Endurance Limit

The endurance limit is defined as the stress range below which there is no crack
growth and the material presents an infinite life under cyclic stresses.

From: Surface Modification of Magnesium and its Alloys for Biomedical Applica-
tions, 2015

Related terms:

Welds, Hydrogen, High Cycle Fatigue, Fatigue Limit, S-N Curve, Tensile Strength,
Stress Amplitude

View all Topics

Strengths of materials
J. Carvill, in Mechanical Engineer's Data Handbook, 1993

Non-ferrous metals and alloys

There is no endurance limit and the fatigue stress is taken at a definite value of
stress reversals, e.g. 5 × 107. Some typical values are given.

Endurance limit for some steels

Steel Condition Tensile strength, Endurance limit, e/ u

u (N mm−2) e (Nmm−2)

0.4% carbon Normalized 540 270 0.50

(080M40) Hardened and 700 340 0.49
tempered
Carbon, man- Normalized 540 250 0.46
ganese
(150M19) Hardened and 700 325 0.53
tempered
3% Chrome Hardened and 1000 480 0.48
molybdenum tempered
(709M40)
Spring steel Hardened and 1500 650 0.43
(735A50) tempered
18,8 Stainless Cold rolled 1200 490 0.41
Wrought aluminium alloys

Material Tensile strength, u (N Fatigue stress, FS (N u/ FS

mm−2) mm−2), (5 × 107 cycles)
N3 non-heat-treated 110 48 0.44
130 55 0.42
175 70 0.40
H9 heat treated 155 80 0.52
240 85 0.35

> Read full chapter

Sucker-Rod Pumping System Compo-

nents and Their Operation
Gabor Takacs PhD, in Sucker-Rod Pumping Handbook, 2015

3.5.5.3.1 Steel rods

It is well known that most of the sucker rod breaks are fatigue failures, which occur
at rod stresses well below the ultimate tensile strength or even the yield strength of
the steel used. Material fatigue is a plastic tensile failure, due to repeated stresses,
which starts at some stress raiser (a surface imperfection like a nick or a corrosion
pit) on the surface of the rod, and slowly progresses at right angles to the direction
of the stress, i.e., across the rod material. The load-carrying cross-section is thus
progressively reduced until the remaining metal area is overloaded and breaks. Such
failures occur in most mechanical parts subjected to repeated stresses.

The design of any equipment subjected to cyclic loading must consider the nature
of the loading and the appropriate fatigue endurance limit of the material. The
fatigue endurance limit of a given steel material, in a broad sense, is that maximum
stress level at which the equipment will operate under cyclic loading conditions for
a minimum of 10 million complete cycles. If the steel will withstand this number
of cycles, it can be expected to tolerate these stress levels for an extended period of
time. The endurance limit can only be determined empirically by running a series
of experiments under controlled conditions. Steels can have different endurance
limits depending on the nature of loading (e.g., tension–compression, or in case of
rod strings, pulsating tension); thus the actual value of endurance limit is primarily
controlled by the type of loading. Other important effects include the surface im-
perfections (notches, nicks, etc.) and the nature of the operating environment.

API Rod Materials

Fatigue endurance limits for steels under cyclic tension–compression loads (as in
steel structures) have been long established. The fatigue endurance diagram (called
the Goodman diagram), however, could not be used for designing rod strings
because of the differences in the nature of rod string loading and those used in the
experiments. A. A. Hardy, working with an API study group, proposed the following
modifications [81,82] to the original Goodman diagram:

• the maximum tension stress allowed should be less than the yield strength,
as loading of the steel beyond plastic deformation changes the material prop-
erties,
• compression cannot be allowed in rod strings, as this can cause buckling (to
which slender rods are especially vulnerable) and premature failure, and
• an additional safety factor to account for the corrosiveness of the environ-
ment, generally called service factor, was to be included.

With some minor alterations, this modified Goodman diagram was adopted by the
API [52] and is still in use today. The fatigue endurance limit of the steel rod material,
called the allowable stress in this API publication, can be calculated by the following
formula:

(3.21)

where:

Sa = fatigue endurance limit (allowable stress), psi

SF = service factor, –
Ta = minimum tensile strength of the rod material, psi
Smin = minimum rod stress, psi.

The term SF is called service factor, and its use allows for an additional safety
factor in the string design when corrosive fluids are pumped. Its value is best
established from field records because it varies with the nature of well fluids and the
effectiveness of corrosion treatments. Generally accepted service factors for different
environments and rod grades are listed in Table 3.14, taken from Brown [83].

Table 3.14. Generally Accepted Service Factors, According to Brown [83]

Environment Service Factors

Grade C Grade D
Noncorrosive 1.00 1.00
Salt water 0.65 0.90
Hydrogen sulfide 0.50 0.70

Although the rod body is usually the limiting factor in rod string design, slimhole
couplings considerably reduce the string's strength due to their reduced cross-sec-
tional areas. Therefore, if slimhole couplings are used, the allowable stresses calcu-
lated from Eq. (3.21) must be reduced by applying a derating factor. Values of the
slimhole coupling derating factors for different rod sizes and grades are given in
Table 3.15, based on data from Bradley [48].

