300
range of test data
100
50
35-3,37
28-3,2
10
a) 28-3,0
25-3,0
5
1E+04 1E+05 1E+06 1E+07 1E+08 1E+09
ms.dos_6/(eigene)
ERAAS / IIW Doc. XIII-1588-95 Web Attachment With Transverse Fillet, E2 (23-3.37)
EC 9 Welded Attachment, Transverse Weld Toe Cat. 2.6 (20-3.2) for L=100 mm, T=15 mm
or Cat. 2.17 (31-3,5)
IIW Recom 96 Transverse Fillet, As Welded Det. No. 512 (28/25-3) or No. 513 (28-3)

300
100
50 data components
(beams) 23-3,37
31-3,5
20-3,2
10
b) 28-3,0
25-3,0
5
1E+04 1E+05 1E+06 1E+07 1E+08 1E+09
ms.dos_6/(eigene)
53
ERAAS / IIW Doc. XIII-1588-95 Flange Edge Attachment E3 (35-3.37) or E4 (18-3,37)
EC 9 Flange Edge Welded Attachment Cat. 2.11/2.12/2.13 (31/28/25/-3.2) or Cat. 2.10 (18-3,2)
IIW Recom 96 Transverse Fillet, As Welded Det. No. 526 (36/28/22-3) or No. 525 (18/16/14-3)

at 2x1E06
300
35-3,37
data components (beams) 18-3,37
100 w ith radius
31-3,2
28-3,2
50 25-3,2
18-3,2
data components (beams) 36-3,0

w ithout radius 28-3,0
22-3,0
10
c) 18-3,0
16-3,2
14-3,2
5
1E+04 1E+05 1E+06 1E+07 1E+08 1E+09
ms.dos_6/(eigene)
ERAAS / IIW Doc. XIII-1588-95

Vertical-Long. Flange Attachment E5 (35-3.37) or E6 = E7 (18-3,37)
trans. radius extruded or built-up beam
EC 9 Flange Attachment Cat. 2.13 (31-3.2) and Cat. 2.8 (22-3,2)
IIW Recom 96 Det. No. 522 (32-3) or No. 521 (28/25/20/18-3)
at 2x1E06
300
35-3,37
18-3,37
100
data on components (beams)
w ithout any radius 31-3,2
constant amplitude 22-3,2
50
v ariable amplitude
32-3,0
18-3,0
10
d)
5
1E+04 1E+05 1E+06 1E+07 1E+08 1E+09
ms.dos_6/(eigene)
Ò runout
54
ERAAS / IIW Doc. XIII-1588-95
Vertical-Transverse Flange Attachment E8 (23-3.37)
EC 9 Flange Attachment Cat. 2.2 (28-3.2)
IIW Recom 96 Det. No. 511 (36/28/25-3)

at 2x1E06
300
data
small
specimens
100 and
components
(beams) 23-3,37
50
28-3,2
36-3,0
28-3,0
25-3,0
10
e)
5
1E+04 1E+05 1E+06 1E+07 1E+08 1E+09
ms.dos_6/(eigene)
A general comment: in the initial classification according to the European recommendations

ERAAS allowed for somewhat higher values, i.e. higher utilization of the detail capacity. It is
noted that attention had been drawn from the very beginning to the fact that these values are
attainable only under specific manufacturing and quality control procedures stated in the
document. Extrapolations to other conditions of manufacturing and service shall be handled with
care.
To this purpose the current documents (European codes for design and manufacturing/quality for
aluminum structures) give more information.
In the diagrams above a cut-off limit at 5x106, ND, cycles for the IIW assumption of the design
S-N line, contrary to the definition of 108 cycles, NL, for the Eurocode. There is some indication
that the first limit could be true if considering the limited information of runouts in spectrum
loaded and random tests performed in this life region. The Eurocode is more conservative in this
high cycle fatigue region and more information will be necessary under different loading
conditions and patterns to confirm results.
55
8. FRICTION STIR WELDING
This section of the report provides introductory material on Friction Stir Welding, FSW. It
describes the physical processes occurring in the joined members and provides an understanding
of the differences between FSW and traditional fusion welding.
Another important source of information on the subject of FSW is the currently evolving
document ISO TC/ SC N “Welding – Friction Stir Welding of Aluminum – General
Requirements”. The document handles issues of Specification and qualification of welding
procedures, Welding operator qualification, Fabrication, Inspection and testing. It will serve as
an intermediate step toward the integration of FSW in design and execution codes, the Eurocode
or respective specifications, although considerable work has to be performed on data
accumulation and evaluation especially in the area of fatigue behavior.
Friction Stir Welding of aluminum is used in limited applications by various European

shipbuilders for high speed aluminum craft and the discussion presented herein includes the
mechanics of the FSW process as well as some of the reasons why the finished product has
higher static and fatigue design properties than their fusion welded counterparts. It also presents
a summary of the continuing studies into the use and application of friction stir welding.
8.1 General Information on Friction Stir Welding
8.1.1 The Process
Friction Stir Welding is a solid state joining process invented by The Welding Institute (TWI) in
1991 and is rapidly emerging as a viable alternative to fusion welding for joining a variety of
structural alloys. FSW can best be described as a combination of extrusion and forging of metals
at elevated temperatures. It is considered a solid state process, and it does not normally require
any edge preparation of the joint, shielding gases or consumable filler metals.
The process is suitable for welding butt joints, corner sections, T-sections and different lap-joint
configurations, and offers new possibilities for fabrication of large aluminum sections.
Among the benefits of FSW are the ability to weld difficult-to-weld aluminum alloys such as
7xxx series, better retention of baseline material properties and improved dimensional stability of
the welded structure. Since it is essentially solid-state, i.e. without melting in the Heat Affected
Zone HAZ, high quality weld can generally be fabricated with fewer weld defects, low residual
stresses, absence of solidification cracking, porosity and oxidation.
The process is attractive for several other reasons. First, the friction heating is generated locally,
so there is no widespread softening of the assembly. The weld is formed across the entire cross-
sectional area of the interface in a single shot process. The technique is capable of joining
dissimilar materials. Finally, the process is completed in a few seconds with very high
reproducibility – an essential requirement for a mass production industry.
57
8.1.2 The Principle of Operation
In Friction Stir Welding, Figure 33 & Figure 34, the plates to be joined are clamped on a
backing plate to resist the vertical, longitudinal and lateral forces, trying to lift and push them
apart. A cylindrical shoulder tool with a specially designed and profiled probe, Figure 35, made
from a hard, wear resistant material relative to the material being welded, is rotated at a high
speed and slowly plunged into the abutting edges of the parts to be joined.
Figure 33 Schematic Illustration of Friction Stir Welding
58
FSW - Principles of the Process
1.Plates 2.Tool
The plates have to The FSW-Tool is
be rigidly clamped placed over the starting
to a backing plate, point of the joint
which is not shown.
3.Plunging
4.Shoulder
The rotating FSW
Touchdown
Tool is pressed
into the work
piece under high Local material is heated
axial load. and plasticized by the
friction produced by the
shoulder, which is in
contact with the work
piece.
5.Welding
The tool is moved along the

