Received: 28 February 2018 Revised: 16 May 2018 Accepted: 23 May 2018

DOI: 10.1111/ffe.12867

SPECIAL ISSUE ‐ MANUFACTURING INFLUENCE ON

FATIGUE PROPERTIES

A comparison of rolling contact fatigue behaviour of

17NiCrMo6‐4 case‐hardened disc specimens and gears

A. Terrin1,2 | G. Meneghetti1

1
University of Padova, Department of
Industrial Engineering, Via Venezia, 1,
Abstract
35131 Padova, Italy A two‐disc test rig employed to assess the contact fatigue performance of gear
2
Engineering and Innovation materials is presented. Tests were performed on both shot‐peened and un‐
Department, Carraro S.p.A., Via Olmo 37,
peened 17NiCrMo6‐4 discs, designed to recreate the meshing conditions
35010 Campodarsego, PD, Italy
between the sun and planet gears of planetary gear sets in the wheel hubs of
Correspondence off‐highway vehicles. The mechanisms leading to pitting damage in disc spec-
A. Terrin, Department of Industrial
imens are discussed. Disc specimens with the same slide‐to‐roll ratio as that of
Engineering, University of Padova, Via
Venezia, 1, 35131 Padova, Italy. the sun gears exhibited similar crater morphology but had 44% higher capacity
Email: [email protected] in terms of contact pressure. This difference was attributed to changes in the
disc geometry due to wear and micropitting during the tests, which enhanced
Funding information
Carraro S.p.A. the conformity between the contact surfaces and thus reduced the contact pres-
sure for a given applied load.

KEYWORDS
contact fatigue, gears, micropitting, pitting, shot peening, slide‐to‐roll ratio

1 | INTRODUCTION gears with poor lubricating conditions.4,5 This process

produces craters that are typically 10 to 20 μm deep and
Pitting is a contact fatigue phenomenon resulting from a few tens of microns wide and are only visible to the
the repetitive application of high contact stresses on the naked eye when several micropits coalesce to form a matt
tooth surface, which results in the formation of craters surface.6 By contrast, macropitting is characterized by
that lead to noise and vibration and eventually act as craters ranging from 0.5 to 1 mm in size.
structural failure initiation points. Pitting can occur in When isolated macropits form, they assume a seashell
case‐hardened gears in two forms, micropitting and shape with two nearly straight sides meeting at the origin
macropitting (Figure 1). The damaging mechanism that at angles of 70° to 140° (Murakami et al7). With further
leads to the formation of craters initiates with the plastic operation, macropitting usually spreads rapidly over a
deformation of the tooth surface and the flattening of the wide area. Although micropitting is much less detrimen-
asperities in the direction of the friction force (surface dis- tal for the normal functioning of gears than macropitting,
tress). This phase is accompanied by mild wear, which macropitting can originate from the craters produced by
produces a mirrorlike surface, and by the formation of micropitting. Moreover, the deterioration of the surface
shallow cracks characterized by their direction to the sur- finish due to the presence of microcraters enhances the
face, generally at angles below 30°1-3 (Figure 2A,B). The wear rate causing substantial modifications of the surface
cracks may eventually propagate until they turn or topography, which may in turn affect the contact
branch towards the surface and debris material is pressure.
detached, leaving a crater on the tooth flank (Figure 2 The design against macropitting is based on Hertzian
B). Micropitting is particularly common in carburized contact stresses, whose limit values should be derived

Fatigue Fract Eng Mater Struct. 2018;1–17. wileyonlinelibrary.com/journal/ffe © 2018 Wiley Publishing Ltd. 1
2 TERRIN AND MENEGHETTI

FIGURE 1 Macropit generated at the pitch line of the sun gear of a planetary gear set in an off‐highway axle. The contact stress
concentration at the upper edge of the micropitted area caused the macrocrater to form [Colour figure can be viewed at
wileyonlinelibrary.com]

