QUALITY AND RELIABILITY ENGINEERING INTERNATIONAL

Qual. Reliab. Engng. Int. 2009; 25:921–932

Published online 4 March 2009 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/qre.1010

Research Applicability of Published Data

for Fatigue-limited Design
Martin Leary1, ∗, † and Colin Burvill2
1 Department of Aerospace, Mechanical and Manufacturing Engineering, RMIT University, Bundoora,
Victoria 3083, Australia
2 Department of Mechanical Engineering, University of Melbourne, Parkville, Victoria 3010, Australia

The use of published fatigue data provides an expedient basis for fatigue-limited
engineering design by alleviating the necessity of explicit testing. However, published
fatigue data often exhibits incomplete documentation of the associated test condi-
tions. Incomplete documentation introduces uncertainties in fatigue life prediction
that may limit the applicability of the published fatigue data for design applica-
tions. Characterization of the applicability of published fatigue data is critical for
robust fatigue-limited design. However, no quantitative methods have been identified
which respond to this requirement. A novel method has been developed to provide
a systematic characterization of the applicability of published fatigue data based
on internationally recognized standards. This method provides a conceptual mecha-
nism to:

• identify the applicability of published fatigue test data for specific design
scenarios—thereby informing engineers of potential limitations of published
data and allowing prioritization of multiple data sources;
• identify material domains of insufficient applicability—thereby providing a
robust basis for identifying beneficial fatigue test programs;
• compromise between design complexity and the uncertainties inherent in
fatigue life prediction;
• define a framework for the appropriate documentation of published fatigue
data.
A sample of published fatigue data sources associated with a specific fatigue-limited
safety–critical design scenario was assessed by the method presented in this paper. For
the majority of the sampled references, the associated documentation was insufficient
to allow the fatigue test data to be confidently applied to the subsequent design activity.
Copyright © 2009 John Wiley & Sons, Ltd.

KEY WORDS: fatigue design; life prediction; reliability analysis; knowledge management; expert systems

∗ Correspondence to: Martin Leary, Department of Aerospace, Mechanical and Manufacturing Engineering, RMIT University, Bundoora,
Victoria 3083, Australia.
† E-mail: martin.leary@rmit.edu.au

Copyright 2009 John Wiley & Sons, Ltd.

922 M. LEARY AND C. BURVILL

1. INTRODUCTION

F
atigue life prediction is achieved with reference to an estimate of the material response to dynamic
loading. This estimate may be obtained by the following methods, presented in order of increasing
design complexity and reducing uncertainty in the ensuing fatigue life prediction1 :
1. empirical correlation with other material properties;
2. published fatigue test data;
3. explicit testing of representative specimens;
4. explicit testing of actual components.
For example, explicit component testing minimizes uncertainty in fatigue life prediction, but requires
extant components and experimental apparatus to emulate the actual loading. The use of published fatigue
test data for the fatigue-limited design provides an opportunity to compromise between design complexity
and the associated uncertainties of fatigue life prediction. This compromise is suited to industry sectors
with production imperatives that require component design to be achieved within a limited implemen-
tation schedule and design budget2,3 , and is typically applied in preliminary design investigation, for
example4,5 .
The impetus for the work reported in this paper was an attempt to reduce the mass of a fatigue-limited
safety–critical‡ (FLSC) automotive component by the substitution of ferrous metals with light alloys1 . It
was hoped that publicly available fatigue data would provide a robust basis for this design task.
Publicly available fatigue test data provides an opportunity to reduce design complexity by alleviating
the necessity of explicit testing. However, such data often contain insufficient associated documentation
to adequately characterize the associated test conditions6 . If the documentation associated with published
fatigue test data is incomplete, the mismatch between the test attributes and the attributes of the component
being designed is ill-defined, and the reported fatigue test data are likely to be of limited applicability7 .
Characterization of the applicability of fatigue test data is critical to robust fatigue-limited design8 . Despite
a significant review of the available literature, no methods have been identified which formally characterize
the applicability of published fatigue data. Applicability is defined as the level of uncertainty the designer
must accept when applying published fatigue test data to a design task. This paper reports on a novel approach
to systematically characterize the applicability of published fatigue test data for fatigue-limited design. To
assess the applicability of fatigue test data reported in the literature, a series of applicability metrics have
been generated, based on standards developed by the American Society for Testing and Materials (ASTM).
These applicability metrics, in combination with the documentation clarity (DC) associated with the sampled
fatigue data, provide a basis for an applicability rating (AR) that characterizes the applicability of fatigue
data for design purposes. This method provides a robust mechanism to:
• identify the applicability of published fatigue test data for specific design scenarios—thereby informing
engineers of potential limitations of published data and allowing prioritization of multiple data
sources;
• compromise between design complexity and the uncertainties inherent in fatigue life prediction;
• identify material domains of insufficient applicability—thereby providing a robust basis for identifying
beneficial fatigue test programs;
• define a framework for the appropriate documentation of published fatigue data.
A sample of published fatigue data associated with the authors’ light-weight automotive component
redesign project was assessed according to the proposed AR method (Section 3). This assessment formally
identified that, for the majority of the sampled references, the associated documentation was insufficient to
allow the fatigue test data to be confidently applied to subsequent design activity.

