This international standard was developed in accordance with internationally recognized principles on standardization established in the Decision on Principles

for the
Development of International Standards, Guides and Recommendations issued by the World Trade Organization Technical Barriers to Trade (TBT) Committee.

Designation: E739 10 (Reapproved 2015)

Standard Practice for

Statistical Analysis of Linear or Linearized Stress-Life (S-N)
and Strain-Life (-N) Fatigue Data1
This standard is issued under the fixed designation E739; the number immediately following the designation indicates the year of
original adoption or, in the case of revision, the year of last revision. A number in parentheses indicates the year of last reapproval. A
superscript epsilon () indicates an editorial change since the last revision or reapproval.

1. Scope 3. Terminology
1.1 This practice covers only S-N and -N relationships that 3.1 The terms used in this practice shall be used as defined
may be reasonably approximated by a straight line (on appro- in Definitions E206 and E513. In addition, the following
priate coordinates) for a specific interval of stress or strain. It terminology is used:
presents elementary procedures that presently reflect good 3.1.1 dependent variablethe fatigue life N (or the loga-
practice in modeling and analysis. However, because the actual rithm of the fatigue life).
S-N or -N relationship is approximated by a straight line only 3.1.1.1 DiscussionLog (N) is denoted Y in this practice.
within a specific interval of stress or strain, and because the 3.1.2 independent variablethe selected and controlled
actual fatigue life distribution is unknown, it is not recom- variable (namely, stress or strain). It is denoted X in this
mended that (a) the S-N or -N curve be extrapolated outside practice when plotted on appropriate coordinates.
the interval of testing, or (b) the fatigue life at a specific stress 3.1.3 log-normal distributionthe distribution of N when
log (N) is normally distributed. (Accordingly, it is convenient
or strain amplitude be estimated below approximately the fifth
to analyze log (N) using methods based on the normal
percentile (P . 0.05). As alternative fatigue models and
distribution.)
statistical analyses are continually being developed, later
3.1.4 replicate (repeat) testsnominally identical tests on
revisions of this practice may subsequently present analyses
different randomly selected test specimens conducted at the
that permit more complete interpretation of S-N and -N data. same nominal value of the independent variable X. Such
replicate or repeat tests should be conducted independently; for
2. Referenced Documents example, each replicate test should involve a separate set of the
2.1 ASTM Standards:2 test machine and its settings.
E206 Definitions of Terms Relating to Fatigue Testing and 3.1.5 run outno failure at a specified number of load
the Statistical Analysis of Fatigue Data; Replaced by cycles (Practice E468).
E 1150 (Withdrawn 1988)3 3.1.5.1 DiscussionThe analyses illustrated in this practice
E468 Practice for Presentation of Constant Amplitude Fa- do not apply when the data include either run-outs (or
tigue Test Results for Metallic Materials suspended tests). Moreover, the straight-line approximation of
E513 Definitions of Terms Relating to Constant-Amplitude, the S-N or -N relationship may not be appropriate at long lives
Low-Cycle Fatigue Testing; Replaced by E 1150 (With- when run-outs are likely.
drawn 1988)3 3.1.5.2 DiscussionFor purposes of statistical analysis, a
E606/E606M Test Method for Strain-Controlled Fatigue run-out may be viewed as a test specimen that has either been
Testing removed from the test or is still running at the time of the data
analysis.

4. Significance and Use

1
This practice is under the jurisdiction of ASTM Committee E08 on Fatigue and
Fracture and is the direct responsibility of Subcommittee E08.04 on Structural 4.1 Materials scientists and engineers are making increased
Applications. use of statistical analyses in interpreting S-N and -N fatigue
Current edition approved Oct. 1, 2015. Published November 2015. Originally data. Statistical analysis applies when the given data can be
approved in 1980. Last previous edition approved in 2010 as E739 10. DOI:
10.1520/E0739-10R15. reasonably assumed to be a random sample of (or representa-
2
For referenced ASTM standards, visit the ASTM website, www.astm.org, or tion of) some specific defined population or universe of
contact ASTM Customer Service at [email protected]. For Annual Book of ASTM material of interest (under specific test conditions), and it is
Standards volume information, refer to the standards Document Summary page on
desired either to characterize the material or to predict the
the ASTM website.
3
The last approved version of this historical standard is referenced on performance of future random samples of the material (under
www.astm.org. similar test conditions), or both.