FIGURE 2 A, Surface cracks in the addendum of a gear tooth. B, Surface cracks and micropit in the dedendum of a gear tooth. C, Top view of
a macropit on a sun gear tooth flank observed with a scanning electron microscope

experimentally by tests on reference gears under refer- fatigue performances of different materials relative to
ence working conditions.8 However, the high costs and one another is universally accepted, many
time consumption related to the execution of test cam- authors9,12,14,16-18 claim that disc specimens are more
paigns on gears have encouraged the use of simpler and durable than gears. Nevertheless, even when the transfer
faster tests on rolling pairs of discs in loaded contact. Disc of the absolute values of endurable contact stresses from
tests have been widely used to assess the contact fatigue disc tests to actual gears was questionable, good relative
behaviours of different materials9-12 and the influences correlations were proven in several works.14,15,17
of several parameters that affect contact fatigue strength, Notably, the pitting durability of gears for the with-
such as case carburizing depth,13 residual stress,14 slide‐ drawn standard BS43619 was based on the results of disc
to‐roll ratio (SRR),9,11,15 and lubrication condition.16 tests, and the currently used standard ISO 6336‐1:20068
Although the use of disc tests to compare the surface still mentions the possibility of using disc tests to
TERRIN AND MENEGHETTI 3

determine permissible contact stress values for gear mate- 2 | T H E 2‐ DI SC T E ST RI G

rials, although it informs the designer about the higher
load carrying capacity of discs with respect to gears. The specimens are mounted on a pair of spindles driven
In this work, sun gears in off‐highway axles are by an electric motor, as shown in Figure 3.
treated, which typically fail by pitting because of the A gear box with a unitary gear ratio connects the
high applied torque despite their small dimensions, ie, driver and the driven shaft. The load is applied to the
small curvature radii of the flanks resulting in high con- shafts by means of a screw and a spring mounted against
tact pressure. Moreover, although the frequency of load a load cell. Because the two spindles rotate with the same
cycles in sun gears is high due to the concurrent angular speed ω (in the range 840 − 2000 rpm), the com-
meshing with 3 or 4 planet gears, the speed of rotation bined rolling and sliding motion of the gear teeth is
is relatively low; therefore, the capability of the teeth obtained by using two discs with different diameters.
surfaces to drag the oil film in the contact area is limited Figure 4 shows the peripheral speeds of the discs,
and results in a lubricant film that cannot separate the which depend on their radii Rx1 and Rx2. The ratio
mating surfaces. A two‐disc test rig is presented along between the difference of the peripheral speeds (sliding
with the adopted specimen design criteria to closely velocity, vS) and their arithmetic mean (rolling velocity,
match the operating conditions of the sun gears in a v1 þ v2
medium‐power axle for agricultural applications. The vR ¼ ) is defined as the SRR and represents a major
2
results of tests on both shot‐peened and un‐peened disc factor that influences the development of rolling contact
specimens are compared with gear failures in terms of fatigue phenomena. The versus of the rolling speed vR is
durability, damaging mechanisms, and morphology. given by the direction of motion of the point of contact
Finally, the differences between the durability of discs over the mating surfaces and is reported in Figure 4,
and gears are tentatively explained considering the mod- along with the directions of the friction forces F f 1 and
ifications in the geometry of the contact surfaces due to F f 2 acting on disc 1 and 2, respectively. A negative sign
wear and micropitting. is attributed to the SRR of the smaller disc, where F f 1

FIGURE 3 Layout of the test rig

[Colour figure can be viewed at
wileyonlinelibrary.com]

FIGURE 4 Rolling and sliding speeds

on the discs surfaces [Colour figure can be
viewed at wileyonlinelibrary.com]
4 TERRIN AND MENEGHETTI