‡ I.e. fatigue is the dominant failure mode and component failure can cause injury or death.

Copyright 2009 John Wiley & Sons, Ltd. Qual. Reliab. Engng. Int. 2009; 25:921–932
DOI: 10.1002/qre
APPLICABILITY OF PUBLISHED DATA 923

2. APPLICABILITY METRICS
ASTM Standard E468 defines the ‘standard practice for presentation of constant amplitude fatigue test
results for metallic materials’ and ‘establishes the desirable and minimum information to be communicated
between the originator and the user of data’7 . The practices of E468 have been discretized into representative
applicability metrics (Table I). Modifications and omissions from E468 in the applicability metrics are:

• sections have been omitted to avoid repetition, e.g. Section 4.1 of7 ;
• owing to the observed lack of reporting of data, that is ‘desirable but not required’, only ‘minimum
information’ requirements have been included7 ;
• the requirement that terminology and nomenclature be compatible with ASTM standard E18239 was
not assessed.

2.1. Documentation clarity

In this work, the DC of a specific set of fatigue test data has been numerically evaluated with three scaling
factors, a null option is included to allow for unnecessary data10 :

• unambiguous (10);
• ambiguous (5);
• non-existent (0);
• unnecessary (-).
The scaling factor magnitude was agreed by the stakeholders associated with the design scenario inves-
tigated in this paper (Section 3). Alternate magnitude ranges can be applied as required for a specific
scenario.
Not all applicability metrics are necessary for all test scenarios. An applicability metric is unnecessary if
it does not contribute to the subsequent AR (Table III):

• When an omitted applicability metric negates the necessity of a subsequent metric, for example, if yield
strength is not reported, the associated strain offset is redundant. Potentially redundant applicability
metrics are identified with an ‘a’ suffix, for example in metric B2a.
• When an applicability metric is unnecessary within the specific fatigue test program, for example,
if no run-outs (i.e. no fatigue failure at a specified number of load cycles) were observed, a run-out
specification was deemed unnecessary. These redundant applicability metrics are identified with a ‘b’
suffix, for example in metric G5b.
Reference to external sources was not considered sufficient to contribute to the DC as the external sources
may be proprietary, difficult to obtain or unavailable (Section 3).

2.2. Applicability rating

The AR of a specific set of fatigue test data is evaluated by summing the DC associated with each of the
relevant applicability metrics. Each AR is expressed as a percentage of the maximum possible AR and may
be evaluated for two scenarios of importance:

• Scenario #1: all applicability metrics are of equal merit

The AR for Scenario #1 (A R1 ) is based on the ‘minimum information’ requirements of ASTM Standard
E468 with each applicability metric contributes equally to the AR.
• Scenario #2: core applicability metrics only
It is recognized that a differential weighting scheme may provide a more appropriate measure of
the applicability of fatigue test data for design applications. For example, the applicability metrics
associated with the material designation (A1, Table I) and the inclusion of a fitted curve (J2, Table I)
both contribute equally to the AR for Scenario #1 when, in fact, the contribution of the material

Copyright 2009 John Wiley & Sons, Ltd. Qual. Reliab. Engng. Int. 2009; 25:921–932
DOI: 10.1002/qre
924 M. LEARY AND C. BURVILL

designation to applicability is far greater than the inclusion of a fitted curve. ASTM Standard E468
reports that the graphical representation of fatigue data should include a caption (J3, Table I). This
caption is understood to represent the set of core applicability metrics that maximize the applicability
of the fatigue data when exported from its original context, for example, to a compilation of fatigue
data, such as4,5,11,12 . In Scenario #2 (A R2 ) only the core applicability metrics (Table I) are relevant
and non-core applicability metrics do not contribute to the AR (i.e. they have zero weighting).