joint line, transporting the
plasticized material around the
pin.
Figure 34 FSW – Principles of the Friction Stir Weld Process
59
Pin
Shoulder
Figure 35 Schematic for the Tip of the FSW Tool
The rotating pin produces the stirring action in the material along the bond line and produces the
required thermo-mechanical deformation. During welding, the probe first makes contact as it is
plunged into the joint region. This initial plunging friction heats a cylindrical column of metal
underneath the probe: the material softens without reaching the melting point and allows
traversing of the tool along the welding line. The depth of penetration is controlled by the length
of the probe below the shoulder of the tool. The contacting shoulder applies additional frictional
heat to the weld region and prevents highly plasticized material from being expelled during the
welding operation. Once the shoulder makes contact the adjacent thermally softened region takes
up a frustum shape corresponding to that of the overall tool geometry, Figure 36.
Figure 36 Samples of FSW Welds
Typically, the surface appearance is a regular series of partial circular ripples, which point
towards the start of the weld. These ripples are essentially cycloidal and are produced by the
final sweep of the trailing circumferential edge of the shoulder. The combined frictional heat
from the probe and the shoulder creates a plasticized, almost hydrostatic condition, around the
immersed probe and the contacting surface of the shouldered region of the work piece top
surface. The soft material is mashed by the leading face of the pin profile and transported to the
trailing face of the pin where it consolidates and cools to produce a high integrity weld. The
60
process can be regarded as a solid phase keyhole welding technique since a hole to accommodate
the probe is generated, then filled during the welding sequence. The consolidated welds are
solid-phase in nature and do not show fusion welding defects. The distortion is significantly less
than that caused by any fusion welding technique. The properties of the weld are closely related
to the tool technology. The tool bit shape and material determines the heating, plastic flow and
forging pattern. Usually, the pin is almost brought into contact with the backing plate, as close as
a few tenths of a millimeter. The stronger the base material, the closer the pin to the backing
plate, in order to ensure complete penetration of the weld through the thickness.
Friction stir welds are not symmetric about the weld centreline due to the tool rotation: the side
of the weld on which the rotational velocity of the tool has the same direction as the welding
velocity is designated the advancing side of the weld; the side of the weld on which the two
velocities have opposite direction is the retreating side of the weld, Figure 37.
Figure 37 Definitions of FSW terminology
The use of an effective fixture, Figure 38, is vital for the success of the process: if the clamping
is not firm enough, a lack of material is experienced in the weld zone.
Figure 38 FSW Setup
61
The forces associated with such softening and mass transfer are significant and will act to push
the softened material out of the joint line. Special care must be taken not to set the tool too deep
which can result in mixing and joining of the backing bar and aluminum plate materials. This
would be unacceptable to the joint and would also damage the probe.
8.1.3 Friction Stir Welding Process Advantages
The process advantages result essentially from the fact that the FSW takes place in the solid
phase, below the melting point of the material to be joined. The benefits include the ability to
join materials that are difficult to fusion weld, for example 2000 and 7000 aluminum alloys.
Friction Stir Welding can use existing available machine tool technology and is also suitable for
automation and adaptable for robotic use. Among its main advantages are:
• Low distortion and shrinkage, even in long welds;

• Excellent mechanical properties as proven by fatigue, tensile and bend tests;
• No fumes, sparks, porosity or spatter;
• Environmentally friendly;
• Can operate in all positions and is energy efficient.
• Non-consumable tool;
• One tool can typically be used for up to 1000m of weld in 6000 series aluminum alloys;
• No need for shielding gas or filler wire;
• No welder certification required;
• Some tolerance to imperfect weld preparations – thin oxide layers can be accepted;
• No grinding, brushing or pickling required in mass production.
• Weight savings compared to fusion welding due to the lack of any required consumables
to complete the weld.
The main limitations of the FSW process are at present:
• Keyhole at the end of each weld;

• Welding speeds are moderately slower than those of some fusion welding processes (up
to 750mm/min for welding 5mm thick 6000 series aluminum alloy on commercially
available machines);
• High axial and transverse loads applied: need for stable backing and clamping elements;
• High degree of stiffness required from the handling system used;
• Limited flexibility when compared to fusion welding processes.
• Too complex for repair welds in the field.
8.1.4 Materials and Thickness
Friction stir welding can be used for joining many types of materials and material combinations,
if tool materials and designs can be found which operate at the forging temperature of the pieces.
It can weld all aluminum alloys, including those that cannot normally be joined by conventional
fusion techniques. Up to the present day, TWI has concentrated most of its efforts on optimizing
the process for the joining of aluminum and its alloys. A major Group Sponsored Project
62
undertaken for TWI‘s Industrial Members demonstrated that the following aluminum alloys
could be successfully welded to yield reproducible, high integrity welds within defined
parametric tolerances:
• 2000 series aluminum (Al-Cu)

• 5000 series aluminum (Al-Mg)
• 6000 series aluminum (Al-Mg-Si)
• 7000 series aluminum (Al-Zn)
• 8000 series aluminum (Al-Li)
This work primarily investigated welding of wrought and extruded alloys. However, subsequent
studies have shown that cast to cast, and cast to extruded (wrought) combinations in similar and
dissimilar aluminum alloys are equally possible.
The stirring effect of the tool is clearly visible in transverse macrosections if different types of
materials have been welded such as wrought aluminum sheets to cast aluminium, Figure 39, or
extrusions to wrought sheets. The onion ring like structure of the nugget is typical of high quality
stir welds in which no porosity or internal voids are detectable.
Figure 39 Transverse Section of 6mm Wrought Aluminum Welded to Cast Aluminum
Continuing development of the FSW tool, its design and materials have allowed preliminary
welds to be successfully produced in:
• Copper and its alloys

• Lead
• Titanium and its alloys
• Magnesium alloys
63
• Zinc
• Aluminum alloys of the 1000 (commercially pure), 3000 (Al-Mn) and 4000 (Al-Si) series
• Plastics
• Mild steel
Preliminary trials have also yielded encouraging results when FSW was used to join aluminum
based metal matrix composites (MMCs), and when the process was applied to dissimilar
materials such as cast magnesium alloy to extruded aluminum alloy.
Single pass butt joints with aluminum alloys have been made in thicknesses ranging from 1.2 to
50 mm without the need for edge preparation. Parameters for butt welding of most aluminum
alloys have been optimized in a thickness range from 1.6 to 10 mm. Special lap joining tools
have been developed for aluminum with thicknesses of 1.2 - 6.4 mm. Thicknesses of up to 100
mm can be welded using two passes, one from each side, with 6082 aluminum alloy, Figure 40.
Figure 40 Double Sided Friction Stir Weld in 75 mm Thick Aluminum Extrusion
8.1.5 Weld Properties and Characteristics
Since traditional heating methods are not employed, the properties of the metal in the joined area
are higher than those from any other known welding process and distortion is virtually
eliminated. The repeatable quality of the solid-phase welds can improve existing products and
lead to a number of new product designs previously not possible. The crushing, stirring and
forging action of the FSW tool produces a weld with a finer microstructure than the parent
material.
64
The first attempt at classifying microstructures was made by P L Threadgrill (Bulletin, March
1997). This work was based solely on information available from aluminum alloys. However, it
has become evident from work on other materials that the behaviour of aluminum alloys is not
typical of most metallic materials, and therefore the scheme cannot be broadened to encompass
all materials. It has been proposed to use the following scheme:
When a cross-section is taken through a friction stir weld, a unique structure is seen that is
comprised of four characteristic regions, Figure 41. The region far from the weld center is
Parent material (base material), which is unaffected by heat or mechanical deformation. The
grains are elongated as a consequence of the earlier rolling operation. In the Heat affected zone
(HAZ), which is closer to the weld center, the material has experienced a thermal cycle, which
has modified the microstructure and the mechanical properties. However, there is no plastic
deformation occurring in this area. This region is similar to the heat-affected zone in a fusion
weld but the peak temperatures are lower. In this zone optical microscopy shows no apparent
difference from the parent material but in age hardened alloys and mechanically hardened alloys,
the hardness is lower in this area: this shows that heat from the welding process either causes
over aging, or lowering of dislocation density, and probably both in fully aged alloys.
Figure 41 Illustration of the Regions Associated with a Friction Stir Weld
In the Thermo-mechanically affected zone (TMAZ) the material has been plastically deformed by
the FSW tool, and the heat from the process will also exert some influence on the material. In the
case of aluminum, it is possible to get significant plastic strain without recrystallization in this
region, and there is generally a distinct boundary between the recrystallized zone and the
deformed zones of the TMAZ. In the earlier classification, these two sub-zones were treated as
distinct regions of microstructure. However, subsequent work on other materials has shown that
aluminum behaves in a different manner to most other materials, in that it can be extensively
65
deformed at high temperature without recrystallization. In other materials, the distinct
recrystallized region (the nugget) is absent, and the whole TMAZ appears to be recrystallized.
The center of the weld experiences plastic flow and recrystallization: this zone is known as the
nugget, Figure 42. The nugget has an asymmetric shape caused by material being preferentially
sheared from one side of the tool and drawn into the centre. The diameter is usually slightly
greater than the diameter of the pin. The weld nugget is the region where full dynamic
recrystallization occurs and is comprised of a fine equi-axed grain structure. Grain size depends
on the alloy and the welding procedure. Typically, it is less than about ten microns (10 μm). For
example, fine equiaxed grains of 2-4 μm in diameter are reported for 7075-T6 alloy while grains
of 10 μm are reported in the weld zone of 6061-T6 Al, in contrast with an average grain size of
100 μm in the base material. In addition, the dislocation density can be significantly reduced.
Electron diffraction indicates that the grain boundaries are of the high angle type, which means
the structure is really formed by grains, not by subgrains, characterized by low-angle boundaries.
Typically, the parent metal chemistry is retained, without any segregation of alloying elements.
Figure 42 Nugget at Center of Friction Stir Weld
Each region has different mechanical properties resulting from the local thermal and mechanical
processing cycles.
The weld nugget strength in the as-welded condition can be in excess of that in the heat affected
zone. In the case of annealed materials, tensile tests usually fail in the unaffected material well
away from the weld and heat affected zone. The welding properties of fully hardened (cold
worked or heat treated) alloys can be further improved by controlling the thermal cycle, in
particular by reducing the annealing and over aging effects in the thermo-mechanically affected
66
zone, where the lowest hardness and strength are found after welding. For optimum properties, it
would seem that, for the latter, a heat treatment after welding is the best choice, although it is
recognised that this will not be a practical solution for many applications.
Typical tensile properties of friction stir welded 5000, 6000 and 7000 series alloys are given in
Table 18. The studies have been conducted by TWI [46], Granges technology [36], Finspǻng,
Sweden and Hidro Aluminum [47] in Hǻvik, Norway. They show that for solution treated plus
artificially aged 6082-T6 aluminum by post weld heat treatment a tensile strength similar to that
of the parent material could be achieved, although the ductility was not fully restored. A further
improvement was possible when weld specimens where made from solution treated and naturally
aged 6082 base metal in the T4 condition and then after welding subjected to normal aging.
Natural aging at room temperature led, in the recently developed 7108 aluminum alloy, to a
similar effect which resulted in a tensile strength of 95% of that of the base material.
Table 18 Typical Mechanical Properties of Friction Stir Welded Aluminum Specimens
Material 0.2% Proof strength Tensile strength Elongation Welding factor