and vR have opposite directions. Conversely, the larger and maintain the required working temperature. During
disc is characterized by a positive SRR. periodic pauses of the tests, the specimens are automati-
Pitting is more prone to occur on surfaces subjected to cally scanned by a dedicated vision system. The discs
negative SRR values, where friction forces exercised by perform a complete revolution in 14 small steps while a
the surface of the mating body always act in the opposite digital microscope grabs pictures of their surfaces. The
direction with respect to the motion of the point of con- images are processed to identify the possible presence of
tact (namely, the rolling direction). This process is attrib- pitting craters and to assess whether the failure criterion
uted to the action of the lubricant, which is more prone to has been achieved. Further details about the test rig and
leak in cracks formed in surfaces with negative SRR, the image processing algorithm can be found in
where the material ahead of the contact point is stretched Meneghetti et al.25
and forces the mouth of the crack to open. During the
subsequent contact, the pressure of the lubricant pro-
motes crack propagation.20-23 The opposite process takes
3 | S P E C I M E N S DE S I G N AN D T E S T
place on the surfaces characterized by positive SRR
PARAMETERS
values, where friction forces drag the surface, keeping
the cracks ahead of the contact point closed and
The existing literature about two‐disc tests shows how the
preventing oil entrapment and pressurization. Hence, in
pitting durability of case‐hardened steel specimens can
most cases, the smaller disc manifests pitting before the
vary depending on the specimen features and the test
larger disc.
parameters. Particularly, contact fatigue phenomena can
Because the centre distance of the discs is fixed
be affected by the curvature radii of the surfaces, sliding
(Rx1 + Rx2 = 70 mm), the difference between the disc
conditions, lubricant viscosity and temperature, rough-
radii required to achieve the desired SRR can be deter-
ness, case hardening depth, and the residual stresses
mined by considering the following equation:
induced by the heat treatment or by shot peening. In
Figure 5, experimental data and fatigue curves from vari-
vS ωðRx2 −Rx1 Þ 2ðRx2 −Rx1 Þ ous publications (see Table 1 for references) are com-
SRR ¼ ¼ ¼ (1)
vR 1=2ωðRx1 þ Rx2 Þ ðRx1 þ Rx2 Þ pared with the pitting life of shot‐peened and un‐peened
case‐hardened 17NiCrMo6–4 sun gears in the planetary
where ω is the angular speed of the spindles. gear sets of off‐highway axles published in previous
In the test rig, specimens are lubricated by a continu- works.28-30 The pressure values reported in the chart were
ous stream through a nozzle located over the contact calculated by tooth contact analysis of the meshing
area. The oil temperature is an important test parameter between the sun and planet gears. In contrast with previ-
because it strongly affects the viscosity of the lubricant24; ous publications,29,30 the data were not corrected with the
therefore, a heater is provided for the bench to achieve factors provided by ISO standard 6336 to account for the

FIGURE 5 Survey of the literature data on disc tests compared with the pitting test results from the sun gears of off‐highway axles [Colour
figure can be viewed at wileyonlinelibrary.com]
TABLE 1 Details of tests reported in Figure 5

Kinematic
Type Viscosity of the
Test of Lubricant at End of
Reference Series Test SRR Temperature Ehta Material 40°C Test Criteria Contact
13
Rosh (1976) 1 ND −24% 50°C 0.15 mm 16MnCr5 100 cS ND Line contact
2 0.3 mm R′ = 10.5 mm
TERRIN AND MENEGHETTI

3 0.5 mm
4 1 mm
Graham et al (1980)15 5 4 discs −5.2% ND 0.65 mm EN39b 25 cS Torque measurement Line contact
6 −56.4% 25 cS (lower threshold) R′ = 15.4 mm
7 −5.2% 36.2 cS
8 −56.4% 36.2 cS
Kobayashi and Hasegawa (1990)14 9 2 discs −40% 80°C 0.7 mm Case‐hardened 87 cS Change on reflectivity Point contact
Cr‐Mo steel measured by optical R′ = 18.9 mm
(shot peened) fibre and
10 Case‐hardened accelerometers
Cr‐Mo steel
(un‐peened)
Wilkinson and Olver (1999)10 11 4 discs −5.2% ND 1.5 mm BS156 92 cS RMS of vibration Point contact
12 −29% signal R′ = 5.56 mm
Niemann et al (2005)26 13 ND 0 ND 0.75 mm 16MnCr5 ND ND Line contact
14 −10%
15 −30‐50%
Li and Kahraman (2011)27 16 2 discs −25% 90°C ND SAE 4620 M 6.61 cS at 90°C First macropit greater Exponential
17 60°C 15.39 cS at 60°C than 1 mm2 correction
of profile
Dengo et al (2015),28 Terrin et al. (2017)29 18 Gears ND 80°C 0.6–1.1 17NiCrMo6‐4 or ND Incipient macropitting Gears flanks
20MnCr5,
shot peened
19 17NiCrMo6‐4
or 20MnCr5,
un‐peened
a
Effective hardening depth: distance from the surface to the point where the hardness is 550 HV1.
5
6 TERRIN AND MENEGHETTI