Copyright ASTM International, 100 Barr Harbor Drive, PO Box C700, West Conshohocken, PA 19428-2959. United States

1
 E739 10 (2015)
5. Types of S-N and -N Curves Considered 5.1.1 The fatigue life N is the dependent (random) variable
5.1 It is well known that the shape of S-N and -N curves in S-N and -N tests, whereas S or is the independent
can depend markedly on the material and test conditions. This (controlled) variable.
practice is restricted to linear or linearized S-N and -N NOTE 2In certain cases, the independent variable used in analysis is
relationships, for example, not literally the variable controlled during testing. For example, it is
log N 5 A1B ~ S ! or (1) common practice to analyze low-cycle fatigue data treating the range of
plastic strain as the controlled variable, when in fact the range of total
strain was actually controlled during testing. Although there may be some
log N 5 A1B ~ ! or
question regarding the exact nature of the controlled variable in certain
log N 5 A1B ~ logS ! or (2) S-N and -N tests, there is never any doubt that the fatigue life is the
dependent variable.
log N 5 A1B ~ log ! NOTE 3In plotting S-N and -N curves, the independent variables S
in which S and may refer to (a) the maximum value of and are plotted along the ordinate, with life (the dependent variable)
constant-amplitude cyclic stress or strain, given a specific plotted along the abscissa. Refer, for example, to Fig. 1.
value of the stress or strain ratio, or of the minimum cyclic
stress or strain, (b) the amplitude or the range of the 5.1.2 The distribution of fatigue life (in any test) is unknown
constant-amplitude cyclic stress or strain, given a specific (and indeed may be quite complex in certain situations). For
value of the mean stress or strain, or (c) analogous informa- the purposes of simplifying the analysis (while maintaining
tion stated in terms of some appropriate independent (con- sound statistical procedures), it is assumed in this practice that
trolled) variable. the logarithms of the fatigue lives are normally distributed, that
NOTE 1In certain cases, the amplitude of the stress or strain is not
constant during the entire test for a given specimen. In such cases some
is, the fatigue life is log-normally distributed, and that the
effective (equivalent) value of S or must be established for use in variance of log life is constant over the entire range of the
analysis. independent variable used in testing (that is, the scatter in log

NOTE 1The 95 % confidence band for the -N curve as a whole is based on Eq 10. (Note that the dependent variable, fatigue life, is plotted here along
the abscissa to conform to engineering convention.)
FIG. 1 Fitted Relationship Between the Fatigue Life N (Y) and the Plastic Strain Amplitude p/2 (X) for the Example Data Given

2
 E739 10 (2015)
N is assumed to be the same at low S and levels as at high Minimum Number
Type of Test
levels of S or ). Accordingly, log N is used as the dependent of SpecimensA
(random) variable in analysis. It is denoted Y. The independent Preliminary and exploratory (exploratory research and 6 to 12
variable is denoted X. It may be either S or , or log S or log , development tests)
respectively, depending on which appears to produce a straight Research and development testing of components and 6 to 12
specimens
line plot for the interval of S or of interest. Thus Eq 1 and Eq Design allowables data 12 to 24
2 may be re-expressed as Reliability data 12 to 24

Y 5 A1BX (3) A
If the variability is large, a wide confidence band will be obtained unless a large
Eq 3 is used in subsequent analysis. It may be stated more number of specimens are tested (See 8.1.1).

precisely as Y ? X 5A1BX, where Y ? X is the expected value of Y 7.1.2 ReplicationThe replication guidelines given in
given X. Chapter 3 of Ref (1) are based on the following definition:
NOTE 4For testing the adequacy of the linear model, see 8.2.
NOTE 5The expected value is the mean of the conceptual population % replication = 100 [1 (total number of different stress or strain levels used
in testing/total number of specimens tested)]
of all Ys given a specific level of X. (The median and mean are identical
for the symmetrical normal distribution assumed in this practice for Y.) Type of Test Percent ReplicationA