Table 3.15. Slimhole Coupling Derating Factors Based on Data from Bradley [48]

Rod Size, in Derating Factors for Rod Grades

K C D
5/8 – 0.97 0.77
3/4 – – 0.86
7/8 0.93 0.88 0.69
1 – – 0.89

The modified Goodman diagram for the API rod grades and the high-strength
E-rods is shown in Fig. 3.69. The allowable stress values, as plotted, were calculated
using Eq. (3.21) for a service factor of 1.0. As discussed in Section 3.5.2.3, the
high-strength E rods can sustain a constant allowable stress. In addition, these
rods do not necessitate the use of service factors to increase the safety of design
in corrosive wells.

Figure 3.69. The modified Goodman diagram for various steel sucker-rod grades.

Figure 3.70 illustrates a nondimensional presentation of the modified Goodman

diagram, which is constructed by plotting the maximum stresses against the min-
imum stresses, both expressed as a percentage of the material's minimum tensile
strength. This diagram can be universally used for all API rod grades and enables
the analysis of different rod grades on the same diagram.
Figure 3.70. A universal modified Goodman diagram.

The safe loading limits on the modified Goodman diagram (see Fig. 3.70) are above
the Smax = 0 line (as below this value the rods are in compression, which was not
included), and below the Sa (allowable stress) line valid for the given service factor.
If maximum rod stress values, plotted against the appropriate minimum stresses,
fall inside these limits, the rod string design is considered a safe one, allowing for
failure-free continuous operation. Therefore, the aim of any string design method
is to keep the stresses in the different taper sections within these safe operating
limits.

It is also apparent from the diagram that, as minimum stress increases, the range
of allowable stresses, i.e., the difference between maximum and minimum stresses,
decreases. This means that when sucker rods operate under high stress levels, the
minimum stresses must be kept fairly high. Fortunately, this is the situation in deep
wells, where the weight of the long rod string constitutes the major part of the
maximum rod load. For the same reasons, in medium-depth wells pumping with
large plungers, resulting in fairly low minimum and high peak loads, the maximum
rod stress may have to be limited in order not to overload the string.

High-Strength Non-API Rod Materials

The manufacturers of Norris 97, LTV HS, and Trico 66 rods claim that their rods
can handle wider ranges of stresses than regular API rods because of their greater
strength. They recommend the following modification to Eq. (3.21) to calculate the
maximum allowable stress in the rod material, where the minimum tensile strength
of these rod materials equals Ta = 140,000 psi:

(3.22)

where:
Sa = fatigue endurance limit (allowable stress), psi
SF = service factor, –
Ta = minimum tensile strength of the rod material, psi
Smin = minimum rod stress, psi.

Norris was the first company to come up with this type of rod (Norris 97) and was also
the first to use this method of stress analysis. The other companies, after developing
rods similar to Norris 97 rods, recommend the use of the same stress range diagram.

Rods with Tenaris Premium Connections

As described in Section 3.5.2.2.2, the revolutionary new thread form developed by

Tenaris, if used on sucker rods made of any conventional material, dramatically
increases the rod string's fatigue life. This improvement is due to the fact that the
majority of rod failures normally occur in the joints. Thanks to the exceptional
strength of the new connection, rods can withstand much higher stress ranges than
the strongest API material. The safe operating range of these rods, therefore, exceeds
that of the rods made of Grade D material. Their maximum allowable stresses are
found from the modified Goodman diagram utilizing the following formula and
Ta = 125,000 psi as the tensile strength:

(3.23)

where:

Sa = fatigue endurance limit (allowable stress), psi

SF = service factor, –
Ta = minimum tensile strength of the rod material, psi
Smin = minimum rod stress, psi.

Figure 3.71 presents a comparison of the modified Goodman diagram for Grade D,

the high-strength materials, and for rods with Tenaris connections. As shown, the
use of the new rod materials and premium connections greatly increases the stress
range allowed in the sucker-rod string and can thus sufficiently raise the lifting
capacity of rod-pumping installations.
Figure 3.71. Comparison of the modified Goodman diagram valid for different rod
materials.

Example 3.8

The Size of the Top Section in a Tapered Sucker-Rod String is 1 in, Rod Material
is Grade D, and the Well Fluid is Salt Water. The Maximum and Minimum Rod
Loads Were Measured as 26,300 lb and 15,700 lb, Respectively. Check Whether the
Rod Section Can Safely Be Operated under These Conditions, if (1) Full-Size and (2)
Slimhole Couplings are Used

Solution
The minimum and maximum rod stresses are calculated with the cross-sectional
area of the 1 in rod found in Table 3.9:

The value of the service factor is found as 0.90 from Table 3.14, and the slimhole
derating factor equals 0.89, taken from Table 3.15.

First, the full-size coupling case is checked, when no derating factor is to be applied
to the allowable stress. The allowable stress is found from Eq. (3.21), where the
minimum tensile strength of Grade D rods is 115,000 psi:

Since this value is higher than the peak stress, the rod section is safe.

In the case of slimhole couplings, the allowable stress is reduced by using the
derating factor:

In this case, the maximum rod stress is higher than the allowed stress and the same
rod section is likely to fail if slimhole couplings are used.
> Read full chapter

Multiaxial Fatigue and Fracture

Thierry Palin-Luc, Serge Lasserre, in European Structural Integrity Society, 1999

UNIAXIAL STRESS STATE

At the endurance limit at the critical point Ci, the quantity a(Ci) is supposed to be
constant. If we note aD (Uniax) its value at the endurance limit for uniaxial stress
state our criterion can be written by inequality (5). Failure occurs if this inequality is
not satisfied.