different working conditions of gears in terms of lubri- approaches the cylindrical shape at the centre of the
cant, speed, and roughness. Because micropitting does contact area. However, to be effective, such correction
not prevent the operation of power transmission systems requires high accuracy in the angular positioning of
in heavy vehicles, gears were considered to fail when the specimens, which is not easy to guarantee when
incipient macropitting was observed on at least one tooth the discs are supported by cantilever shafts.
during periodic inspections. For consistency with the • Higher temperature and lower lubricant viscosities
criterion applied to actual gears, only the macropitting generally lead to shorter lifetimes because the capacity
failures relevant to disc tests were reported in Figure 5 of the lubricant to keep the mating surfaces separated
along with the gear data. The results relevant to case‐ is reduced.
hardened steel have been grouped in a single chart • The SRR plays a fundamental role in determining the
regardless of the material composition according to the durability of specimens, such that the lifetime is
ISO 6336‐5:2016.31 reduced as the percentage of sliding increases; in
Figure 5 highlights the wide scatter of results gener- some cases, increased sliding leads to different damag-
ated from the disc tests on case‐hardened steels, which ing modes, such as scuffing.9
in most cases showed much higher fatigue performances • The scatter of disc durability revealed by the literature
than gears. Based on the literature review of disc tests, review emphasizes that, to obtain representative data
the following considerations may be drawn: from disc tests, the manufacturing process and heat
treatment of the specimens, as well as the test param-
• Tests performed on cylindrical discs (usually with eters, must be accurately chosen to accurately recreate
chamfered edges) to obtain a line contact between the stress state, material properties, and working con-
the surfaces often lead to the formation of craters at ditions of the gear pair that are intended to be
the sides of the contact area due to the concentration resembled.
of stress caused by edge loading.10,27
• Tests on couples of discs where at least one disc is Notably, in the case of gears, the radii of curvature
crowned to obtain a point contact avoid the edge‐ and the SRR vary during the contact path. In the absence
induced damage but are subjected to a pressure gradi- of other causes of stress concentration (such as tip to root
ent in the transverse direction with respect to the interference, the edges of tip and root reliefs, and the fur-
motion, which causes uneven damage to the surface. rows caused by handling or previously formed micropits),
Thus, if micropitting or wear occurs, a groove may pitting usually forms first in the dedendum of gears,
form on the contact surfaces and cause an uncon- where negative values of the SRR exist. The SRR van-
trolled reduction in the contact pressure for a given ishes in correspondence with the pitch line, where the
applied load. relative motion between the mating teeth is from pure
• As a consequence of the previously mentioned issues, rolling and then finally increases positively throughout
specimens with crowned surfaces (point contacts) are the addendum.
typically more durable than cylindrical specimens In light of the previous considerations, the radii of the
(line contacts), as shown in Figure 5. Li and discs were chosen according to Figure 6 to obtain the
Kahraman27 recently proposed using a cylindrical disc same SRR calculated at the lowest point of single tooth
coupled with a second disc with rounded sides and a contact (LPSTC) of the sun gear teeth of a planetary gear
curvature that decreases exponentially as it set mounted in the wheel hub of a medium‐power axle

FIGURE 6 Design of the case‐hardened discs

TERRIN AND MENEGHETTI 7

for agricultural vehicles. That point was deemed the most Part of the disc specimens was shot peened with 100%
critical one point along the meshing path because it was coverage using a peen size of 0.7 mm and an Almen
subjected to the highest value of contact pressure occur- intensity of 20 A, to evaluate the possible beneficial
ring in the meshing path and had a negative SRR of effects induced by the treatment. The compressive
−24%. residual stresses introduced in the discs were measured
Despite point contact disc pairs tend to be more dura- by X‐ray diffraction at different depths of up to 300 μm
ble than line contact specimens, one of the discs was after removing the surface material from a small spot
crowned to prevent edge loading and to reduce the on the disc by electropolishing. As reported in Figure 8,
amount of force needed to generate a given contact pres- the residual stress profiles relevant to the disc specimens
sure in order to match the load capacity of the test rig are comparable to the stress profile measured near the
better. The 15‐mm crown radius of the larger disc was operating pitch diameter of the sun gears. No substantial
chosen to obtain patterns of the principal stresses close difference was found between the micro hardness profiles
to the ones experienced by gear flanks, at least beneath of case hardened and case hardened plus shot peened disc
the point of first contact,25 although the small crown specimens and gears.
radius causes a more severe pressure gradient across the Additional details about the specimen design, such as
width of the elliptical contact area. After carburizing, the stress field generated beneath the surface by contact,
quenching, and tempering, the surfaces of the discs were can be found elsewhere.25,30 The lubricant used for the
ground to obtain the desired geometry and eliminate tests is an API GL‐4 oil that is typically used for planetary
the surface layer where inter‐granular oxidation can be gear sets and has a kinematic viscosity of ν100 = 9.3 mm2/
induced by the carburizing process, which has detrimen- s at 100°C and ν40 = 56 mm2/s at 40° C. The oil temper-
tal effects on the surface fatigue strength.14 A comparison ature was measured at 87 ± 3°C at the outlet of the
of in‐depth Vickers hardness tests performed on the disc contact area by means of a heater and a flow adjuster.
specimens and sun gear teeth by using a 1‐kg weight The rig was set to automatically stop the test if the tem-
showed substantially identical hardness profiles, as perature reached 91°C.
shown in Figure 7.