6. Test Planning Preliminary and exploratory (research and development 17 to 33 min

tests)
6.1 Test planning for S-N and -N test programs is discussed Research and development testing of components and 33 to 50 min
in Chapter 3 of Ref (1).4 Planned grouping (blocking) and specimens
randomization are essential features of a well-planned test Design allowables data 50 to 75 min
Reliability data 75 to 88 min
program. In particular, good test methodology involves use of
planned grouping to (a) balance potentially spurious effects of A
Note that percent replication indicates the portion of the total number of
nuisance variables (for example, laboratory humidity) and (b) specimens tested that may be used for obtaining an estimate of the variability of
allow for possible test equipment malfunction during the test replicate tests.
program. 7.1.2.1 Replication ExamplesGood replication: Suppose
7. Sampling that ten specimens are used in research and development for
the testing of a component. If two specimens are tested at each
7.1 It is vital that sampling procedures be adopted that of five stress or strain amplitudes, the test program involves
assure a random sample of the material being tested. A random 50 % replications. This percent replication is considered ad-
sample is required to state that the test specimens are repre- equate for most research and development applications. Poor
sentative of the conceptual universe about which both statisti- replication: Suppose eight different stress or strain amplitudes
cal and engineering inference will be made. are used in testing, with two replicates at each of two stress or
NOTE 6A random sampling procedure provides each specimen that strain amplitudes (and no replication at the other six stress or
conceivably could be selected (tested) an equal (or known) opportunity of strain amplitudes). This test program involves only 20 %
actually being selected at each stage of the sampling process. Thus, it is replication, which is not generally considered adequate.
poor practice to use specimens from a single source (plate, heat, supplier)
when seeking a random sample of the material being tested unless that
particular source is of specific interest. 8. Statistical Analysis (Linear Model Y = A + BX, Log-
NOTE 7Procedures for using random numbers to obtain random Normal Fatigue Life Distribution with Constant
samples and to assign stress or strain amplitudes to specimens (and to Variance Along the Entire Interval of X Used in
establish the time order of testing) are given in Chapter 4 of Ref (2). Testing, No Runouts or Suspended Tests or Both,
7.1.1 Sample SizeThe minimum number of specimens Completely Randomized Design Test Program)
required in S-N (and -N) testing depends on the type of test
8.1 For the case where (a) the fatigue life data pertain to a
program conducted. The following guidelines given in Chapter
random sample (all Y i are independent), (b) there are neither
3 of Ref (1) appear reasonable.
run-outs nor suspended tests and where, for the entire interval
of X used in testing, (c) the S-N or -N relationship is described
by the linear model Y = A + BX (more precisely by Y ? X 5
A + BX), (d) the (two parameter) log-normal distribution
describes the fatigue life N, and (e) the variance of the
log-normal distribution is constant, the maximum likelihood
estimators of A and B are as follows:

A 5 Y
2 B X (4)
k

( ~ X 2 X ! ~ Y 2 Y !
i51
i i
B 5 k (5)
2
( ~ X 2 X !
i51
i
4
The boldface numbers in parentheses refer to the list of references appended to
this standard. where the symbol caret ( ^ ) denotes estimate (estimator),

3
 E739 10 (2015)

the symbol overbar () denotes average (for example, Y dence interval. This table has one entry parameter (the statis-
k k tical degrees of freedom, n, for t ). For Eq 7 and Eq 8, n =
5 ( Yi/k and X 5 ( Xi/k), Yi = log Ni, Xi = Si or i, or log Si or k 2.
i51 i51
log i (refer to Eq 1 and Eq 2), and k is the total number of test NOTE 9The confidence intervals for A and B are exact if conditions
specimens (the total sample size). The recommended expres- (a) through (e) in 8.1 are met exactly. However, these intervals are still
sion for estimating the variance of the normal distribution for reasonably accurate when the actual life distribution differs slightly from
log N is the (two-parameter) log-normal distribution, that is, when only condition
(d) is not met exactly, due to the robustness of the t statistic.
k
2 NOTE 10Because the actual median S-N or -N relationship is only
2
( ~ Y 2 Y !
i51
i i approximated by a straight line within a specific interval of stress or strain,
5 (6) confidence intervals for A and B that pertain to confidence levels greater
k22
than approximately 0.95 are not recommended.
in which i = + BXi and the (k 2) term in the denomi-
nator is used instead of k to make 2 an unbiased estimator of 8.1.1.1 The meaning of the confidence interval associated
the normal population variance 2. with, say, Eq 8 is as follows (Note 11). If the values of tp given
in Table 1 for, say, P = 95 % are used in a series of analyses
NOTE 8An assumption of constant variance is usually reasonable for involving the estimation of B from independent data sets, then
notched and joint specimens up to about 106 cycles to failure. The variance
of unnotched specimens generally increases with decreasing stress (strain) in the long run we may expect 95 % of the computed intervals
level (see Section 9). If the assumption of constant variance appears to be to include the value B. If in each instance we were to assert that
dubious, the reader is referred to Ref (3) for the appropriate statistical test. B lies within the interval computed, we should expect to be
8.1.1 Confidence Intervals for Parameters A and BThe correct 95 times in 100 and in error 5 times in 100: that is, the
estimators and B are normally distributed with expected statement B lies within the computed interval has a 95 %
values A and B, respectively, (regardless of total sample size k) probability of being correct. But there would be no operational
when conditions (a) through (e) in 8.1 are met. Accordingly, meaning in the following statement made in any one instance:
confidence intervals for parameters A and B can be established The probability is 95 % that B falls within the computed
using the t distribution, Table 1. The confidence interval for A interval in this case since B either does or does not fall within
is given by 6 tp , or the interval. It should also be emphasized that even in
1
X 2
independent samples from the same universe, the intervals
A 6t p
3k
1 k