(5)

To apply this criterion, the following parameters have to be identified: *, Wa*(uniax)

and aD (Uniax). As there is no stress gradient along the longitudinal axis of a smooth
cylindrical specimen under tension, the volume V* can be reduced to the surface S*
inside the specimen cross-section, Eqs (3) and (4) become:

(6)

(7)

The Wa distributions are axisymmetric in tension and in four point rotating bending
(Fig.1b), for this reason these two sinusoidal loadings are taken in reference to
identify *. In tension, all the points of the cross-section of the specimen have the
same Wa value (Fig.2), expression (7) becomes:

Fig.2. Iso-Wa lines and S* surfaces on the cross-section of a specimen loaded in

tension, four points rotating bending or plane bending.

(8)
In four point rotating bending S* is a crown shape, the iso-Wa lines are circular as
shown in Fig.2. For such a loading on smooth cylindrical specimens a is given by (9)
where RotBend is the maximum stress due to rotating bending on the cross-section
(0 ≤ r ≤ R)and * is the radius of the circle representing the iso-Wa* line (Fig.2).

(9)

At the endurance limit, a(Ci) is supposed to be constant whatever the uniaxial

stress state at the critical point Ci: Thus Eq.(10) is obtained from (8) and (9); it is
a convenient expression for design. From (2) and (8) it is easy to prove that at the
endurance limit Wa*(uniax) is given by (11). aD (Uniax) can be calculated by (12).

(10)

(11)

(12)

RotBendD can be considered as a material parameter if the radius of the specimen is

larger than about 5 mm as shown by Pogorotskii and Karpenko (from Papadopoulos
and Panoskaltsis (12)). TracD is not dependent of the size of the specimen, thus *
can be identified as a material parameter. Equations (10) and (11) are available if ;
according to the authors this condition is true on all metallic materials, usually this
ratio is less than 1.3 (12).

> Read full chapter

Fatigue, Creep and Fracture

E.J. Hearn PhD; BSc(Eng) Hons; CEng; FIMechE; FIProdE; FIDiagE, in Mechanics of
Materials 2 (Third Edition), 1997

Example 11.4
The values of the endurance limits at various stress amplitude levels for low-alloy
constructional steel fatigue specimens are given below:

a (MN/m2) Nf (cycles)
550 1 500
510 10 050
480 20 800
450 50 500
410 1 25 000
380 2 75 000

A similar specimen is subjected to the following programme of cycles at the stress

amplitudes stated; 3000 at 510 MN/m2, 12000 at 450 MN/m2 and 80000 at 380
MN/m2, after which the sample remained unbroken. How many additional cycles
would the specimen withstand at 480 MN/m2 prior to failure? Assume zero mean
stress conditions.

Solution
From Miner's Rule, eqn. (11.14), with X the required number of cycles:

> Read full chapter

Bioengineering
S.-H. Teoh, in Comprehensive Structural Integrity, 2003

9.01.5.2 Polymeric Implants

In polymers, the fatigue endurance limits are difficult to determine. First, because
of the low thermal conductivity of polymers, the limits obtained by fatigue experi-
ments are not representative of actual durability of the material, as the mechanism
of failure is associated with localized thermal fatigue. Long-term failure of polymers
is associated with environmental stress cracking and aging mechanisms. Second,
the cost of conducting low-cycle fatigue experiments, especially in a simulated
physiological environment (such as in saline and cholesterol lipid solution for med-
ical plastics), is too high. Failure by a creep-elated fracture process is of primary
importance in engineering polymers. In fact, some researchers have termed such
tests as static fatigue experiments. Previous work (Teoh, 1990, 1993; Teoh et al.,
1992) has shown that the lower stress limits of many polymers can be obtained by
nonlinear computational modeling of the creep rupture time using a three-element
mechanical model having a rate-activated dashpot to simulate plastic flow, in con-
junction with a critical elastic energy criterion. The model equation relating to the
life span (tf) of the material, can be written as

(18)

where C is a constant related to the activation energy, B is a constant related to the

activation volume, H= ap−[2Ea(R− ap2/Ee)]1/2, Ea is anelastic modulus, Ee the elastic
modulus, R the resilience of the material (defined as the maximum elastic stored
energy before failure), and ap is the applied stress. It can be seen that when the
applied stress reaches [Ee, R]1/2, immediate fracture occurs and when it approaches
{R/[1/Ee+1/(2Ea)]}1/2, the material sustains the load indefinitely.

Equation (18) can best be solved by computational nonlinear regression analysis;

it defines two important limits, the upper stress limit (SX) and the lower stress
limit (SN). The SN values at room temperature for a number of polymers such as
polyacetal have been shown to correspond to the endurance stress limits obtained by
a conventional fatigue tester. Figure 22 shows some examples of the fit of Equation
(18) to a number of polymers at 37°C in saline solution. Good fits can be observed
in all cases. Table 5 shows the results from an earlier work (Teoh, 1994) on the
prediction of the lower stress limits of some medical plastics at 20°C. Results for
UHMWPE and polyacetal showed that the lower stress limits were no more than
12 MPa, in saline solution, 37°C. This may well account for the known problems in
wear debris formation and failure of the acetabular UHMWPE cup used in many
hip joint prosthesis, where the contact stresses can exceed 30 MPa. Other new
potential medical plastics, such as polyether ether ketone (PEEK) and polysulfone,
have also been modeled, giving SN values of 75 MPa and 45 MPa, respectively. These
values are much higher than polyethylene and may be more suitable for implant
applications where high bearing stresses are concerned. However, polysulfone,
being amorphous, has been shown to be poor in wear resistance (Teoh et al., 1994).