4 | TEST RESULTS

Tests were conducted at contact pressures of 2650, 2850,

3000, 3150, and 3380 MPa, as reported in Table 2. The
spindle speed was adjusted in the range 1000 to
2000 rpm, depending on the applied pressure and ambi-
ent temperature, to prevent lubricant overheating.
In most cases, macropitting craters formed on the sur-
face of the smaller (cylindrical) disc, which was subjected
to the negative SRR.
The craters had the typical seashell shape with rela-
tively straight edges branching from the origin towards
FIGURE 7 Micro‐hardness profile of a sun gear tooth and a pair the rolling direction, as reported in Figure 9. A cross
of disc specimens (all the components were un‐peened) section of the specimens showed micropitting cracks

FIGURE 8 Residual stresses in the un‐

peened and shot‐peened gear teeth and
the disc specimens
8 TERRIN AND MENEGHETTI

TABLE 2 Summary of the pitting test results on the case‐hardened disc specimens
TERRIN AND MENEGHETTI 9

FIGURE 9 Top view (A) and section (B) of a macropit and top view (C) and section (D) of micropitting craters and surface distress cracks
in the smaller disc (SRR = −24%) of test #1 after 4.6·106 cycles at 3000 MPa

extending up to a depth of 20 μm, and the macropit was rare, was also reported by Murakami et al.33 The authors
generated by a crack that propagated to a depth of up to noticed that such craters showed a different morphology
200 μm and then branched and curved up towards the compared with those of the seashell‐shaped pits observed
surface. Similarly to what was observed in the gears,30,32 in the negative SRR discs and suggested that their origin
the crack often propagated through the subsurface even could have been in the subsurface. A section of the
after the formation of the first crater, branching and crowned disc of test #5 did not show any clear evidence
interacting with several other surface cracks in a more of the subsurface origin, although a small white etching
advanced location. Surface distress cracks and crack was observed at the bottom of the main crater. In
microcraters were observed also in the greater discs, as the components subjected to rolling contact fatigue, the
shown in Figure 10. material surrounding a crack often undergoes a micro-
The first crater manifested on the surface of the larger structural modification that prevents the etching of the
(crowned) disc only in test #5, which was performed with structure, leading to a characteristic white appearance
a nominal contact pressure of 3000 MPa, as shown in in microscopical observations. The resulting structure is
Figure 11. composed of equiaxed nanoferrite grains with a fine
As discussed earlier, the formation of macropits in distribution of carbide particles36-40 and can be 10% to
surfaces characterized by a positive SRR is unlikely. 50% harder than the matrix.40 However, whether a
Murakami and coworkers33 reported that surface cracks pre‐existing crack causes the microstructural change or
that had developed under negative SRR conditions during a crack develops as a consequence of the different hard-
disc tests stopped growing when the sign of the SRR was ness between the white etching area and the martensite
changed by reversing the rotational speed or exchanging matrix is still a controversial issue.40 On the other hand,
the follower and driver discs. Similar results were the crack could be related to the presence of a defect or
observed by Fujita and Yoshida34 and by Tyfour and inclusion at a location other than the plane examined in
Beynon.35 However, the formation of pitting craters on the studied section. As in the previously analysed cases,
the positive SRR specimen during the disc tests, although the surface microcracks were oriented from the surface
10 TERRIN AND MENEGHETTI