( ~X
i51
i 2 X!
2
4 , (7) given by Eq 8 will vary both in width and position from sample
to sample. (This variation will be particularly noticeable for
small samples.) It is this series of (random) intervals fluctu-
and for B is given by B 6 tp B, or ating in size and position that will include, ideally, the value
B 6t p F(~
k

i51
X i 2 X !
2
G 2

(8)
B 95 times out of 100 for P = 95 %. Similar interpretations
hold for confidence intervals associated with other confidence
in which the value of tp is read from Table 1 for the desired levels. For a given total sample size k, it is evident that the
value of P, the confidence level associated with the confi- width of the confidence interval for B will be a minimum
whenever
k
2

TABLE 1 Values of tp (Abstracted from STP 313 (4)) ( ~ X 2 X !

i51
i (9)

P, %B is a maximum. Since the Xi levels are selected by the

nA investigator, the width of confidence interval for B may be
90 95
reduced by appropriate test planning. For example, the width
4 2.1318 2.7764
5 2.0150 2.5706
of the interval will be minimized when, for a fixed number
6 1.9432 2.4469 of available test specimens, k, half are tested at each of the
7 1.8946 2.3646 extreme levels Xmin and X max. However, this allocation
8 1.8595 2.3060 should be used only when there is strong a priori knowledge
9 1.8331 2.2622
10 1.8125 2.2281
that the S-N or -N curve is indeed linearbecause this allo-
11 1.7959 2.2010 cation precludes a statistical test for linearity (8.2). See
12 1.7823 2.1788 Chapter 3 of Ref (1) for a further discussion of efficient se-
13 1.7709 2.1604
14 1.7613 2.1448
lection of stress (or strain) levels and the related specimen
15 1.7530 2.1315 allocations to these stress (or strain) levels.
16 1.7459 2.1199 NOTE 11This explanation is similar to that of STP 313 (4).
17 1.7396 2.1098
18 1.7341 2.1009 8.1.2 Confidence Band for the Entire Median S-N or -N
19 1.7291 2.0930
20 1.7247 2.0860
Curve (that is, for the Median S-N or -N Curve as a Whole)
21 1.7207 2.0796 If conditions (a) through (e) in 8.1 are met, an exact confidence
22 1.7171 2.0739 band for the entire median S-N or -N curve (that is, all points
A
B
n is not sample size, but the degrees of freedom of t, that is, n = k 2. on the linear or linearized median S-N or -N curve considered
P is the probability in percent that the random variable t lies in the interval
from tp to +tp. simultaneously) may be computed using the following equa-
tion:

4
 E739 10 (2015)

1 ~ X 2 X ! 2
8.2 Testing the Adequacy of the Linear ModelIn 8.1, it
A 1B X6 =2F p
k
3
1 k