Table I. Applicability metrics for reported fatigue test data, with reference
to the relevant section of Reference7
Code Applicability metric
E468 Section 5.1.1 Material prior to test specimen preparation √
A1 ‘the designation or specification . . . for example, A 441’
A2 ‘form of product (for example, plate, bar, casting, and so forth)’
A3 ‘heat number’∗
A4 ‘melting practice’
A5 ‘last mechanical working’
A6 ‘last heat treatment’
A7 ‘chemical composition’
A8 ‘surface condition . . . for example, rolled and descaled, ground’
E468 Section 5.1.2.1 Mechanical properties √
B1 ‘ tensile strength’
B2 ‘yield point or yield strength. . .’
B2a ‘. . .at a specified onset’
B3 ‘elongation. . .’
B3a ‘. . .in a specified gauge length’
B4 ‘the designation of the test used to procure the mechanical properties’
E468 Section 5.2.1 Shape, size, and dimensions
C1 ‘drawing showing shape, size, and dimensions of the fatigue test specimen’
C2 ‘details [tolerances] on test section, grip section, fillets, radii. . .’
C3 ‘orientation. . . with respect to the direction of maximum working’
E468 Section 5.3 Listing of information on specimen preparation†
D1 ‘process used to form the specimen (e.g. milling, turning, grinding, etc.)’
D2 ‘thermal treatment’ √
D3 ‘surface treatment (for example, shot-peening, nitriding. . .)’
√
D4b ‘the polishing sequence and direction should be listed’‡
D5 ‘procedures. . . used to eliminate. . . defects (that have occurred in storage)’
E468 Section 6.1 Design of the fatigue test program
E1 ‘statistical techniques. . . used to design the fatigue test program’
E468 Section 6.2 Fatigue testing machine
F1 ‘the type [manufacturer] of testing machine’
F2 ‘the functional characteristic. . . for
√ example, electrohydraulic’
F3 ‘frequency of force application’ √
F4 ‘forcing function (for example, sine, square, etc)’
F5 ‘the number of testing machines used’
F6 ‘method of dynamic force verification and force monitoring procedures’
E468 Section 6.3 Fatigue test √
G1 ‘type of test (axial, rotary bending, plane bending, or torsion)’
G2 ‘derivation (or method of computation) of. . . dynamic stresses’
G3b ‘when applicable, the experimental stress analysis techniques’
G4 ‘failure criterion’
G5b ‘Where required. . . the number of cycles to run-out’

Copyright 2009 John Wiley & Sons, Ltd. Qual. Reliab. Engng. Int. 2009; 25:921–932
DOI: 10.1002/qre
APPLICABILITY OF PUBLISHED DATA 925

Table I. Continued
Code Applicability metric
E468 Section 6.4 Ambient conditions during the fatigue test √
H1 ‘average value and ranges of both temperature and relative humidity’
E468 Section 6.5 Results of post-test examination
I1 ‘the reason for test termination, either achievement of the failure criterion or
run-out, and, if applicable, a description of the failure surface’
E468 Section 7.2 Graphical presentation √
J1 ‘graphical presentation of. . . all of the test results’
J2 ‘the faired or fitted curve’
J3 ‘diagram [J1] should include the designation [A1]. . .tensile strength [B1]. . .
surface condition [D3, D4]. . . type of fatigue test [G1]. . . stress parameter [F4]. . .
test frequency [F3], environment [J4] and temperature [H1]’
√
J4 ‘environment’§
√
Nomenclature: core applicability metric ( ).
∗ In this work, ‘A3’ is defined as the batch number associated with thermal processing of the material prior to
specimen preparation.
† This listing should occur in chronological order of specimen preparation 7. ASTM, E468-90(2004 )e1 Standard
Practice for Presentation of Constant Amplitude Fatigue Test Results for Metallic Materials, in Annual Book
of ASTM Standards. 2005, ASTM International: West Conshohocken, Pennsylvania, USA.
‡ ASTM E-606 recommends that specimens be ground and longitudinally lapped to a surface finish of 8 micro-
inches or better, with no circumferential grinding marks visible at 20× magnification 15. ASTM, E606-92(1998 )
Standard Practice for Strain-Controlled Fatigue Testing. 2005, ASTM International: West Conshohocken,
Pennsylvania13 .
§ In this work, J4 requires an explicit definition of the specimen operating environment, not inclusive of the
temperature and humidity requirements of H1.