Figure 22. (a) Creep rupture modeling of some medical plastics and (b) experimental
setup for creep rupture testing in a saline solution environment.

Table 5. Lower stress limits at 20°C of some medical plastics as estimated by

Equation (1).

Plastics Lower stress limit (MPa)

Polyacetal 30
Polycarbonate 47
Polyethylene 8
Polymethylmetacrylate 34
Polypropylene 10
Polysulphone 42
Polyvinyl chloride 35

After Teoh (1994).

Fatigue fracture and wear of polymeric materials used in implants are perhaps the
most difficult to understand and over the years numerous reports have been
produced on the subject. In biomedical applications, such as occluders in mechanical
heart valves and joint prostheses, fatigue fracture and wear of the polymers have
been considered to be an important factor in determining the durability of the
prostheses. Factors influencing the wear properties of UHMWPE, which has been
used in many hip joint prostheses, were examined by Trainor and Haward (1980).
Their results indicated that a significant improvement in the behavior of wear (using
a pin on plate or a rotating shaft on plate system in a medium of distilled water)
was obtained by molding the UHMWPE between 190°C and 200°C. The addition of
some antioxidants also appeared to improve the wear resistance. Molding at higher
pressures and increasing the molecular weight were reported to be detrimental.
Nonetheless, there is a possibility that there could be an optimum processing condi-
tion and molecular weight distribution that could give the best wear characteristics.
More recent work (Pruitt and Bailey, 1998) has shown that processing conditions play
a vital role on the cyclic fatigue of UHMWPE. In particular -radiation and oxidative
aging are very detrimental to the fatigue threshold and crack propagation resistance
(Table 6). Compression molding appears to give a better fatigue resistance when
compared to extrusion.

Table 6. Effect of processing condition on the fatigue threshold (ΔKth) of UHMWPE.

Condition ΔKth

Compression molded 1.8

Compression molded -air 1.2

Extruded 90° 1.7

Extr0uded 0° nonsterilized 1.3
Extruded 0° -air 1.0

Extruded 0° -peroxide 1.1

After Pruitt and Bailey (1998).

Extensive work has been carried out to study the wear and degradation of re-
trieved polymeric implants (Gibbons et al., 1979). This interesting piece of work
examined 30 implants ranging from UHMWPE to silicone occluders. It reported
wear mechanisms related to abrasive wear and environmental stress cracking of the
incompletely sintered UHMWPE powder. For polymeric valve occluders, abrasive
wear was predominant. Such conclusions were also reported for polyacetal (Teoh
et al., 1990). In an examination of an explanted valve (Björk–Shiley polyacetal disk
mechanical heart valve) that had been in a patient for more than 17 years, abrasive
and static wear marks, arising from plastic deformation and surface material flow
and polymer debris adhesion on the metallic struts, were observed. Clarke and
McKellop (1980) compared the wear of polyacetal with UHMWPE, polyester, and
teflon (PTFE). A pin (polymer)-on-disk (316 stainless steel) in bovine serum solution
was used. Their results indicated polyacetal, polyester, and PTFE wear 60, 2,576,
and 4,986 times more than UHMWPE, respectively. In some studies on total knee
joint prostheses, it has been shown that UHMWPE debris can contribute to implant
loosening (St. John, 1992; Engh et al., 1992). The UHMWPE debris can migrate
down from the bulk component to the bone–cement or bone–implant interface and
provoke a host response resulting in bone resorption.

It has been noted that many accelerated fatigue testers could not reproduce the in
vivo performance, and caution needs to be exercised when interpreting the results
especially for heart valves. In the case of mechanical heart valves with a single
tilting disk design, Teoh et al. (1994) have noted that the in vivo loading consists
of impact-cum-sliding action and proposed a new impact-cum-sliding accelerated
wear test to evaluate polyacetal, UHMWPE, polysufone, and PEEK as materials for
the occluder. This is a much simpler and cost-effective method. Their results are
summarized in Figures 23 and 24. On the basis of wear depth and debris morphol-
ogy, polyacetal was concluded to be better than UHMWPE because the debris size
of UHMWPE was large (>100 μm) compared to that of polyacetal ( 30 μm), even
though polyacetal was ranked second after UHMWPE in wear depth penetration. The
large debris morphology was deemed unacceptable in cardiovascular applications.
To ascertain the stress magnitude at the stress concentration areas, Teoh et al. (1993)
also carried out in vitro strain measurements on a St. Vincent's mechanical heart
valve in a pulse simulator. The results were combined with a finite element (FEM)
stress analysis of the titanium valve housing. The imposed stress on the occluder
by the upper strut was less than 2 MPa. This is below the lower stress limit of
polyacetal (5 MPa) (Teoh, 1993) and may explain why no fracture of the polyacetal
disk occluder has been reported. (It needs to be emphasized that the mechanical
polishing of the polyacetal occluder introduces compressive surfaces stresses, which
further enhances the fatigue and wear resistance of the occluder.) This may explain
why polyacetal used in the artificial hip joint prostheses (Dumbleton, 1979) was
known to wear severely and fail by fatigue where the contact stresses exceed this
limit.
Figure 23. Wear depth of polyacetal, UHMWPE, polysuphone, and PEEK subject-
ed under an impact-cum-sliding action (after Teoh et al., 1994. Reprinted, with
permission, from STP 1173–Biomaterials Mechanical Properties, copyright ASTM
International, 100 Barr Harbor Drive, West Conshohocken, PA, 19428, USA.).