( ~X
i51
i 2 X ! 2
4 (10) was assumed that a linear model is valid, namely that Y ? X
5A1BX. If the test program is planned such that there is more
than one observed value of Y at some of the Xi levels where i
in which Fp is given in Table 2. This table involves two en-
try parameters (the statistical degrees of freedom n1 and n2 3, then a statistical test for linearity can be made based on the
for F). For Eq 9, n1 = 2 and n2 = (k 2). For example, when F distribution, Table 2. The log life of the jth replicate
k = 7, F0.95 = 5.7861. specimen tested in the ith level of X is subsequently denoted
8.1.2.1 A 95 % confidence band computed using Eq 10 is Yij.
plotted in Fig. 1 for the example data of 8.3.1. The interpreta- 8.2.1 Suppose that fatigue tests are conducted at l different
tion of this band is similar to that for a confidence interval levels of X and that mi replicate values of Y are observed at
(8.1.1). Namely, if conditions (a) through (e) are met, and if the each Xi. Then the hypothesis of linearity (that Y | X 5A1BX) is
values of Fp given in Table 2 for, say, P = 95 % are used in a rejected when the computed value of
series of analyses involving the construction of confidence l
2
bands using Eq 10 for the entire range of X used in testing; then
in the long run we may expect 95 % of the computed
( m ~ Y 2 Y ! / ~ l 2 2 !
i51
i i i

l mi (11)
hyperbolic bands to include the straight line Y ? X 5A1BX 2

everywhere along the entire range of X used in testing.

( ( ~Y
i51 j51
ij 2 Y i ! / ~ k 2 l !

NOTE 12Because the actual median S-N or -N relationship is only exceeds Fp, where the value of Fp is read from Table 2 for
approximated by a straight line within a specific interval of stress of strain, the desired significance level. (The significance level is defined
confidence bands which pertain to confidence levels greater than approxi-
mately 0.95 are not recommended. as the probability in percent of incorrectly rejecting the
hypothesis of linearity when there is indeed a linear relation-
8.1.2.2 While the hyperbolic confidence bands generated by ship between X and Y | X .) The total number of specimens
Eq 9 and plotted in Fig. 1 are statistically correct, straight-line tested, k, is computed using
confidence and tolerance bands parallel to the fitted line Y ? X
5A 1B are sometimes used. These bands are described in l

Chapter 5 of Ref (2).

k5 (m
i51
i (12)

TABLE 2 Values of FPA (Abstracted from STP 313 (4))

Degrees of Freedom, n1
1 2 3 4
1 h 161.45 199.50 215.71 224.58
4052.2 4999.5 5403.3 5624.6
2 h 18.513 19.000 19.164 19.247
8.503 99.000 99.166 99.249
3 h 10.128 9.5521 9.2766 9.1172
34.116 30.817 29.457 28.710
4 h 7.7086 6.9443 6.5914 6.3883
21.198 18.000 16.694 15.977
5 h 6.6079 5.7861 5.4095 5.1922
16.258 13.274 12.060 11.392
6 h 5.9874 5.1433 4.7571 4.5337
13.745 10.925 9.7795 9.1483
7 h 5.5914 4.7374 4.3468 4.1203
12.246 9.5466 8.4513 7.8467
8 h 5.3177 4.4590 4.0662 3.8378
Degrees of Freedom, n2 11.259 8.6491 7.5910 7.0060
9 h 5.1174 4.2565 3.8626 3.6331
10.561 8.0215 6.9919 6.4221
10 h 4.9646 4.1028 3.7083 3.4780
10.044 7.5594 6.5523 5.9943
11 h 4.8443 3.9823 3.5874 3.3567
9.6460 7.2057 6.2167 5.6683
12 h 4.7472 3.8853 3.4903 3.2592
9.3302 6.9266 5.9526 5.4119
13 h 4.6672 3.8056 3.4105 3.1791
9.0738 6.7010 5.7394 5.2053
14 h 4.6001 3.7389 3.3439 3.1122
8.8616 6.5149 5.5639 5.0354
4.5431 3.6823 3.2874 3.0556
15 h 8.6831 6.3589 5.4170 4.8932

A
In each row, the top figures are values of F corresponding to P = 95 %, the bottom figures correspond to P = 99 %. Thus, the top figures pertain to the 5 % significance
level, whereas the bottom figures pertain to the 1 % significance level. (The bottom figures are not recommended for use in Eq 10.)