3. APPLICABILITY RATING OF SAMPLED REFERENCES

Fatigue test data of relevance to the authors’ FLSC component redesign project was sourced from the
literature (Table II). This sample is representative of the AR for a specific research field, i.e. mass reduction
of FLSC ferrous metal components by substitution with light alloys1 . However, it is not expected that the
sample is representative of the AR for research data available from other fields of fatigue research, or for
the AR for the overall population of available fatigue data.
Furthermore, the presentation forum and research scope of the sampled references have been identified
to allow qualitative insight into correlations between these attributes and the associated ARs. Journal and
conference papers are subject to peer review, whereas there is no expectation that technical papers or books
have undergone a peer review process. The samples were limited to those that meet the scope of ASTM
Standard E468, i.e. ‘data derived from constant-force amplitude axial, bending, or torsional fatigue tests of
metallic materials’7 .
The DC and the associated ARs were evaluated (Table III) for the sampled references. The DC was
identified as ambiguous if the requirements of the applicability metrics were not fully reported, for example:
• Applicability metric C2: Stating only nominal specimen geometry is ambiguous, for example, Reference
#G. An explicit specification of the associated geometric tolerances is unambiguous, for example,
Reference #H (Figure 1).
• Applicability metric H1: Stating operating environment as ‘room temperature’ only is ambiguous,
for example, Reference #F. An explicit specification of the laboratory temperature and humidity is
unambiguous, for example, Reference #D.
• Applicability metric F4: Stating the associated R-ratio only is ambiguous, for example, Reference #N.
An explicit definition of the associated waveform is unambiguous, for example, Reference #C.

Copyright 2009 John Wiley & Sons, Ltd. Qual. Reliab. Engng. Int. 2009; 25:921–932
DOI: 10.1002/qre
926 M. LEARY AND C. BURVILL

Table II. Sample references of fatigue test data and associated attributes

Reference label
(Table III, Figure 2) Forum Research scope

#A T Squeeze cast aluminium FLSC automotive component design14

#B J Mechanism of fatigue failure in the ultra-long life regime15
#C J The reliability of dynamically loaded automotive gearwheels16
#D J Porosity and the fatigue limit of die cast aluminium and magnesium17
#E C Fatigue of squeeze, semisolid, and gravity diecast aluminium18
#F J Effect of surface coatings on fatigue strength19
#G J Influence of load sequence on fatigue life, SLOC20
#H J Influence of bending overloads on fatigue life, SLOC21
#I J Surface roughness and the fatigue life of squeeze cast aluminium22
#J J Influence of size (volume) effects on aluminium fatigue life23
#K T The reliability of automotive aluminium brake calipers24
#L J Effect of shot peening on high cycle fatigue of cast iron25
#M J Defect tolerant design of automotive components26
#N J Influence of surface roughness on forged aluminium fatigue life27
#O C Influence of ECAP∗ processing on 6061 fatigue strength28
#P B Magnesium product design29
#Q B Fatigue characteristics of cast and wrought 1040 steel12
#R B Stress concentration and the fatigue strength of 6061-T6 aluminium30
#S B Fatigue strength of wrought 6061-T6 aluminium5
#T B Stress-life curves for 6061-T6 aluminium31
#U B Fatigue strength of various 6061-T6 aluminium specimens4
#V B Fatigue strength of AZ31-B magnesium4
#W B Fatigue strength of AZ31-B magnesium5

Nomenclature: technical paper—not peer reviewed (T), book (B), conference paper—peer reviewed (C), journal
paper—peer reviewed (J), fatigue-limited safety–critical (FLSC), small load opmission criteria (SLOC).
∗ Equal channel angular pressing (ECAP) is a manufacturing process that significantly refines grain size.

The ARs and average DC of the sampled references are presented graphically in Figures 2 and 3, respectively.
The associated statistical distribution is presented in Table IV.

3.1. Analysis of sampled references

There is significant mismatch between the DC of the sampled references and the minimum requirements
of ASTM Standard E468, and in general the applicability of the sampled references is low. This outcome
supports the assertion that the applicability of published fatigue data is often insufficient, for example8 .
Cross-examination of the sampled data identifies a series of potential bases for this impediment:
Perceived importance: The observed AR (Table IV) is greater for Scenario #2 than for Scenario #1
(Section 2.2), i.e. the documentation provided in the sampled references is skewed towards the core appli-
cability metrics:
• A R1 ∈ (17, 62%), A R 1 = 36%
• A R2 ∈ (20, 95%), A R 2 = 56%
This outcome suggests that the originators of the fatigue test data perceived the core applicability metrics as
being of greater importance to applicability than the other metrics. If Scenario #2 is used as a basis for appli-
cability, some of the sampled data may contain sufficient documentation to be applied for design purposes
without incurring significant uncertainties in fatigue-life prediction. Based on the surveyed references, some
applicability metrics were perceived to have little to no importance. For example: heat number (A3), melting
practice (A4), storage defect procedures (D5), statistical techniques (E1), and a fully documented graphical
presentation (J3) were not documented in any of the sampled references.