Mpa Mpa % UTSFSW /UTSPARENT
5083-0 Parent 148 298 23,5 N/A
5083-0 FSWed 141 298 23 (1.00)
5083-H321 Parent 249 336 16,5 N/A
5083-H321FSWed 153 305 22,5 0.91
6082-T6 Parent 286 301 10,4 N/A
6082-T6 FSWed 160 254 4,85 0.83
6082-T6 FSWed and aged 274 300 6,4 (1.00)
6082-T4 Parent 149 260 22,9 N/A
6082-T4 FSWed 138 244 18,8 (0.93)
6082-T4 FSWed and aged 285 310 9,9 (1.19)
7108-T79 Parent 295 370 14 N/A
7108-T79 FSWed 210 320 12 (0.86)
7108-T79 FSWed naturally aged 245 350 11 (0.95)
Fatigue tests on friction stir welds made from 6 mm thick 5083-0 and 2014-T6 have been
conducted [46]. The fatigue performance of friction stir butt welds in alloy 5083-0 was
comparable to that of the parent material when tested using a stress ratio of R=0.1. Despite the
fact that the fatigue tested friction stir welds were produced by a single pass from one side, the
results have substantially exceeded design recommendations for fusion welded joints [1].
Analysis of the available fatigue data has shown that the performance of friction stir welds is
comparable with that of fusion welds, and in most cases substantially better.
The outstanding fatigue results can only be achieved if the root of butt welds is fully bonded. As
known from other welding processes, it is also essential in FSW to avoid root flaws. If the pin is
too short for the actual material thickness, the work pieces are only forged together without
stirring up the oxide layers. These flaws are difficult to detect by non-destructive testing. In case
of large variations in sheet thickness, it could be necessary to have extendible pins, which can be
adjusted depending on the actual sheet thickness.
67
8.1.6 Welding Parameters
There are a number of variables that need to be controlled when performing friction stir welds:
• Tool plunge depth: the tool probe is kept at a small distance above the backing bar
(typically 0.2 mm). If the distance is greater than 0.2 mm the stirring action will not
proceed down to the backing bar. High pressures will be transmitted to the backing bar
and cause the root area to be pressure bonded. Pressure bonds are weaker than both
stirred material and parent plate. The tool plunge depth may be influenced by the
thickness of material. If the material is thinner than expected, the probe may gouge into
the backing bar. If the material is thicker than expected then pressure bonds will result.
• Machine parameters: the speed of rotation and tool movement along the weld has an
important effect on weld quality. Tools travelling too fast may not allow consolidation of
plasticized material and the tool could rise similar to a hydrofoil. If speed is too slow,
then material may not become plasticized or be heated to sufficiently high temperatures.
• Plate positioning: the positioning of the plates relative to each other and the tool is
significant. First, the tool probe is of a small diameter and must be positioned over the
center of the joint line. If this is offset then the amount of plasticized material on one side
of the joint will be too small. This leads to weaker joints with the possibility of only
pressure bonding the plates together in extreme circumstances. The plates need to be in
contact: small gaps between the plates are closed by the tool, appearing to zip the plates
together. However, if the gap between the plates is greater than 10% of the tool diameter
the joint strength and elongation will be reduced. As the gap between plates increases
there will be insufficient material to fill the gap resulting in void formation.
Care should be taken during the set-up to ensure optimum positioning of the tool and plates.
8.1.7 Joint Geometries
The process has been used for the manufacture of different kind of welds, Figure 43, i.e. butt
welds, overlap welds, T-sections, fillet and corner welds.
Figure 43 Typical Joint Configurations for Friction Stir Welds
68
The joint configurations shown in Figure 43 correspond to:
a. Square butt
b. Combined butt and lap
c. Single lap
d. Multiple lap
e. 3 piece T butt
f. 2 piece T butt
g. Edge butt
h. Corner Fillet
For each of these joint geometries specific tool designs are required which are being further
developed and optimized.
The Friction Stir Welding process can also be used for circumferential, angular, non-linear, and
three-dimensional welds, Figure 44.
Figure 44 Samples of FSW Components
Since gravity has no influence on the solid-phase welding process, it can be used in all positions:
horizontal, vertical, overhead and orbital.
8.1.8 Applications in Shipbuilding and Marine Industries
The shipbuilding and marine industries are two of the first industry sectors to have adopted the
process for commercial applications, Figure 45.
Friction stir welding has been used in the construction of fast ferries and cruise ships. Fabricators
construct components, which are then delivered to shipyards and fitted directly into place. With
component generation being done away from the shipyards and leaving them with final assembly
of sections to the main structure; shipyards can have faster turn-around times.
69
Figure 45 Deck panels made from FSW profiles
The initial commercial application of friction stir welding involved the manufacture of hollow
aluminum panels for deep freezing of fish. The minimal distortion and high reproducibility make
FSW both technically and economically attractive for production of these stiff panels.
To date, the main application of friction stir welding has been to join extruded sections for deck
structures of fast ferries and helo-decks. Another application has been joining extruded sections
in cruise ship fabrication. Pre-fabricated wide aluminum panels for high speed ferry boats are
already commercially available. The panels are made by joining extrusions which can be
produced in standard size extrusion presses: compared to fusion welding, the heat input is very
low and this results in low distortion and reduced thermal stresses. However, there may be a
number of other applications for friction stir welding in the marine sector. Hull plates may one
day be joined by this process. Another application of friction stir welding may be in the repair of
fusion welds: defects such as cracks and porosity may be stir welded to give a worked
microstructure free from the original imperfections.
8.2 Fatigue Behavior of Friction Stir Welds in 6000 Series Aluminum Alloys
This section presents the literature survey carried out in order to summarize the current state of
knowledge on the fatigue behavior of friction stir welds in aluminum alloys 6XXX (Magnesium
and Silicon). Using computer analyses of all data, many Wöhler Diagrams have been obtained
for different series alloys and different testing parameters. The majority of the data available
was from studies on 6082 aluminum alloy. All these collected data were on butt-welded joint
tests. In most cases, testing was conducted with an R value equal to 0.1 and a frequency value
between 10 and 20 Hz. A comparison between friction stir welds and design curves for
conventional welds has been presented in order to highlight the enhancement of fatigue behavior.
Finally, a diagram summarizes the bulk of the available data.
8.2.1 Fatigue Behavior of Friction Stir Welds in Aluminum Alloy 6082
8.2.1.1 First experimental data on aluminum alloy 6082-T6
Initial experimental data deals with fatigue tests performed on friction stir butt welds in
aluminum alloy 6082-T6. The chemical composition of the parent material is shown in Table 19.
70
Table 19 Chemical Composition of AA6082-T6 in Initial FSW Fatigue Tests
Compositions
Element [wt. %]
Silicon (Si) 0.9

Iron (Fe) 0.50
Copper (Cu) 0.10
Manganese (Mn) 0.40
Magnesium (Mg) 1.0
Chromium (Cr) 0.25
Zinc (Zn) 0.20
Titanium (Ti) 0.10
The mechanical properties of the material are listed inTable 20.
Table 20 Mechanical Properties of AA6082-T6
Yield Ultimate Elongation