5
 E739 10 (2015)
8.2.2 Table 2 involves two entry parameters (the statistical Also, from Eq 6:
degrees of freedom n1 and n2 for F). For Eq 11, n1 = (l 2), 2 5 0.07837/7 5 0.011195 (14)
and n2 = ( k l). For example, F0.95 = 6.9443 when k = 8 and
l = 4. or,
8.2.3 The F test (Eq 11) compares the variability of average 5 0.1058 (15)
value about the fitted straight line, as measured by their mean
square (Note 14) (the numerator in Eq 11) to the variability 8.3.1.4 Accordingly, using Eq 7, the 95 % confidence inter-
among replicates, as measured by their mean square (the val for A is (tp = 2.3646) [0.6435, 0.1540], and, using Eq 8,
denominator in Eq 11). The latter mean square is independent the 95 % confidence interval for B is [1.6054, 1.2974].
of the form of the model assumed for the S-N or -N 8.3.1.5 The fitted line = log N = 0.24474 1.45144 log
relationship. If the relationship between Y ? X and X is indeed (p/2) = 0.24474 1.45144X is displayed in Fig. 1, where
linear, Eq 11 follows the F distribution with degrees of the 95 % confidence band computed using Eq 10 is also
freedom, (l 2) and (k l ). Otherwise Eq 11 is larger on the plotted. (For example, when p/2 = 0.01, X = 2.000,
average than would be expected by random sampling from this = 2.65814, lower band = 2.65814 0.15215 = 2.50599, and up-
F distribution. Thus the hypothesis of a linear model is rejected per band = 2.65814 + 0.15215 = 2.81029.)

if the observed value of F (Eq 11) exceeds the tabulated value 8.3.1.6 The fitted line can be transformed to the form given
Fp. If the linear model is rejected, it is recommended that a in Appendix X1 of Practice E606/E606M as follows:
nonlinear model be considered, for example: logN 5 20.24474 2 1.45144log~ p /2 ! (16)
2
Y ? X A1BX1CX (13)
log~ p /2 ! 5 20.16862 2 0.68897logN
NOTE 13Some readers may be tempted to use existing digital
computer software which calculates a value of r, the so-called correlation p /2 5 0.67823 ~ N ! 20.68897
coefficient, or r2, the coefficient of determination, to ascertain the
suitability of the linear model. This approach is not recommended. (For Substituting cycles (N) to reversals (2Nf) gives
example, r = 0.993 with F = 3.62 for the example of 8.3.1, whereas
r = 0.988 and F = 21.5 for similar data set generated during the 1976
E09.08 low-cycle fatigue round robin.)
NOTE 14A mean square value is a specific sum of squares divided by
p /2 5 0.67823 S D
2 N f
2
20.68897

(17)

its statistical degrees of freedom.

p /2 5 0.67823 ~ 1/2 ! 20.68897 ~ 2N f ! 20.68897
8.3 Numerical Examples:
20.68897
8.3.1 Example 1: Consider the following low-cycle fatigue p /2 5 1.09340 ~ 2N f !
data (taken from a 1976 E09.08 round-robin test program
(laboratory 43): The above alternative equation is shown on Fig. 1.
p/2 N
8.3.1.7 Ancillary Calculations:
Plastic Strain Amplitude Fatigue Life
Unitless Cycles X 5 22.53172 Y 5 3.42990 (18)
0.01636 168 9
0.01609
0.00675
200
1 000
( ~X 2 X!
i51
i
2
5 2.63892 (19)
0.00682 1 180 9
0.00179
0.00160
4 730
8 035 ( ~ X 2 X ! ~ Y 2 Y ! 5 23.83023
i51
i i (20)
0.00165 5 254