Copyright 2009 John Wiley & Sons, Ltd. Qual. Reliab. Engng. Int. 2009; 25:921–932
DOI: 10.1002/qre
 Copyright
Table III. Applicability ratings assessed for the sampled references of fatigue test data (Tables II and III)
Applicability metric
√ √ √ √ √ √ √ √ √ √
AR1 AR2
A1 A2 A3 A4 A5 A6 A7 A8 B1 B2 B2a B3 B3a B4 C1 C2 C3 D1 D2 D3 D4b D5 E1 F1 F2 F3 F4 F5 F6 G1 G2 G3b G4 G5b H1 I1 J1 J2 J3 J4 (%) (%)
Reference label
#A U U X X A U X U U U X U X X X X X X U X X X X X X X X X X X X N X N X X X U X X 25 20
#B U X X X X U U X X X N X N N U U X X U U N X X X X X A X X U X N X U X X U X X X 30 50
APPLICABILITY OF PUBLISHED DATA

#C U X X X X U X U X X N X N N A X X U U U N X X X U X U X X U U N U N X X U U X X 40 56

2009 John Wiley & Sons, Ltd.

#D U U X X U X U U U U U U X X U A X U X U U X X X U U A X U U U N X A U A U U X U 62 95
#E U U X X A U U U U U U U X X X X U U U X X X X X X X A X X X X N X U U X U U X X 44 45
#F U U X X U U A U U U X X N X U U X U U U N X X X X X U X X U X N U N A X U U X X 50 72
#G U U X X X U X X X X N X N N U A X U U U N X X U X U U U X U X N U N X X U X X X 43 67
#H U X X X X U U X X X X X X N U U X U U U A X X U U U A U U U X N U A X X U U X X 49 60
#I U U X X X U U X U U X X N X X X X U U U U X X X X X A X X U X N X A X X U X X X 34 65
#J U U X X A U X X U U X X N X X X X X U X X X X X X X A X X X X N U A A X U U X U 32 50
#K U U X X U A U U U U U U U U U U X X X X X X X X X U U X U U X N X N A X U U X X 53 65
#L U U X X X X U X X X N X N X U A X A X U N X X X X U X X X U X N X A X X U U X X 29 56
#M U U X X U U X X U U U X N X U A X U U U N X X X U U X U X U X N X A X X U X X X 46 67
#N U U X X U U X U U U N X N X X X X X U U A X X X X X A X X U X N U A X X U X X X 36 60
#O U U X X U U X U U U X X N X X X X X U X X X X X X U U X X X X N X N A X U X X U 34 65
#P U U X X A U U X U X X X X X X X X X X U X X X X U X U X X U X N U N X X X X X X 28 50
#Q U U X X U U X X U U X U X X X X X X X X N X X X U X X X X U X N X N X X X U X X 27 33
#R U X X X X U X X X X N X N N X X X X X X X X X X X U X X X U X N X A X X U U X X 18 40
#S U U X X A U X U U U A X N X X X X X X X X X X X X U X X X U X N X A A X U U X U 34 65
#T U X X X X U X X X X N X N X X X X X X U X X X X U X U X X U X N X N X X U U X X 22 50
#U U U X X X U X X X X N X N X X X X X X U X X X X X X X X X U X N X N X X X X X U 17 40
#V U U X X A U U X U U X U X X X X X X X U X X X X U X U X X U X N U N X X X X X X 33 50
#W U U X X A U X U U U A U X X X X X X U X U X X X X U X X X U X N X A A X U U X U 41 75
Average applicability metric response (%)
100 80 0 0 50 83 50 53 67 67 40 27 20 9 57 40 7 57 80 67 33 0 0 13 27 47 57 20 20 73 13 E 40 61 27 3 93 60 0 20
Nomenclature: applicability rating (AR), AR for Scenario #1 (A R1 ), AR for Scenario #2 (A R2 ), documentation clarity (DC), unambiguous DC (U), ambiguous

required for any of the sampled references (E), core applicability metric ( ). Numerical DC weighting identified in Section 2.1.
DC (A), non-existent DC (X), no documentation required as applicability√metric unnecessary (N), error in computing average as the applicability metric was not

DOI: 10.1002/qre
Qual. Reliab. Engng. Int. 2009; 25:921–932
927
928 M. LEARY AND C. BURVILL