2 2
[N/mm ] [N/mm ] [%]
140 240 23
The specimen’s geometry is shown in Figure 46. The plate thickness was 6 mm.
Figure 46 Specimen Geometry
71
Table 21 Experimental points from FSW-joints fatigue tests
Material: Aluminum alloy 6082-T6

R = + 0.1
Specimen Stress Range, S Cycles to Remarks
2
Nr. [N/mm ] Failure, N
Fatigue failure
1 100 275043
in parent material
Fatigue failure
2 100 243673
in parent material
Fatigue failure
3 100 263799
in parent material
Fatigue failure
4 140 60007
in parent material
Fatigue failure
5 140 64802
in parent material
Fatigue failure
6 150 63440
in parent material
Fatigue failure
7 150 46555
in parent material
Fatigue failure
8 200 80000
at the weld toe
Fatigue failure
9 150 400000
at the weld toe
Fatigue failure
10 150 500000
at the weld toe
The specimens were tested under axial loading in a servo-hydraulic machine at a frequency of 14
Hz. The R value was equal to +0.1. Unfortunately, no information on welding conditions is
available. The test results are presented inTable 21. The fatigue failures in the first seven
specimens initiated in the parent material as a result of grain imperfections on their surfaces. In
fact, the heat input caused by friction stir welding had improved the fatigue strength properties in
the welding zone. In the last three specimens, flushed to eliminate such imperfections, fatigue
failure initiated, as expected, at the weld toe. In the statistical evaluations of results, these last
three specimens were considered to obtain information about fatigue strength of the friction stir
weld.
The equation of the regression line (1) and other information about the regression analysis are
presented below, Table 22. Broken points were excluded from analysis.
LogN = −5.98 × LogS + 18 .67 (1)
The Wöhler Diagram is presented in Figure 47.
72
Figure 47 Wöhler Diagram for friction stir butt weld in aluminum alloy 6082-T6
73
Table 22 Results of Linear Regression Analysis
Average
Values:
Mean Mean
Log[Stress]: 2.22 Log[Cycles]: 5.40
Variance and Standard Deviations:
SSR: 0.00 Std. Dev. Log S: 0.01
Variance of
LogN: 0.00 Std. Dev. Slope: 0.67
Std. Dev.
Std. Dev. LogN: 0.07 Intercept : 1.49
95% Confidence Intervals for Slope and Intercept
Parameters:
-28.40 < Slope < 16.43 -31.06 < Intercept < 68.39
Estimated
Values:
Estimated Mean Stress at 2E+06 Cycles to Failure
[MPa]: 116.78
Estimated Mean Stress at 1E+05 Cycles to Failure
[MPa]: 192.68
Estimated LogN Stress Range at 30 679181223
MPa: LogN: 9.83 Cycles: 5
Estimated LogN Stress Range at 50
MPa: LogN: 8.50 Cycles: 319751641
Estimated LogN Stress Range at 100
MPa: LogN: 6.70 Cycles: 5057680
Probability of Survival:
97.5% Probability of Survival at 2E+06 Cycles
in [MPa]: 48.43
97.5% Probability of Survival at 1E+05 Cycles
in [MPa]: 79.91
In order to make a comparison with the fatigue strength of the parent material, a regression
analysis was developed on the seven points where the failure occurred in the parent material.
The equation of the regression line (2) and other information about the regression analysis are
presented below, Table 23. Broken points were excluded from analysis.
LogN = −4.00 × LogS + 13 .41 (2)
74
The Wöhler Diagram is presented in Figure 48. The two Wöhler Diagrams are then compared in
Figure 49, which shows the regression line obtained from the tests on the last three specimens is
higher than the other because of the fatigue strength enhancement obtained by flushing. These
can be considered only as indicative results because of the low number of specimens.
Average Values:
Mean Log[Stress]: 2.09 Mean Log[Cycles]: 5.04
Variance of LogN: 0.00 Std. Dev. Slope: 0.25
Std. Dev. LogN: 0.05 Std. Dev. Intercept : 0.53
95% Confidence Intervals for Slope and Intercept Parameters:
-4.98 < Slope < -3.01 11.35 < Intercept < 15.47
Estimated Values:
Estimated Mean Stress at 2E+06 Cycles to Failure [MPa]: 59.94
Estimated LogN Stress Range at 30 MPa: LogN: 7.50 Cycles: 31848916
97.5% Probability of Survival at 2E+06 Cycles in [MPa]: 53.12
75
Figure 48 Wöhler Diagram for AA6082-T6 (failure occurred in parent material)
76
Figure 49 Comparison between the two Wöhler Diagrams
77
8.2.1.2 Data on Friction Stir Welds in Aluminum 6082 in the T4 and T6 Tempers
Other experimental data on AA6082 are presented in Table 24 through Table 28. In this case
AA6082 was friction stir welded in the T4 and T6 tempers. The aim was to determine fatigue
properties of friction stir welded aluminum alloy 6082, T4 and T6 tempers when subjected to a
post weld aging treatment (PWAT) to improve the static properties. To increase the ductility and
toughness of the alloy, small amounts of manganese were added (typically around 0.7wt%). A
solution heat treatment was executed at 530-550°C, followed by quenching to room temperature.
The solid solution then became supersaturated. The T4 temper is referred to as the condition
obtained if the material is allowed to age naturally at room temperature. The T6 condition is
obtained through artificial aging at an elevated temperature of 170-200°C. The welds in the T4
alloy were further post weld heat treated (PWAT), which enhances yield- and tensile properties
to those of the base material in the T6 condition. The post weld heat treatment consisted of
artificially aging at 185°C for 5 hours, which gives re-precipitation of the hardening particles.
Through this process about 90% HAZ strength recovery can be achieved, resulting in a
considerable increase in the strength of the material. A servo-hydraulic testing machine equipped
with an actuator of 250kN load capacity was used to determine the fatigue properties of the
welds. The dimensions of the pieces tested were 260 x 70 x 5.8mm (length x width x thickness).
The stress ratio R of the sinusoidal curve function was set to +0.5. Average stresses in the range
of 105 to 165 MPa were tested. The frequency was adjusted in the range from 9 Hz to 15 Hz.
Material: Aluminum Alloy 6082-T4

R = + 0.5
Specimen Stress Range, S Cycles to Remarks
Nr. [N/mm2] Failure, N
1 70 1400000 Failure
2 70 1500000 Failure
3 80 1000000 Failure
4 80 900000 Failure
5 80 700000 Failure
6 90 500000 Failure
7 90 450000 Failure
8 90 400000 Failure
9 100 400000 Failure
10 100 250000 Failure
11 100 200000 Failure
12 110 120000 Failure
For T4+PWAT the fractures, for more than half of the specimens, were in the weld area.
They occurred near the weld center or halfway between the center and the weld/HAZ border on
the shear side of the weld. For the rest of the specimens the fracture was located in the thermo-
mechanically affected zone, or in some case in the base material, probably as a consequence of
78
PWAT treatment. The equation of the regression line (3) and other information about the
regression analysis are presented below, Table 25.
LogN = −5.18 xLogS + 15.76 (3)
Average Values:
-6.55 < Slope <-3.80 13.09 < Intercept < 18.42
Estimated Values:
79
Figure 50 Wöhler Diagram for AA6082-T4 Friction Stir Welds R=0.5
80
The process parameters concerning the friction stir welding processes are listed in Table 26.
Table 26 Friction Stir Welding Process Parameters
FSW Parameters Value Description
Travel speed 350 mm/min Horizontal speed

Rotating speed 1000 rpm Rotating speed of pin tool
Shoulder Diameter 20 mm Shoulder diameter of the tool