Figure 24. Scanning electron micrographs of: (a) polysuphone, (b) PEEK, (c) polyac-
etal, and (d) UHMWPE subjected under an impact-cum-sliding action (after Teoh
et al., 1994. Reprinted, with permission, from STP 1173–Biomaterials Mechani-
cal Properties, copyright ASTM International, 100 Barr Harbor Drive, West Con-
shohocken, PA, 19428, USA.).

> Read full chapter

Multiaxial Fatigue and Fracture

Franck MOREL, ... André BIGNONNET, in European Structural Integrity Society,
1999

Fatigue life prediction steps (Fig.7)

Required material fatigue characteristics are two endurance limits, an S-N curve
and a particular two constant amplitude blocks loading test. The first step of the
fatigue life prediction procedure is once again the location of the critical material
plane through a maximisation of the parameter T rrns. After TΣrms Prms, TΣlim, lim and
computations, damage accumulation can be estimated from the evolution of the
macroscopic resolved shear stress T(t) on a direction of the critical plane. The number
of sequences to crack initiation is deduced from a calculation on the direction
leading to the highest accumulated plastic mesostrain.

Fig.7. Diagram of algorithm for evaluation of fatigue life of metals under(uniaxial or

multiaxial) variable amplitude loading

> Read full chapter

Fatigue of aerospace materials

In Introduction to Aerospace Materials, 2012

20.6.3 Mechanical properties of fatigued composites

When the fatigue stress exceeds the endurance limit, the mechanical properties of
composites degrade over their fatigue life owing to the initiation and spread of
damage. Figure 20.18 shows the reductions to the Young’s modulus and tensile
strength that can occur over the fatigue life. The stiffness decreases early in life,
but usually by only a few percent, and then remains reasonably steady for most of
the fatigue life. Towards the end of life, the stiffness drops sharply owing to the
failure of fibres and, within a short time, the material fails completely. The change in
strength follows a different trend to stiffness. The strength remains unchanged for
a large period of the fatigue life, and it is only when fibres begin to break well into
the life-span of the composite that the strength begins to fall. The strength drops as
more and more fibres are broken under cyclic loading until eventually the composite
is completely broken.

20.18. Reduction in stiffness and strength of carbon–epoxy with increasing number

of tensile load cycles.

> Read full chapter

Fundamental Theories and Mecha-

nisms of Failure
S.C. Tjong, Y.-W. Mai, in Comprehensive Structural Integrity, 2007

2.11.5.2 High-Cycle Fatigue

MMCs generally have superior room-temperature endurance limits compared to
unreinforced metals owing to their higher mechanical strengths (Figure 55). Several
intrinsic factors such as particle size and volume fraction, matrix microstructure, and
particle–matrix bonding could affect the high-cycle fatigue (HCF) performance of
particulate-reinforced MMCs. In addition, extrinsic test parameters such as stress
ratio or mean stress, temperature, and environment have large influences on the
HCF life of MMCs (Tjong et al., 2005; Bonnen et al., 1991; Nieh et al., 1995;
Srivatsan and Al-Hajri, 2002). Fatigue cracks can be initiated in various defects in
MMCs depending upon the composite processing route. These defects could be
large intermetallic inclusions, large ceramic reinforcing particulates, agglomeration
of ceramic particles, porosity, etc. (Bonnen et al., 1991; Nieh et al., 1995; Srivatsan
and Al-Hajri, 2002).

Figure 55. Fatigue S–N curves for pure Ti and Ti/TiBw composite (Tjong and Wang,
2005). Reproduced from Tjong, S. C. and Wang, G. S. 2005. Cyclic deformation
behavior of titanium-matrix composite reinforced with in situ TiB whiskers. Adv. Eng.
Mater. 7, 63–68.

Figure 56 shows the S–N curves for 6061Al/SiC-T6 composites with different volume
fractions of reinforcing phase and unreinforced 6061-T6 under stress-controlled
fatigue at ambient temperature (Srivatsan et al., 2002). Apparently, the incorporation
of SiC particulates improves the fatigue strength, taken as the highest stress at
which the specimen endures 106 cycles. Moreover, the fatigue endurance limit tends
to increase with increasing particulate content. This behavior is expected since the
ceramic particulates can carry more applied stress during cyclic deformation with
increasing reinforcement volume content. Chawla et al. (1998) and Chawla and Shen
(2001) have studied the effect of aging on the HCF behavior of 2080Al/SiCp(5 μm)
composites. The fatigue endurance limit subjected to thermomechanical treatment
(T8) is 230 MPa at 107 cycles. A T8 treatment produces homogeneous fine dis-
tribution of S precipitates in the matrix of the composite. The fatigue strength
of 2080Al/SiCp-T8 composites is much higher than unreinforced 2080 alloy, and
increases with increasing particle content (Chawla et al., 1998). But the fatigue limit is
dramatically reduced to 150 MPa for the composite treated at T8 and then overaged
at 250 °C for 24 h. This is because overaging leads to precipitate coarsening and
increase in precipitate spacing (Chawla and Shen, 2001).
Figure 56. Maximum stress versus fatigue life for T6 treated unreinforced 6061Al
alloy and 6061Al/SiCp composites reinforced with 10% and 15% SiCp under
stress-controlled condition (Srivatsan et al., 2002).