Figure 1. Left: ambiguous documentation clarity, i.e. only nominal geometry documented, Reference #G. Right: unam-
biguous documentation clarity, i.e. nominal and allowable geometry explicitly documented, Reference #H

Figure 2. ARs for the sampled references (Table II) categorized as journal or non-journal and sorted in descending order
of A R2 . A R1 = 100% corresponds to unambiguous documentation of the minimum requirements of ASTM Standard
E468. A R2 = 100% corresponds to unambiguous documentation of the core requirements of ASTM Standard E468.
Nomenclature: applicability rating (AR), AR for Scenario #1 (A R1 ), AR for Scenario #2 (A R2 ), mean AR of journal
for Scenario #2 (A R 2J ), mean AR of non-journal for scenario #2 (A R 2N J ). Histogram legend: A R1 (black bar), A R2
(white bar). Dashed lines indicate 95% distribution around the AR means

Presentation forum: It appears that journal proceedings are a more fertile source for applicable fatigue
test data than non-journal data. However, independently of the forum of the published fatigue data, the low
applicability of the sampled references indicates that, in general, the designer must accept uncertainties in
the associated test conditions if published fatigue data are used in a design task.
Research intent: A qualitative review of the sampled references suggests that AR typically reflects the
associated research intent, for example (Table I):
• Tensile properties are not reported in Reference #H as they are irrelevant to the research intent, i.e. the
influence of load sequence on the fatigue properties of otherwise similar test specimens.
• Where the influence of surface finish is of particular interest, the associated applicability metrics are
well defined, for example, Reference #I.
For design tasks that match the research intent associated with the test data under consideration, the scenario-
specific applicability may be greater than that identified by the proposed AR.
External referencing: Several of the sampled references referred to external sources to provide further
information. However, external sources may be difficult to obtain, proprietary or unavailable, severely

Copyright 2009 John Wiley & Sons, Ltd. Qual. Reliab. Engng. Int. 2009; 25:921–932
DOI: 10.1002/qre
APPLICABILITY OF PUBLISHED DATA 929

Figure 3. Mean documentation clarity of the applicability metrics for the sampled references (Table II).
Nomenclature: documentation clarity (DC), mean DC for Scenario #1 (DC 1 ), mean DC for Scenario #2
√
(DC 2 ), core applicability metric ( )

Table IV. Statistical distribution of sampled references of fatigue test data (Table II)

Minimum Maximum Mean s 95% range

A R1 17 62 36 11 13 59
A R2 20 95 56 16 25 88
A R1J 29 62 41 10 21 61
A R2J 50 95 63 13 38 89
A R1N J 17 53 31 11 10 53
A R2N J 20 75 50 16 19 81

Nomenclature: applicability rating (AR), AR for Scenario #1 (A R1 ), AR for Scenario #2

(A R2 ), standard deviation (s), journal (J), non-journal (NJ). 95% range is based on a normal
distribution, i.e. mean ±2s.

limiting the applicability of the presented data for design. For example, Reference #V pertains to the fatigue
strength of AZ31-B magnesium; however, the associated applicability rating is low. Reference is made to
an external source29 that in turn provides incomplete documentation and refers to internal company reports
for further details. As these external references are unlikely to be available, this fatigue test data are unlikely
to be sufficient for design applications. If no more robust data source can be identified for the material of
interest (for example, Reference #W), the designer must accept either the additional complexity of explicit
testing, or the reduced certainty in fatigue life prediction associated with using the poorly documented data.

4. CONCLUSIONS

An extensive set of documentation is required to fully define the test conditions of relevance to the fatigue
failure mode. Incomplete documentation introduces uncertainties in fatigue life. These uncertainties may
lead to isolated domains of applicable fatigue test data, where the originator of the test data is aware of the

Copyright 2009 John Wiley & Sons, Ltd. Qual. Reliab. Engng. Int. 2009; 25:921–932
DOI: 10.1002/qre
930 M. LEARY AND C. BURVILL