R = + 0.5
Specimens Stress Range, S Cycles to Remarks
2
Nr. [N/mm ] Failure, N
1 110 100000 Failure
2 110 180000 Failure
3 110 270000 Failure
4 110 300000 Failure
5 110 350000 Failure
6 100 120000 Failure
7 100 180000 Failure
8 100 200000 Failure
9 100 480000 Failure
10 100 600000 Failure
11 92 800000 Failure
12 90 250000 Failure
13 90 500000 Failure
14 90 700000 Failure
15 90 800000 Failure
16 90 900000 Failure
17 80 480000 Failure
18 80 1400000 Failure
19 80 1500000 Failure
20 80 1800000 Failure
21 80 1850000 Failure
22 80 2100000 Failure
23 70 2000000 Failure
24 68 700000 Failure
25 68 2500000 Failure
26 68 3000000 Failure
27 68 > 4500000 Run Out
28 50 > 3500000 Run Out
81
For the T6 material all samples went to fracture at the side of the weld that contains the rougher
welding edge, resulting from the rotating action of the tool. This is the shear side of the weld,
where the relative difference in velocity between tool and work piece is the largest and thereby
also the welding induced residual stresses. The fracture is in the border area weld/HAZ, which is
the softest area in the material. Fracture has in some cases (high stress) been initiated in the weld,
slightly on the inside of the rough edge. The cracks initiated at the top or root edges of the
specimens, alternatively grew out to the edge and then went into the material again. The equation
of the regression line (4) and other information about the regression analysis are presented
below, Table 28.
LogN = −5.06 xLogS + 15.65 (4)
Average Values:
-6.55 < Slope <-3.80 13.09 < Intercept < 18.42
Estimated Values:
The results show that the fatigue strength of T4 + PWAT is lower than for T6. This was not
expected since the T4 + PWAT welded material is statically stronger.
82
Figure 51 Wöhler Diagram for AA6082-T6 Friction Stir Welds (R=0.5)
83
8.2.1.3 Additional Wöhler Diagram for Aluminum 6082-T4
Another Wöhler Diagram was found in the literature for AA6082-T4. The Diagram derives from
a testing program to determine the fatigue properties of transverse butt welding of extruded
plates in aluminum alloy 6082-T4. The plate thickness was 5 mm. The results of mechanical
strength tests on these specimens are listed inTable 29.
Table 29 Mechanical Properties of AA6082

8.2.1.3.1
Yield Ultimate Elongation
2
[N/mm ] [N/mm2] [%]
153 258 26
The geometry and dimensions of the specimens used in S-N tests are shown in Figure 52.
In this case there is also information on process parameters: the Friction Stir Welding was
performed by means of a 7.5 kW Köpings milling machine. The rotating tool consists of a 15
mm diameter cylindrical part made of high strength die steel, H13. The high strength steel pin
had a diameter of 6 mm. Other information on FSW process is listed in Table 30. The specimens
were tested under axial loading in a servo-hydraulic fatigue testing machine equipped with an
actuator of 10kN load capacity. Testing was performed at a room temperature of approximately
20°C. Test frequency was 10 Hz.
The tests were run at a load ratio of R = +0.5. The applied load ranges were selected to produce
fatigue lives in the range of 105 to 106 cycles. Failure was defined to have taken place when the
specimen had separated into two parts. The Wöhler Diagram is presented in Figure 53.
The regression line is (5):
LogN = −5.33xLogS + 15.62 (5)
84
Other information about the regression analysis is presented in Table 31.
Table 30 Friction Stir Welding Process parameters
FSW Parameters Value Description
Travel speed 500 mm/min Millimeter-per-minute horizontal speed

Rotating speed 1150 rev./min Rotating speed of pin tool
Table 31 Details of mean life S-N curve obtained from regression analysis of test results
S-N curve: Standard Fatigue strength

N(∆S)m = C deviation of at 2*106 cycles
logN Stress Percent of
m C range base material
∆S2mill fatigue strength
5.35 4.164*1015 0.05 56 50
85
Figure 53 Wöhler Diagram for friction stir welds in aluminum alloy 6082-T4
86
8.2.2 Fatigue Behavior of Friction Stir Welds on Other 6XXX Aluminum Alloys
8.2.2.1 Experimental Data on Aluminum Alloy A6N01-T5
Experimental data from fatigue tests on an aluminum deck fabricated by friction stir welding
[61] were carried out for the deck and “beam-type” specimens provided by cutting the deck in
the transverse direction. The aluminum is A6N01S-T5. The chemical composition of the
aluminum alloy is in Table 32 together with the values specified in the Japanese Industrial
Standard (JIS).
Table 32 Chemical Composition of A6N01S-T5
Si Fe Cu Mn Mg Cr Zn Ti
Measured
0.51-0.53 0.16 0.08-0.09 0.14-0.15 0.69 0.01 0.01 0.02-0.03
values (%)
JIS-values (%) 0.40-0.90 ≤ 0.35 ≤ 0.35 ≤ 0.50 0.40-0.80 ≤ 0.30 ≤ 0.25 ≤ 0.10
The chemical composition of this aluminum alloy is similar to the one specified for aluminum
alloy 6008, Table 33.
Table 33 Tab. 15 Chemical Composition of Aluminum Alloy 6008
Si Fe Cu Mn Mg Cr Zn Ti
0.50-0.90 ≤ 0.35 ≤ 0.30 ≤ 0.30 0.40-0.70 ≤ 0.30 ≤ 0.20 ≤ 0.10
The beam-type specimens, provided by cutting the deck in the transverse direction, have the
friction stir weld in the middle of the top and bottom flanges. The load of R = +0.1 was applied
on the 10 cm width at the span centre, Figure 54. The fatigue crack was initiated on the lower
surface of the bottom flange at the FSW. It propagated on the cross section of the specimen, and
the specimen broke into two pieces due to brittle fracture.
The equation of the regression line for the beam-type specimens (6) shows a slope approximately
equal to 3.
LogN = −3.23 xLogS + 12.3 (6)
The standard deviation ξN of logN about the arbitrary logS is 0.106.
87
Figure 54 Beam-type Specimen
In this case, the friction stir weld is loaded longitudinally. The results of this fatigue test are
much higher than the S-N curve of the longitudinal butt welds specified in “EC 9- Proposal for
NAD and Corrections” (see Table Detail Category Δσ – m1:60-4.3) as shown in Figure 55.
88
Figure 55 Comparison between Wöhler Diagram for friction stir welds in A6N01-T5
(Japanese Industrial Standard) and design curve for longitudinal butt welds
89
8.2.3 Experimental Data from Fatigue Tests on Transverse FSW Butt Welds on
Aluminum Alloy 6013-T6
Experimental data for the fatigue strength of friction stir welds on aluminum alloy 6013-T6 are
presented below [63]. The objective was to characterize the mechanical properties of butt welds
produced in three different aerospace aluminum alloys using Friction Stir Welding. The FSW
was performed at SAPA Finspång. The mechanical properties evaluated were tensile strength at
room temperature, bend performance and fatigue strength. The properties were tested transverse
to the weld. Static strength and bending properties were found superior to what is usually
achieved with conventional welding methods. The fatigue strength was tested using test
specimens in the as-welded condition as well as after surface milling of the top weld. Strips of
thin sheet material were FSW welded at SAPA Finspång. The strips were welded in a fixture
using square butt joints oriented in the longitudinal direction of the sheet. The welded blanks had
the dimensions 140 x 700 mm and were used for fabrication of test specimens. The welding tool
is referred to as “Standard Tool” in “Patent no. US 5813592”. The pin diameter was 0.4 mm.
Table 34 outlines the welding parameters used.
Table 34 Welding Data
Alloy and Type of Rotational

condition alloy Material Speed Traveling Speed
At welding thickness (rpm) (mm/min)
(mm)
6013 – T4 AlMgSi 1.6 2000 208
Welded blanks of each alloy were inspected using visual inspection and radiography. The visual
inspection revealed that the degree of burrs formed on the advancing side of the weld was very
high for the 6013 alloy. Radiographs on welds of 6013 were without remarks. The blanks of
6013 T4 were artificially aged to the T6 condition. Plain un-notched specimens were used in the
fatigue testing of the FSW welds, as shown in Figure 56. The specimens were tested both in the
as-welded condition and after flush milling of both the weld topside and the root side. By the
milling operation 0.10 to 0.15 mm material was removed from the weld and the sheet adjacent to
the weld. The fatigue testing was carried out with constant amplitude at the stress ratio R = +0.1.
The loading frequency was 25 Hz.
Two different regression analyses have been carried out for the specimens in the as-welded
condition and for the specimens after flush milling.
90
The test results for the specimens in the as welded conditions are presented in Table 35.
Table 35 Experimental points from fatigue tests – FSW as welded

R = + 0.1
2
Nr. [N/mm ] Failure
1 225 50000 Failure
2 225 60000 Failure
3 205 95000 Failure
4 205 100000 Failure
5 190 160000 Failure
6 190 250000 Failure
7 170 210000 Failure
8 155 600000 Failure
9 155 2000000 Failure
The equation of the regression line (7) and the Wöhler Diagram, Figure 57, are presented below.
LogN = −7.61xLogS + 22.60 (7)
91
Figure 57 Wöhler Diagram for FSW butt welds in the as welded conditions on aluminum
alloy 6013–T6
92
Other information about the regression analysis is listed in Table 36.
Average Values:
-11.48 < Slope <-3.74 13.78 < Intercept < 31.41
Estimated Values:
The test results for the specimens in the as welded conditions are presented in Table 37.
Table 37 Experimental points from fatigue tests – FSW as milled