F G
1
0.00053 28 617 1 ~ 22.53172! 2 2
0.00054 32 650 A 5 1 5 0.1686 (21)
9 2.63892
8.3.1.1 Estimate parameters A and B and the respective 1
95 % confidence intervals.
B 5 @ 2.63892# 2 2 5 0.06513 (22)
8.3.1.2 First, restate (transform) the data in terms of loga- 8.3.1.8 Test for linearity at the 5 % significance level.
rithms (base 10 used in this practice due to its wide use in 8.3.1.9 We shall ignore the slight differences among the
practice). amplitudes of plastic strain and assume that l = 4 and k= 9.
Xi = log (pi 2) Y i = log N i Then, at each of the four Xi levels, we shall compute i using
(Independent Variable) (Dependent Variable)
1.78622 2.22531
i = 0.24414 1.45144Xi and Yi using Yi = Y ij /m i .
1.79344 2.30103 Accordingly, F0.95 = 5.79, whereas F computed (using Eq
2.17070 3.00000 11) = 3.62. Hence, we do not reject the linear model in this
2.16622 3.07188
2.74715 3.67486 example.
2.79588 3.90499 8.3.1.10 Ancillary Calculations:
2.78252 3.72049
3.27572 4.45662 Numerator ~ F ! 5 0.0532/2 (23)
3.26761 4.51388
8.3.1.3 Then, from Eq 4 and Eq 5: Denominator ~ F ! 5 0.0368/5

= 0.24474 B = 1.45144 8.3.2 Example 2: Consider the following low-cycle fatigue

Or, as expressed in the form of Eq 2b: data (also taken from a 1976 E09.08 round-robin test program
= 0.24474 1.45144 log (p/2)
logN (laboratory 34)):

6
 E739 10 (2015)

p/2 N 9. Other Statistical Analyses

Plastic Strain Amplitude Fatigue Life
Unitless Cycles 9.1 When the Weibull distribution is assumed to describe
0.0164 153 the distribution of fatigue life at a given stress or strain
0.0164 153
0.0069 563 amplitude, or when the fatigue data include either run-outs or
0.0069 694 suspended tests (or when the variance of log life increases
0.00185 3 515
0.00175 3 860
noticeably as life increases), the appropriate statistical analyses
0.00054 17 500 are more complicated than illustrated in this practice. The
0.00058 20 330 reader is referred to Ref (5) for an example of relevant digital
0.000006 60 350
0.000006 121 500 computer software.
8.3.2.1 The F test (Eq 11) in this case indicates that the NOTE 15It is not good practice either to ignore run-outs or to treat
linear model should be rejected at the 5 % significance level them as if they were failures. Rather, maximum likelihood analyses of the
(that is, F calculated = 39.36, where F3,5,0.95 = 5.41). Hence type illustrated in Ref (5) are recommended.
estimation of A and B for the linear model is not recommended.
Rather, a nonlinear model should be considered in analysis.

REFERENCES

(1) Manual on Statistical Planning and Analysis for Fatigue Experiments, (5) Nelson, W. B., et al., STATPAC SimplifiedA Short Introduction To
STP 588, ASTM International, 1975. How To Run STATPAC, A General Statistical Package for Data
(2) Little, R. E., and Jebe, E. H., Statistical Design of Fatigue Analysis, Technical Information Series Report 73CRD 046, July,
Experiments, Applied Science Publishers, London, 1975. 1973, General Electric Co., Corporate Research and Development,
(3) Brownlee, K. A., Statistical Theory and Methodology in Science and Schenectady, NY.
Engineering, John Wiley and Sons, New York, NY, 2nd Ed. 1965.
(4) ASTM Manual on Fitting Straight Lines, STP 313, ASTM
International, 1962.

ASTM International takes no position respecting the validity of any patent rights asserted in connection with any item mentioned
in this standard. Users of this standard are expressly advised that determination of the validity of any such patent rights, and the risk
of infringement of such rights, are entirely their own responsibility.

This standard is subject to revision at any time by the responsible technical committee and must be reviewed every five years and
if not revised, either reapproved or withdrawn. Your comments are invited either for revision of this standard or for additional standards
and should be addressed to ASTM International Headquarters. Your comments will receive careful consideration at a meeting of the
responsible technical committee, which you may attend. If you feel that your comments have not received a fair hearing you should
make your views known to the ASTM Committee on Standards, at the address shown below.

This standard is copyrighted by ASTM International, 100 Barr Harbor Drive, PO Box C700, West Conshohocken, PA 19428-2959,
United States. Individual reprints (single or multiple copies) of this standard may be obtained by contacting ASTM at the above
address or at 610-832-9585 (phone), 610-832-9555 (fax), or [email protected] (e-mail); or through the ASTM website
(www.astm.org). Permission rights to photocopy the standard may also be secured from the Copyright Clearance Center, 222
Rosewood Drive, Danvers, MA 01923, Tel: (978) 646-2600; http://www.copyright.com/