Figure 57 shows typical S–N curves for 6090Al/25vol.%SiC-T6 composite at room

temperature and 300 °C. The fatigue strength at 20 and 300 °C for 107 cycles are
200 and 63 MPa, respectively. The large loss of fatigue strength at 300 °C is due
to overaging of the precipitates in T6 matrix (Nieh et al., 1995). It is noted that
6090Al/25vol.%SiC-T6 composites exhibit higher fatigue strength/tensile strength
ratio at 300 °C (63/126 MPa; 0.5) compared to that at 20 °C (200/541 MPa; 0.37). They
attributed this to the higher tensile ductility (23%) and lower tensile strength of the
composite at 300 °C. This means the composite can accommodate larger amounts
of deformation strains at higher temperatures. Moreover, fatigue damage is initiated
in the composite matrix at 300 °C rather than particle cracking that commonly takes
place in MMCs at room temperature. Thus, the fatigue failure mechanisms in MMCs
under HCF change from particle fracture at room temperature to plastic deformation
of matrix at high temperature.

Figure 57. Fatigue S–N curves for 6090Al/25vol.%SiC-T6 composite at 20 and 300 °C
(Nieh et al., 1995).

It should be noted that large ceramic reinforcing particles tend to act as stress
concentrators during cycling, leading to inferior fatigue performance of MMCs
at room temperature (Srivatsan et al., 2002). In this regard, in situ MMCs rein-
forced with fine ceramic particles are considered to exhibit better HCF performance
than ex situ composites. Vyletel et al. (1995a) conducted a preliminary study on
HCF behavior of 2219Al/15vol.%TiC alloy reinforced with in situ TiC particles. They
demonstrated that the TiC particles appeared to have no significant effect on
the stress-controlled fatigue life of the in situ composites. The fatigue strength of
naturally aged (T4) composite at 107 cycles is 150 MPa. However, the overaged MMC
failed at 105 cycles and the fatigue strength could not be determined, since fatigue
endurance limit is generally estimated at ≥ 107 cycles. The absence of improvements
in fatigue life of in situ 2219Al/15vol.%TiC composite is attributed to the high
Al3Ti intermetallic content and high density of TiC clusters leading to preferential
fatigue-crack initiation at these defects (Vyletel et al., 1995a). Hence, formation of the
intermetallic Al3Ti compound is detrimental to the fatigue performance of the in situ
Al-based composites. Tjong and Wang also found that intermetallic Al3Ti compound
degrades the HCF resistance of in situ Al-based composites (Tjong and Wang, 2004).
Figure 58 shows the fatigue S–N curves for in situ Al/20vol.%(TiB2 + Al2O3) and
Al/21.6vol.%(TiB2 + Al2O3 + Al3Ti) composites. Apparently, as expected, elimination
of the Al3Ti phase leads to higher fatigue endurance limit.

Figure 58. Fatigue S–N curves for pure Al and in situ Al-based composites. Composite
1 (Al/20 vol.%(TiB2 + Al2O3)) and composite 2 (Al/21.6vol.%(TiB2 + Al2O3 + Al3Ti))
were prepared from Al–TiO2–B system with B/TiO2 molecular weight ratio of 2 and
5/3, respectively. Composite 3 was prepared from Al–TiO2–B2O3 system (Tjong and
Wang, 2004).

There is little information regarding the fatigue performance of submicron and

nanocomposites. Chen et al. have reported that the fatigue strength and fatigue life
of ingot-cast 6061Al/10vol.%Al2O3 composite are improved significantly via ECAP
treatment. The improvement in the HCF performance is due to the grain refinement
introduced by ECAP (Chen et al., 2005). In sharp contrast, the tensile strength and
fatigue strength of ECAP 6061Al alloy are lower than 6061 alloy without ECAP (Figure
59). This is because 6061Al alloy is subjected to T6 treatment (solution heat-treated
at 530 °C and aging at 175 °C for 8 h) to induce Mg2Si precipitates prior to ECAP.
The ECAP of 6061Al alloy is conducted at 260 °C and this is much higher than
the aging temperature of 175 °C in T6 treatment. The loss of tensile and fatigue
strengths of 6061Al alloy after ECAP is attributed to the coarsening of precipitates.
The composite was processed by ECAP without any prior heat treatment. It is noticed
that as-received 6061Al/10vol.%Al2O3 composite exhibits much lower yield stress
( 130 MPa) compared to that of as-received 6061Al alloy subject to T6 treatment
( 260 MPa). Accordingly, as-received 6061Al alloy exhibits much larger fatigue
strength than as-received 6061Al/10vol.%Al2O3 composite. Because the composite
was not subject to T6 treatment prior to ECAP as in the case of as-received 6061Al
alloy, it is difficult to compare directly the differences in the tensile and fatigue
behavior between them before and after ECAP treatments.

Figure 59. Fatigue S–N curves for 6061Al alloy and its composite before and after
ECAP (Chen et al., 2005).