R = + 0.1
2
Nr. [N/mm ] Failure
1 240 40000 Failure
2 240 50000 Failure
3 225 110000 Failure
4 225 130000 Failure
5 205 300000 Failure
6 205 220000 Failure
7 190 350000 Failure
8 190 500000 Failure
9 190 2000000 Failure
10 190 > 6000000 Run Out
The equation of the regression line (8) and the Wöhler Diagram, Figure 58, are presented below.
The unbroken specimen was excluded from analysis.
93
LogN = −11.35 xLogS + 31.70 (8)
Other information about the regression analysis is presented in Table 38.
Average Values:
-17.73 < Slope <-4.97 16.86 < Intercept < 46.53
Estimated Values:
These results highlight that, although the fatigue strength of friction stir welds is always very
high, for optimal fatigue properties milling of the FSW top surface is necessary. Actually, the
fatigue tests indicated that milling of the topside of the weld increased the fatigue strength to
approach that of parent material.
94
Figure 58 Wöhler Diagram for FSW butt welds in the as milled conditions on aluminum
alloy 6013 – T6
95
8.3 Conclusions on Friction Stir Welding
All the experimental points and the regression lines (mean values) found in the literature for
transverse friction stir welds without surface post-treatment on aluminum alloys 6XXX have
been compared with the fatigue strength of butt welds obtained by means of other techniques,
Figure 59. The data in the “Aluminum Data Bank” (TUM) came from fatigue tests on small
specimens or on extruded beams in aluminum alloys containing transverse butt welds.
A linear regression analysis has been developed on these experimental points. The equation of
the regression line (9) is shown below.
LogN = −3.77 xLogS + 12.73 (9)
As expected, the experimental points found in literature for transverse butt welds obtained by
means of Friction Stir Welding are in almost all cases above this regression line.
The same experimental points found in literature are shown in Figure 60 in comparison with
Design Standards in order to verify the applicability of existing design rules to this relatively
new fabrication process.
The considered design standards are the ones stated in “Proposal for NAD and Corrections
(November 1998)”. Different detail categories were considered in order to make a comparison
(see “Proposal for NAD and Corrections (November 1998)” Table 5.14 “Detail Categories for
Welded Joints between Members”). The design curve for the detail category Δσ-m1 55-7.0
means MIG transverse butt welds with overfill dressed flush from both sides is generally higher
than the experimental points found in literature, but can not be accepted as a design curve for
friction stir welds without surface post-treatment.
The best choice seems to be the detail category 40 – 4.3 the single sided butt weld unbacked with
full penetration.
If a general conclusion must be reached from this literature survey on friction stir welds, it has to
be stated that friction stir welding is an excellent way to join aluminum alloys. After a post-
treatment of the FSW top surface, such as milling, the fatigue strength is further increased and
approaches that of parent material.
96
Figure 59 Transverse friction stir welds no surface post-treatment compared to fatigue
strength of transverse butt welds (Aluminum Data Bank – TUM)
97
Figure 60 Transverse friction stir welds without surface post-treatment on aluminum alloys
6XXX compared with Design Standards
98
9. CONCLUSIONS AND RECOMMENDATIONS
Their is a significant volume of additional work that has been done and information that is
available regarding fatigue of aluminum joints, friction stir welding and a comparison of fatigue
performances using FSW and traditional fusion welding procedures. In addition, a lot of work
has also been developed on damage tolerance and crack propagation in aluminum details. The
Eurocode 9: Design of Aluminum Structures – Part 2: Structures Susceptible to Fatigue is an
evolving design standard with procedures for fatigue evaluation of structural joints fabricated
from aluminum. It also includes design curves and standards for damage tolerance and crack
propagation providing a designer with the tools that are required to assess the likelihood of
exceeding a critical crack length by a certain time in the design life of a cracked joint.
The objectives of this project associated with the survey of aluminum structural details still
require work. The use of aluminum in marine structural applications continues to increase in
both commercial and naval applications, with particular interest in high speed craft. This growth
also brings an increased database of available information and experience although the
competitive nature of the commercial industry may restrict access to and publication of detail
performance. Naval programs, such as Littoral Combat Ship, also suggest the development of
aluminum detail performance in combatant environments with different design procedures and
requirements than typified by commercial design.
99
10. REFERENCES
10.1 European Standards
[1] “European Recommendations for Aluminum Alloy Structures Fatigue Design”, European
Convention for Constructional Steelwork, Doc. No. 68, First Edition, Brussels 1992.
[2] Kosteas, D., Uhry, A.: „Fatigue Behaviour of Welded Aluminum Structures - European
Full-Scale Tests“, Part 1 (Kosteas): Analysis Results and Proposals, Part 2 (Uhry):
Statistical Analysis, Part 3 (Kosteas): Data Base. Commission of the European
Communities - COST 506 Program - European Aluminum Association. Munich/Voreppe,
March/April 1989.
[3] Jaccard, R.: ”Alusuisse Data Package onb Welded Aluminum Beams”, Internal report,
May 1988
[4] “AMAG Data on Welded Aluminum Beams”, Laboratory Test Report, Wien, 02.05.1984
[5] Poalas, K., Kosteas, D.: „Voraussage des Ermüdungsverhaltens geschweißter
Aluminumbauteile“. LKI-Forschungsbericht Nr. 120 zum Forschungsvorhaben AIF 5723,
Lehrstuhl für Stahlbau der Technischen Universität München, 1986.
[6] Ondra, R., Kosteas, D.: “Untersuchungen zum Schwingfestigkeitsverhalten geschweißter
Verbindungen in Aluminum-Großbauteilen“. LKI-Forschungsbericht Nr. 897-On zum
Forschungsvorhaben AIF 7331, Lehrstuhl für Stahlbau der Technischen Univbersität
München, 1991.
[7] Jaccard, R., Kosteas, D., Ondra, R.: „Data Re-Evaluation and Finalization of ERAAS
Fatigue“ , Part 1: Main Volume/Commentary, Part 2: Documents M-1 to M49, First
Fatigue Workshop Munich February 1990, Part 3: Documents Z-1 to Z-135, Second
Fatigue Workshop Zürich March 1990, Part 4: Preliminary Documents for Workshops Part
5: Data Sheets as of 12.07.1990, Part 6: Documents A-1 to A-35, Reclassification of
Weldments,Part 7: Comparison of Structural Details Fatigue Strength Values, Internal
Report of a joint TUM/A-L Study 1990/1991.
[8] R. Fisher, C. Menzemer: „Fatigue test results on welded aluminum beams“, internal data
exchange report, 1992.
[9] British Standard: „BS 8118- Structural Use of Aluminum“, Part 1: Code of Practice for
Design, 1992,
[10] Aluminum Association: “Specifications for Aluminum Structures“, Washington, 1986
[11] “Association American Railcars”, March 1988
[12] “Ontario Highway Bridge Code”, Ontario, 1979
[13] Deutscher Verein für Schweißtechnik: “Richtlinien DVS 1608“, Düsseldorf, 1969
[14] DIN 15018, „Krane“, Ausgabe 1984
[15] Richtlinien der Deutschen Bundesbahn: “DS 804 - Vorschriften für Eisenbahnbrücken und
sonstige Ingenieurbauten“, Ausgabe Januar 1983
[16] Deutscher Ausschuß für Stahlbauten: “DASt-Richtlinie 011 - Feinkorn Baustähle“, Februar
1979
[17] ENV 1993 -1 -1: “Eurocode 3 - Bemessung und Konstruktion von Stahlbauten“, Brüssel,
April 1992
[18] Voutaz B., Smith I., Hirt M.: Comportement à la fatigue de poutres en aluminum munies
de goussets doudés“, ICOM-Construction Métallique, Ecole Polytechnique Fédéral de
Lausanne, ICOM 285, April 1993
101
[19] van Straalen IJ. J., Soetens F., Dijkstra O. D.: ”EUREKA 269 - Fatigue tests on transverse
butt welds“, TNO Building and Research, 13. Oct. 1994
[20] van Straalen IJ. J., Soetens F., Dijkstra O. D.: ”EUREKA 269 - Fatigue tests on
longitudinal nonload carrying fillet welds“, TNO Building and Research, 13. Oct. 1994
[21] van Straalen IJ. J., Soetens F., Dijkstra O. D.: ”EUREKA 269 - Fatigue tests on transverse
load carrying fillet welds“, TNO Building and Research, 13. Oct. 1994
[22] van Straalen IJ. J., Soetens F., Dijkstra O. D.: ”EUREKA 269 - Fatigue tests on
transverse non-load carrying fillet welds“, TNO Building and Research, 13. Oct.
[23] prENV 1999-2 (Sept. 1997) / Eurocode 9: Design of aluminum structures - Part 2:
Structures susceptible to fatigue.
[24] Jaccard, R., Ogle, M. H.: Aluminum Eurocode - Design for Fatigue. In: WRC
Proceedings IIW, p. 205-224.
[25] Jaccard, R., Kosteas, D., Ondra, R.: Background Document to Fatigue Design Curves for
Welded Aluminum Components. IIW Doc. No. XIII-1588-95.
10.2 Friction Stir Welding
[26] Kallee S., 1999 Reibrührschweißen – Verfahren und Einsatzfälle, GKSS/TWI Workshop
Mai
[27] Wayne T., Dawes C., Gittos M., Andrews D., 1998, Friction stir – where we are, and
where we’re going…, TWI Bulletin May/June
[28] Przydatek J., 2000, Friction Stir Welding with Lloyd’s register, International Conference