test conditions, but this knowledge is not documented, and the applicability of the associated fatigue test
data is limited.
The proposed AR offers a novel basis to quantify the applicability of publicly available fatigue test data
based on internationally recognized standards for fatigue test documentation. Two variants of the AR have
been proposed, additional variants may be developed for scenarios of importance to particular research and
design applications.
The published fatigue data reviewed in this paper exhibit incomplete documentation. This outcome matches
assertions in the literature that published fatigue test data are often of limited applicability due to insufficient
documentation of the test conditions. The ensuing uncertainties are not compatible with the production
imperatives of FLSC component design and may necessitate a repetition of fatigue testing, or limit the
range of materials that may be confidently considered for material selection and optimization. Analysis of
the sampled data identified a series of potential bases for limited applicability, including perceptions of
importance, presentation forum, research intent, and the use of external referencing.
Furthermore, fatigue failure is a ‘weak-link’ mechanism, where minor variation in processing, loading,
and environment can result in significant statistical scatter in fatigue life. In addition to ensuring sufficient
test documentation by the AR method defined in this paper, a robust fatigue design strategy requires that
the fatigue test scope sufficiently captures the statistical scatter of the specimens. This can be achieved by
evaluating the sample size and percentage replication according to ASTM requirements32 . This outcome is
especially important for safety–critical applications and scenarios where fatigue failure incurs significant
cost or time penalties.
When reporting the outcomes of an experimental fatigue test program, the AR should be assessed to
confirm the applicability of the published data for fatigue-limited design.

REFERENCES
1. Leary M. Mass Reduction of Fatigue-Limited, Safety-Critical (FLSC ), Ferrous Metal Automotive Components by
Forged Light Alloy Substitution, Melbourne University, Melbourne, Australia, 2006.
2. Veers PS. Statistical considerations in fatigue. ASM Handbook, Volume 19: Fatigue and Fracture. ASM International:
Materials Park, OH, U.S.A., 1996.
3. Fermer M, Svensson H. Industrial experiences of FE-based fatigue life predictions of welded structures. Fatigue and
Fracture of Engineering Materials and Structures 2001; 24(7):489–500.
4. Lampman S (ed.). Fatigue Data Book: Light Structural Alloys. ASM International: Materials Park, OH, U.S.A., 1995.
5. Batelle. Military Handbook Metallic Materials and Elements for Aerospace Vehicle Structures, Batelle, 1652, 1998.
6. Rice RC. Fatigue data analysis. Metals Handbook, Volume 8: Mechanical Testing. ASTM: Ohio, 1985.
7. ASTM, E468-90(2004)e1. Standard practice for presentation of constant amplitude fatigue test results for metallic
materials. Annual Book of ASTM Standards. ASTM International: West Conshohocken, Pennsylvania, U.S.A., 2005.
8. Cameron DW, Allegany NY, Hoeppner DW. Fatigue properties in engineering. ASM Handbook, Volume 19: Fatigue
and Fracture. ASM International: Materials Park, OH, U.S.A., 1996; 15–26.
9. ASTM, E1823-96(2002). Standard terminology relating to fatigue and fracture testing. Annual Book of ASTM Standards.
ASTM International: West Conshohocken, Pennsylvania, U.S.A., 2005.
10. Leary M, Burvill C. Enhancing the quality function deployment conceptual design tool. Journal of Mechanical Design
2007; 129(7):701.
11. Brown WF. Aerospace Structural Metals Handbook. Nonferrous Alloys. Mechanical Properties Data Center: Traverse
city Michigan, U.S.A., 1972.
12. Boyer H (ed.). Atlas of Fatigue Curves. ASM International: Materials Park, OH, U.S.A., 1986.
13. ASTM, E606-92(1998). Standard Practice for Strain-Controlled Fatigue Testing. ASTM International: West
Conshohocken, Pennsylvania, 2005.
14. Gerken DT, Neal R. Squeeze Cast (SCPM ) Light Weight Front Knuckle Case Study. SAE: Detroit, 1999.
15. Murakami Y, Yokoyama NN, Nagata J. Mechanism of fatigue failure in the ultralong life regime. Fatigue and Fracture
of Engineering Materials and Structures 2002; 25:735–746.
16. Yang QJ. Fatigue test and reliability design of gears. International Journal of Fatigue 1996; 18(3):171–177.
17. Mayer H, Papakyriacou M, Zettl B, Stanzl-Tschegg SE. Influence of porosity on the fatigue limit of die cast magnesium
and aluminium alloys. International Journal of Fatigue 2003; 25(3):245–256.