> Read full chapter

Nominal stress approach for welded

joints
D. Radaj, ... W. Fricke, in Fatigue Assessment of Welded Joints by Local Approaches
(Second Edition), 2006
2.2.8 Fatigue strength reduction factors
Fatigue strength reduction factors characterise the (technical) endurance limit of
typical uniaxially loaded welded joints in relation to the endurance limit of the
non-welded parent material, Fig. 2.13 and Table 2.1. They were widely used before
the fatigue-relevant codes and guidelines for welded joints were established. They
continue to be a useful means for presenting a well-founded first estimate of
endurable stress amplitudes in welded joints, especially in those engineering areas
which are not strictly bound to code regulations. They are particularly useful for
predicting tendencies and for assessing preliminary designs.

Fig. 2.13. Reduction factor of various types of welded joints made of structural
steel in the as-welded condition ( Y ≈ 250 N/mm2, U ≈ 400 N/mm2); endurance limit
in terms of nominal stress amplitude related to parent material endurance limit (NE
= 2 × 106 cycles, R = − 1); after Stüssi.70

Table 2.1. Survey of reduction factors of defect-free welded joints made of low-car-
bon structural steels ( Y ≈ 250 N/mm2, U ≈ 400 N/mm2); ΣE = 240 N/mm2 at NE =
2 × 106 cycles in mill-finished plate, R = 0, Pf = 10%; after Radaj54 (based on data in
the open literature)

Weld type Reduction factor

Load-carrying welds
Butt weld 0.5–0.9 0.7–0.9 0.6–0.9
Single and double bevel 0.4–0.7 0.6–0.8 0.5–0.7
butt weld
Fillet weld 0.3–0.5a 0.5–0.7 0.4–0.6
0.3–0.5b
0.2–0.4c
Corner weld 0.3–0.5 0.5–0.7 0.4–0.6
Keyhole weld in lap joint 0.2–0.5 0.5–0.7 0.3–0.6
Resistance spot weld 0.1–0.5 0.4–0.5 0.2–0.5
Non-load-carrying
welds
Butt-welded longitudi- 0.2–0.3
nal gusset plate
Fillet-welded attach- 0.4–0.8d 0.4–0.7e 0.3–0.6
ments
Bead-on-plate weld 0.6–0.9 0.6–0.9 0.6–0.9

a Cruciform or lap joint with transverse fillet weld.

b Discontinuous web-to-flange fillet weld.

c Lap joint with side fillet welds.

d Transverse attachment.

e Longitudinal attachment.

Reduction factors , || and are defined for cyclic loading of the weld by nominal
stress amplitudes in the parent material directed perpendicular and parallel to
the weld (normal and shear stresses a, ||a and ||a), evaluating their (technical)
endurance limits A, ||A and ||A in relation to the endurance limit of the non-welded
parent material E or E:

(2.25)

(2.26)

(2.27)

The fatigue limit E or E of the parent material is usually determined by testing

polished specimens (such reference values are used in Fig. 2.13), but another surface
condition may also be chosen, e.g. the mill-finished condition (as in Table 2.1). A
dash is positioned on top of the symbol in the latter case. Also R = − 1 may be
considered instead of R = 0.

Supposing that the fatigue strength of welded joints is independent of the static ten-
sile strength of the material (structural steel or aluminium alloy) – which is a realistic
assumption with regard to severely notched welded joints with correspondingly low
reduction factors – lower reduction factors are derived for high strength materials.
This means that the reduction factors are dependent on the material strength to
some extent.

The reduction factors may also be modified by further influence parameters. Ap-
proximation formulae for endurable nominal stress amplitudes were derived based
on reduction factor considerations which are intended to catch these additional
influences (Radaj54). The reduction factors presented above are valid in the high-cycle
fatigue range. They rise in the medium-cycle fatigue range (or with variable-ampli-
tude loading), especially so if they are originally low. Note that the fatigue strength
of the parent material used for reference also rises. Further important influence
parameters are hardening or softening in the critical cross-section, residual stresses
produced by welding and multiaxiality of the basic stress state. The stress ratio R of
cyclic loading, on the other hand, is considered to be only of minor influence.

> Read full chapter

Mechanics of Materials
DAN B. MARGHITU, ... BOGDAN O. CIOCIRLAN, in Mechanical Engineer's Hand-
book, 2001

3.1 Endurance Limit

The strength of materials acted upon by fatigue loads can be determined by per-
forming a fatigue test provided by R. R. Moore's high-speed rotating beam machine.
During the test, the specimen is subjected to pure bending by using weights
and rotated with constant velocity. For a particular magnitude of the weights, one
records the number of revolutions at which the specimen fails. Then, a second
test is performed for a specimen identical with the first one, but the magnitude
of the weight is reduced. Again, the number of revolutions at which the fatigue
failure occurs is recorded. The process is repeated several times. Finally, the fatigue
strengths considered for each test are plotted against the corresponding number of
revolutions. The resulting chart is called the S–N diagram.