Lightweight Construction – Latest developments 24 and 25 February, London SW1
[29] Seetharaman V., Jata K.V., Semiatin S. L., Plastic flow and microstructure development
during high temperature deformation of a friction stir welded 7050 aluminum alloy, Air
Force Research Laboratory, Materials and Manufacturing Directorate, AFRL/MLLM
Wright-Patterson Air Force Base, OH 45433-7817.
[30] Kallee S., 2000, Application of friction stir welding in the shipbuilding industry,
International Conference Lightweight Construction – Latest Developments, 24 and 25
February London SW1
[31] Shepherd G. E., The potential for using solid phase welding to repair cracks that may
occur on thin aluminum aircraft wing structure, BAE SYSTEMS Airbus UK, Filton,
Bristol.
[32] Lee J.A., Carter R.W., Ding J., 1999, Friction Stir Welding for Aluminum Metal Matrix
Composites (MMC’S), MSFC Center Director’s Discretionary Fund Final Report, Project
No. 98-09
[33] Haagensen P.J., Midling O.T., Ranes M., 1995, Fatigue performance of Friction Stir Butt
Welds in a 6000 Series Aluminum Alloy, 2nd International Conference on Surface
Treatment 95, Milan, Italy, 7-9 June.
[34] Johnson R., Kallee S., 1999, Stirring stuff from friction welding, Materials World,
December.
102
[35] Hannour F., Davenport A.J., Strangwood M., 2000, Corrosion of Friction Stir Welds in
High Strength Aluminum Alloys, 2nd International Symposium on Friction Stir Welding,
26-28 June, Gothenburg, Sweden.
[36] Backlund J., Norlin A., Andersson A., 15-17 April 1998, Friction stir welding – weld
properties and manufacturing techniques, 7th Inalco Conference, Cambridge, ISBN 1
85573 417 6
[37] Thomas W.M., Nicholas E.D., 1997, Friction stir welding for the transportation
industries, Materials & Design, Vol.18, Nos. 4/6, pp.269-273.
[38] Sanderson A., Punshon C.S., Russell J.D., 2000, Advanced welding processes for fusion
reactor fabrication, Fusion Engineering and Design 49-50 77-87
[39] H. S. Yang, 1998, Microstructural Development in Friction Stir Welding of Aluminum
Alloys Proc. of ICAA-6 (6th Int. Conf. Aluminum Alloys), Toyohashi, Japan, 1483-1488.
[40] O. Frigaad, B. Bjorneklett, O. Grong and O. T. Midling, 1998, Process Modelling
Applied to Friction Stir Welding of Al-Mg-Si Alloys, Proc. of ICAA-6 (6th Int. Conf.
Aluminum Alloys), Toyohashi, Japan, 1477-1482.
[41] H. Larsson, L. Karsson and L.-E. Svensson, 1998, Characteristics of Friction Stir Welds
in AA5083 and AA6082 Aluminum Proc. of ICAA-6 (6th Int. Conf. Aluminum Alloys),
Toyohashi, Japan, 1471-1476.
[42] J. Hagstrom and R. Sandstrom, 1998, Static and Dynamic Properties of Joints in Thin-
Walled Aluminum Extrusions, Welded with Different Methods, Proc. of ICAA-6 (6th Int.
Conf. Aluminum Alloys), Toyohashi, Japan, 1447-1452.
[43] M. W. Mahoney, C. G. Rhodes, J. G. Flintoff, R. A. Spurling and W. H. Bingel, 1998,
Properties of Friction Stir welded 7075 -T651 Aluminum, Metall. Mater. Trans., 29A.
[44] G. Rhodes, M. W. Mahoney, W. H. Bingel, R. A. Spurling and C. C. Bampton, 1997,
Effects of Friction Stir Welding on Microstructure of 7075 Aluminum, Scripta Mater.,
36, 69-75.
[45] T. Hashimoto, N. Nishikawa, S. Tazaki and M. Enomoto, 1998, Mechanical properties of
Joints for Aluminum Alloys with Friction Stir Welding Process, Preprints of INALCO
98 (7th Int. Conf. Joints in Aluminum), Cambridge, UK, 237-247.
[46] Dawes C. J., Thomas W. M., 1995, Friction stir joining of aluminum alloys, TWI Bulletin
Nov/Dec
[47] Midling O.T., Oosterkamp L.D., Bersaas J., 15-17 April 1998, Friction stir welding
aluminum – Process and applications, 7th Inalco Conference, Cambridge, ISBN 1 85573
417 6
[48] European Convention for Constructional Steelwork: European recommendations for
aluminum alloy structure fatigue design, No 68
[49] W. M. Thomas E. D. Nicholas, J. C. Needham, M. G. Murch, P. Templesmith and C. J.
Dawes, 1991, Friction Stir Butt Welding, International Patent Application
PCT／GB92／02203, GB Patent Application 9125978.8, 6 Dec. and US Patent
5,460,317
[50] E. D. Nicholas, 1998, Developments in the Friction Stir Welding of Metals, Proc. of
ICAA-6 (6th Int. Conf. Aluminum Alloys), Toyohashi, Japan, 139-151
[51] C. J. Dawes and W. M. Thomas 1995, Friction Stir Joining of Aluminum Alloys, TWI
Bulletin, 36124-127
[52] C. J. Dawes and W. M. Thomas, 1996, Friction Stir Process Welds Aluminum Alloys,
Welding J., March, 41-45
103
[53] P. Threadgill, 1997, Friction Stir Welds in Aluminum Alloys - Preliminary
microstructural Assessment, TWI Bulletin, 38, March/April
[54] C. J. Dawes,1995, An Introduction to Friction Stir Welding and Its Development,
Welding and Metal Fabrication, January, 14-16
[55] H. R. Shercliff and M. J. Russell, 1998, Analytical Modelling of Friction Stir Welding,
Preprints of INALCO 98 (7th Int. Conf. Joints in Aluminum), Cambridge, UK, 185-195.
[56] O. Frigaad, O. Grong and O. T. Midling, 1998, Modelling of Heat Flow Phenomena in
Friction Stir Welding of Aluminum Alloys Preprints of INALCO 98 (7th Int. Conf. Joints
in Aluminum), Cambridge, UK, 197-207.
[57] L. E. Svensson, J. Karlsson, B. Karlsson, H. Larsson and L. Karlsson, 1998,
Microstructure and Properties of Friction Stir Welded Aluminum Alloys Preprints of
INALCO 98 (7th Int. Conf. Joints in Aluminum), Cambridge, UK,221-230.
[58] M. W. Mahoney, W. Harrigan and J. A. Wert, 1998, Friction Stir Welding SiC
Discontinuously Reinforced Aluminuium, Preprints of INALCO 98 (7th Int. Conf. Joints
in Aluminum), Cambridge, UK, 231-236.
[59] R. Tovo, R. De Scicolo, M. Volpone, 1999, Proprietà meccaniche e micro-strutturali di
giunti “Friction Stir Welded” in lega di alluminio.
[60] M. Ericsson, R. Sandström, J. Hagström, Fatigue of friction stir welded AlMgSi-alloy
6082.
[61] I. Okura, M. Naruo, L. Gergely Vigh, N. Hagisawa, H. Toda, 2001, Fatigue of aluminum
deck fabricated by friction Stir Welding, 8th INALCO 2001 International Conference on
Joints in Aluminum, Munich, Germany, 28-30 March 2001
[62] T.Kawasaki, T.Makino, S. Todori, H. Takai, M. Ezumi, Y. Ina, Application of Friction
Stir Welding to the Manufacturing of Next Generation “A-train” type Rolling Stock,
HITACHI Ltd, Kasado administrative division, Higashi-Toyoi 794, Kudamatsu-city,
Yamaguchi, JAPAN 744-8601
[63] L. Magnusson, L. Källman, Mechanical Properties of Friction Stir Welds in Thin Sheet of
Aluminum 2024, 6013 and 7475
104
PROJECT TECHNICAL COMMITTEE MEMBERS
The following persons were members of the committee that represented the Ship Structure
Committee to the Contractor as resident subject matter experts. As such they performed technical
review of the initial proposals to select the contractor, advised the contractor in cognizant matters
pertaining to the contract of which the agencies were aware, performed technical review of the
work in progress and edited the final report.
Chairman Mr. Chao Lin MARAD
Members
Contracting Officer’s Technical Representative:
Ms. Lynell Fox MARAD
Project Technical Committee Members:
Executive Director:
LT Benjamin A. Gates USCG
Administrative Assistant:
Ms. Jeannette Grant USCG
98
SHIP STRUCTURE COMMITTEE LIAISON MEMBERS
LIAISON MEMBERS
American Iron and Steel Institute
American Society for Testing & Materials
American Society of Naval Engineers Captain Dennis K. Kruse (USN Ret.)
American Welding Society
Bath Iron Works Mr. Steve Tarpy
Canada Ctr for Minerals & Energy Technology
Colorado School of Mines Dr. Stephen Liu
Edison Welding Institute Mr. Rich Green
International Maritime Organization Mr. Igor Ponomarev
Int’l Ship and Offshore Structure Congress Dr. Jack Spencer
INTERTANKO Mr. Dragos Rauta
Massachusetts Institute of Technology
Memorial University of Newfoundland Dr. M. R. Haddara
National Cargo Bureau Captain Jim McNamara
Office of Naval Research Dr. Yapa Rajapaksie
Oil Companies International Maritime Forum Mr. Phillip Murphy
Tanker Structure Cooperative Forum
Technical University of Nova Scotia
United States Coast Guard Academy Commander Kurt Colella
United States Merchant Marine Academy
United States Naval Academy Dr. Ramswar Bhattacharyya
University of British Columbia Dr. S. Calisal
University of California Berkeley Dr. Robert Bea
University of Houston - Composites Eng & Appl.
University of Maryland Dr. Bilal Ayyub
University of Michigan Dr. Michael Bernitsas
University of Waterloo
Virginia Polytechnic and State Institute Dr. Alan Brown
Webb Institute Prof. Roger Compton
Welding Research Council
Worchester Polytechnic Institute
Samsung Heavy Industries, Inc. Dr. Satish Kumar
99
RECENT SHIP STRUCTURE COMMITTEE PUBLICATIONS
Ship Structure Committee Publications on the Web - All reports from SSC 392 and forward are
available to be downloaded from the Ship Structure Committee Web Site at URL:
http://www.shipstructure.org
SSC 391 and below are available on the SSC CD-ROM Library. Visit the National Technical
Information Service (NTIS) Web Site for ordering information at URL:
http://www.ntis.gov/fcpc/cpn7833.htm
SSC Report Number Report Bibliography
SSC 445 Structural Survivability of Modern Liners, S. Schilling, 2005
SSC 444 In-Service Non-Destructive Estimation of the Remaining Fatigue Life of