Copyright 2009 John Wiley & Sons, Ltd. Qual. Reliab. Engng. Int. 2009; 25:921–932
DOI: 10.1002/qre
APPLICABILITY OF PUBLISHED DATA 931

18. Davidson CJ, Griffiths JR, Zanada A. Fatigue properties of squeeze, semisolid and gravity diecast Al–Si–Mg alloy.
Structural Integrity and Fracture International Conference (SIF’04 ), Brisbane, Australia, 2004.
19. Nascimento MP, Souza RC, Pigatin WL, Voorwald HJC. Effects of surface treatments on the fatigue strength of AISI
4340 aeronautical steel. International Journal of Fatigue 2001; 23(7):607–618.
20. Ngiau C, Kujawski D. Sequence effects of small amplitude cycles on fatigue crack initiation and propagation in
2024-T351 aluminium. International Journal of Fatigue 2001; 23(9):807–815.
21. Bonnen JJF, Topper TH. The effect of bending overloads on torsional fatigue in normalised 1045 steel. International
Journal of Fatigue 1999; 21(1):23–33.
22. Güngör S, Edwards L. Effect of surface texture on fatigue life in a squeeze-cast 6082 aluminium alloy. Fatigue and
Fracture of Engineering Materials and Structures 1993; 16(4):391–403.
23. Bazios I, Decker M, Gudladt H-J. The influence of size effects on the fatigue behaviour of the aluminium alloy
AlMgSi0.7. Key Engineering Materials 2003; 251–252(203):203–208.
24. Keshavaram BN. Aluminium Alloys for Automotive Disk Callipers. SAE: Detroit, 1999.
25. Ochi Y, Masaki K, Matsumura T, Sekino T. Effect of shot-peening treatment on high cycle fatigue property of ductile
cast iron. International Journal of Fatigue 2001; 23(5):441–448.
26. Beretta S, Blarasin A, Endo M, Giunti T, Murakami Y. Defect tolerant design of automotive components. International
Journal of Fatigue 1997; 19(4):319–333.
27. Güngör S, Edwards L. Effect of surface texture on the initiation and propagation of small fatigue cracks in a forged
6082 aluminium alloy. Materials Science and Engineering 1993; A160(1):17–24.
28. Chung CS, Kim JK, Kim HK, Kim WJ. Enhancement in strength and fatigue resistance of 6061Al alloy produced
by equal-channel angular pressing. Light Materials for Transportation Systems. Pohang University of Science and
Technology: Pusan, Korea, 2001.
29. Busk RS. Magnesium Product Design. Marcel Decker: Ohio, 1987.
30. Madayag AF. Metal Fatigue: Theory and Design. Wiley: New York, 1969.
31. Dowling NE. Mechanical Behaviour of Metals. Prentice-Hall: Englewood Cliffs, NJ, 1993.
32. ASTM, E739-91 (2004). Standard practice for statistical analysis of linear or linearized stress-life (S-N) and strain-life
(E-N) fatigue data. Annual Book of ASTM Standards. ASTM International: West Conshohocken, Pennsylvania, U.S.A.
2005.

APPENDIX: NOMENCLATURE
AR applicability rating
A R1 applicability rating for Scenario #1
A R2 applicability rating for Scenario #2
A R1 mean applicability rating for Scenario #1
A R2 mean applicability rating for Scenario #2
A R 1J mean AR of a journal for Scenario #1
A R 2J mean AR of a journal for Scenario #2
A R 1N J mean AR of a non-Journal for Scenario #1
A R 2N J mean AR of a non-Journal for Scenario #2
DC documentation clarity
DC 1 mean DC for Scenario #1
DC 2 mean DC for Scenario #2
S standard deviation

Authors’ biographies
Dr Martin Leary is a Senior Lecturer in the School of Aerospace, Mechanical and Manufacturing Engi-
neering at the RMIT University. His current undergraduate teaching responsibilities include: design manage-
ment, numeric optimisation and mechanical design. His professional interests focus on the commercialisation

Copyright 2009 John Wiley & Sons, Ltd. Qual. Reliab. Engng. Int. 2009; 25:921–932
DOI: 10.1002/qre
932 M. LEARY AND C. BURVILL

of scholarly research in collaboration with industry partners. Specific research areas include: design opti-
mization, material selection and engineering design education.
Dr Colin Burvill is a Senior Lecturer in the Department of Mechanical and Manufacturing Engineering
at the University of Melbourne. His lectures are in mechanical design and strength of materials at the
undergraduate level. He is the Founding Coordinator of the Capstone Project for final year students. His
research interests focus on issues associated with ‘design for manufacturability’, ‘material performance’
and ‘equine bioengineering’. He has successfully integrated research outcomes with the needs of third-party
collaborators through a series of post-graduate and post-doctoral research and development projects. He is
a member of the Design Engineering Division of the American Society of Mechanical Engineers.

Copyright 2009 John Wiley & Sons, Ltd. Qual. Reliab. Engng. Int. 2009; 25:921–932
DOI: 10.1002/qre