Numerous tests have established that the ferrous materials have an endurance limit
defined as the highest level of alternating stress that can be withstood indefinitely
by a test specimen without failure. The symbol for endurance limit is . The endurance
limit can be related to the tensile strength through some relationships. For example,
for steel, Mischke1 predicted the following relationships

(3.1)

where Sut is the minimum tensile strength. Table 3.1 lists the values of the endurance
limit for various classes of cast iron. The symbol refers to the endurance limit of
the test specimen that can be significantly different from the endurance limit' Se
of any machine element subjected to any kind of loads. The endurance limit Se can
be affected by several factors called modifying factors. Some of these factors are the
surface factor ka, the size factor kb, or the load factor kc. Thus, the endurance limit of
a member can be related to the endurance limit of the test specimen by
Table 3.1. Typical Properties of Gray Cast Iron

Modulus
of
elasticity
(Mpsi)
ASTM Tensile Compres- Shear tension torsion En- Brinell Fatigue
number strength sive modu- durance hard- stress
Sut (kpsi) strength lus of limit Se ness HB concentra-
Suc(kpsi) rupture tion
Ssu(kpsi) kpsi) factor Kf

20 22 83 26 9.6–14 3.9–5.6 10 156 1.00

25 26 97 32 11.5–14.8 4.6–6.0 11.5 174 1.05
30 31 109 40 13–16.4 5.2–6.6 14 201 1.10
35 36.5 124 48.5 14.5–17.2 5.8–6.9 16 212 1.15
40 42.5 140 57 16–20 6.4–7.8 18.5 235 1.25
50 52.5 164 73 18.8–22.8 7.2–8.0 21.5 262 1.35
50 62.5 187.5 88.5 20.4–23.5 7.8–8.5 24.5 302 1.50

Source: Joseph E. Shigley and Charles R. Mischke, Mechanical Engineering Design, 5th
ed., p. 123. McGraw-Hill, New York, 1989. Used with permission.

(3.2)

Some values of the foregoing factors for bending, axial loading, and torsion are listed
in Table 3.2.

Table 3.2. Generalized Fatigue Strength Factors for Ductile Materials

Bending Axial Torsion

a. Endurance limit S-
e = kakbkcS e, where
S e is the specimen en-
durance limit
kc (load factor) kb (gradi- 1 1 0.58
ent factor)
diameter < (0.4 in or 1 0.7–0.9 1
10 mm)
(0.4 in or 10 mm) < 0.9 0.7–0.9 0.9
diameter < (2 in or 50
mm)
ka (surface factor) See Fig. 3.5
b. 103-cycle strength 0.9Su 0.5Su 0.9Sua

a Sus≈ 0.8Su for steel; Sus ≈ 0.7Su for other ductile materials.

Source: R. C. Juvinall and K. M. Marshek, Fundamentals of Machine Component Design.

John Wiley & Sons, New York, 1991. Used with permission.
3.1.1 SURFACE FACTOR ka
The influence of the surface of the specimen is described by the modification factor
ka, which depends upon the quality of the finishing. The following formula describes
the surface factor:

(3.3)

Sut is the tensile strength. Some values for a and b are listed in Table 3.3.

Table 3.3. Surface Finish Factor

Factor a

Surface finish kpsi MPa Exponent b

Ground 1.34 1.58 –0.085

Machined or 2.70 4.51 –0.256
cold-drawn
Hot-rolled 14.4 57.7 –0.718
As forged 39.9 272.0 –0.995

Source: J. E. Shigley and C. R. Mischke, Mechanical Engineering Design. McGraw-Hill,

New York, 1989. Used with permission.

3.1.2 SIZE FACTOR kb

The results of the tests performed to evaluate the size factor in the case of bending
and torsion loading of a bar, for example, can be expressed as

(3.4)

where d is the diameter of the test bar. For larger sizes, the size factor varies from
0.06 to 0.075. The tests also revealed that there is no size effect for axial loading;
thus, kb= 1.

To apply Eq. (3.4) for a nonrotating round bar in bending or for a noncircular cross
section, we need to define the effective dimension de. This dimension is obtained by
considering the volume of material stressed at and above 95% of the maximum
stress and a similar volume in the rotating beam specimen. When these two volumes
are equated, the lengths cancel and only the areas have to be considered. For
example, if we consider a rotating round section (Fig. 3.1a) or a rotating hollow
round, the 95% stress area is a ring having the outside diameter d and the inside
diameter 0.95d. Hence, the 95% stress area is
Figure 3.1. Beam cross-sections. (a) Solid round; (b) rectangular section; (c) channel
section; (d) web section. Used with permission from Ref. 16.

(3.5)

If the solid or hollow rounds do not rotate, the 95% stress area is twice the area
outside two parallel chords having a spacing of 0.95D, where D is the diameter.
Therefore, the 95% stress area in this case is

(3.6)

Setting Eq. (3.5) equal to Eq. (3.6) and solving for d, we obtain the effective diameter
 (3.7)

which is the effective size of the round corresponding to a nonrotating solid or

hollow round.

A rectangular section shown in Fig. 3.1b has A0.95 = 0.05hb and

(3.8)

For a channel section,

(3.9)

where a, b, x, tf are the dimensions of the channel section as depicted in Fig. 3.1c.

The 95% area for an I-beam section is (Fig. 3.1d)

(3.10)

3.1.3 LOAD FACTOR kc

Tests revealed that the load factor has the following values:

(3.11)

> Read full chapter

ScienceDirect is Elsevier’s leading information solution for researchers.

Copyright © 2018 Elsevier B.V. or its licensors or contributors. ScienceDirect ® is a registered trademark of Elsevier B.V. Terms and conditions apply.