Welded Joints, Dexter R.J., Swanson K.M., Shield C.K. 2005
SSC 443 Design Guidelines for Doubler Plate Repairs on Ship Structures
Sensharma P.K., Dinovitzer A., Traynham Y. 2005
SSC 442 Labor-Saving Passive Fire Protection Systems For Aluminum And
Composite Construction E. Greene, 2005
SSC 441 Fire Degradation, Failure Prediction And Qualification Methods For
Fiber Composites R. Asaro, M. Dao, 2005
SSC 440 Deterioration of Structural Integrity Due to Chemical Treatment of

Ballast Water S. Tiku, 2005
SSC 439 Comparative Structural Requirements For High Speed Crafts K. Stone,
2005
SSC 438 Structural Optimization for Conversion of Aluminum Car Ferry to

Support Military Vehicle Payload, R.Kramer, 2005
SSC 437 Modeling Longitudinal Damage in Ship Collisions A.J. Brown, JAW
Sajdak 2005
SSC 436 Effect of Fabrication Tolerances on Fatigue Life of Welded Joints A.

Kendrick, B. Ayyub, I. Assakkaf 2005
SSC 435 Predicting Stable Fatigue Crack Propagation in Stiffened Panels R.J.
Dexter, H.N. Mahmoud 2004
SSC 434 Predicting Motion and Structural Loads in Stranded Ships Phase 1
A.J. Brown, M. Simbulan, J. McQuillan, M. Gutierrez 2004
100

In-Service Performance of Aluminum Structural Details

Uploaded by

Copyright:

Available Formats

In-Service Performance of Aluminum Structural Details

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

In-Service Performance of Aluminum Structural Details

Uploaded by

Copyright:

Available Formats

NTIS #

This document has been approved

SHIP STRUCTURE COMMITTEE

RADM Craig E. Bone

CONTRACTING OFFICER TECHNICAL REP. EXECUTIVE DIRECTOR

SHIP STRUCTURE SUB-COMMITTEE

MARITIME ADMINISTRATION MILITARY SEALIFT COMMAND

17. Key Words 18. Distribution Statement

1. Develop a survey of in-service performance of aluminum structural details for use

2. Construct a database of the in-service information and develop a rating system to

1.1 Determination of Details & Classes of Details for Survey

1.2 Fatigue Behavior of Aluminum Structural Details

1.3 Conclusions and Recommendations

2.1 List of Ship Details & Background Information

• Operating environment of vessel.

• SSC-266 “Review of Ship Structural Details”

2.2 Review of Previous Ship Structure Committee Reports

2.2.1 SSC 266 “Review of Ship Structural Details” 1977

2.2.2 SSC 272 “In-Service Performance of Structural Details” 1978

Table 2 Detail Classifications and Number of Observed Details in SSC 294

2.2.4 Other SSC Reports Reviewed

3.1 Damage Tolerant Design and Fracture Analysis

3.2 History and Development of the European Standards

3.3 Friction Stir Welding

4.1.1 Design S-N Curves in the ERAAS

Figure 1 Equi-distant Design Curve Mesh

equi-distant 18.0 N/mm²

Figure 2 Effect of Classification Considering Equi-Distant and Parallel Design Line

Using an equi-distant parallel S/N curve band result in the following:

At 105cycles and a parallel band slope of 3.40

At 105cycles and a parallel band slope of 4.30

• EPF-Lausanne by Hirt et al [18],

Figure 4 Full-Scale Background Data for ERAAS-Fatigue Document

4.1.3 Base Metal 5000/6000 Series

ERAAS Detail Class A3 (simple extrusions) and A4 (Components)

Figure 5 Base Metal 6005A (AlMgSi0) Small Specimen Data

Figure 6 Base Metal 5083 (AlMg4,5Mn) Full Scale Component Data

4.1.3.1 Butt Weld, Transverse

Figure 10 Two-sided Transverse Butt Weld on Built-up Components

4.1.3.2 Butt Weld, Longitudinal

Figure 12 Longitudinal Butt Weld, As Welded

4.1.3.3 Longitudinal Fillet Welds

Figure 14 Intermittent Longitudinal Fillet Weld

4.1.3.4 Fillet Weld, Transverse, Non-Load-Carrying

Figure 16 Transverse Fillet Weld, Non-Load-Carrying (Web Stiffener), ATLLS-Lehigh

4.1.3.5 Web Attachments

ERAAS detail class E2-23 N/mm² (rectangular, circular, hollow shapes)

Figure 17 Web Attachments

4.1.3.6 Attachment at Edge of Flange

Figure 19 Attachment at Flange Edge, no transition radius

4.1.3.7 Vertical Attachment on Flange

Figure 20 Vertical Attachment (Long’l) - Flange of Built-up Beam, no transition radius

Figure 21 Vertical Attachment (Long’l) - Flange of Built-up/Extruded Beam, no transition

4.1.3.8 Cruciform Joint

Figure 23 Cruciform Joint, Built-up Beams

Figure 24 Cover plate, Built-up Beam, TUM Test Results

Figure 25 Cover plate, Built-up Beam, Lehigh Test Results

• Aluminum Association Specification for Aluminum Structures

5.1 Comparison to Aluminum Codes