Due to the relatively small number of available test

results, an inverse slope of 13.22 was imposed to fit
the experimental data, according to ISO standard 6336.
Figure 12 compares the curves obtained for shot‐peened
and un‐peened specimens with the curve for medium
quality case‐hardened steels proposed by the standard,
which had previously been found to be in good
agreement with the experimental results from tests on
sun gears.28,29
If data relevant to shot‐peened and un‐peened speci-
mens are treated separately, the endurable contact
stresses at 50 million cycles for a 99% survival probability
are 15% and 29% higher, respectively, than the limit
proposed for medium quality steel by ISO standard
6336. Statistical analysis was carried out following the
FIGURE 10 SEM image of surface cracks and micro‐craters in recommendations of ISO standard 12107.41 Because
the larger disc (SRR = 24%) of test #1 after 4.6·106 cycles at the different widths of the scatter bands depend on the
3000 MPa
number of experimental data upon which the statistical
analysis is based, all the data were also analysed together.
to the interior of the material opposite of the sliding The resulting scatter index decreased to Tσ = 1.5, and
direction in the larger disc of test #5. hence, the endurance value at 99% of the survival proba-
Test #2, which was performed at the highest contact bility becomes σHlim disc, 99% = 2153 MPa, which is 44%
pressure of 3380 MPa, led to a failure mode due to higher than the ISO limit. The values of Tσ relevant to
diffuse plastic deformation at the surface rather than pitting tests on gears and discs typically range between
to pitting, The test was stopped due to the excessive 1.3 and 226,42 in the literature. The value Tσ = 1.63, which
vibrations of the rig and will not be considered in the is similar to that calculated in the present work, can be
following analyses. derived from the data reported by Li and Kahraman.27

FIGURE 11 A, Crater on the surface of the larger disc in test #5 after 4.69·106 cycles at 3000 MPa. B, Cracks on the surface of the
specimen. C, Section of the crater. D, White etching crack at the bottom of the crater
TERRIN AND MENEGHETTI 11

FIGURE 12 Fatigue curves obtained from the statistical analysis of the results from shot‐peened and un‐peened specimens and a
comparison with the curve provided by ISO standard 6336

FIGURE 13 Residual stress measured in the discs before and after the test

FIGURE 14 Measurement of the local disc geometry after test #4 (un‐peened discs) [Colour figure can be viewed at wileyonlinelibrary.com]
12 TERRIN AND MENEGHETTI

FIGURE 15 Profile measurements performed on the un‐peened cylindrical disc in tests #3, #4, #6 and #10 [Colour figure can be viewed at
wileyonlinelibrary.com]

FIGURE 16 Profile measurements performed on the shot‐peened cylindrical disc of tests #9, #11, and #12 [Colour figure can be viewed at
wileyonlinelibrary.com]

Based on the data collected within this work, no sig- reported an improvement in the durability of shot
nificant improvement in the limit stress or the statistical peened specimens tested at low pressure values, and a
scatter of the data were observed for the shot peening reduction in life when high pressures were applied.46,47
treatment. The chance to improve the rolling contact Finally, in some cases, shot peening has shown no sub-
fatigue durability of the components by means of shot stantial positive48 or detrimental effect49 on the contact
peening is discussed in the literature. Many fatigue life of discs. Possible negative effects are generally
authors14,43-45 claim that shot peening has beneficial attributed to an increase of roughness caused by the
effects. Conversely, other experimental works have treatment.31,49 Indeed, a roughness measurement per-
formed on 3 shot‐peened and 3 un‐peened discs in the
present work showed average Ra values of 0.70 and
TABLE 3 Estimated groove radius and contact pressure values at
the end of tests #3, #4, and #12 0.57 μm, respectively. Therefore, in the first part of the
test, the higher roughness of the shot‐peened specimens
Pressure at potentially counteracted the benefits of the compressive
Applied Initial Groove the End of
residual stress. Although asperities are flattened in the
Test Load Pressure Radius the Test
early phases of running‐in, X‐ray diffraction measure-
# [N] [MPa] [mm] [MPa] ments performed after tests #7 (shot peened) and #5
3 1940 2650 22 1860 (un‐peened) also showed a relaxation of the compressive
4 2815 3000 18 1770 residual stress in both shot‐peened and un‐peened discs
12 2410 2850 18 1680 (see Figure 13), as reported in previous works.45,46,50

FIGURE 17 Pitting test results for the

case‐hardened disc specimens and
compared with the sun gear data. The
figure shows the effects of contact
geometry modification
TERRIN AND MENEGHETTI 13

FIGURE 18 Geometry modifications of the cylindrical and crowned discs during tests #14 and #15

No measurement was performed in the present work to Kapoor and Johnson52 proposed a method to estimate the
investigate the evolution of residual stresses at interme- changes in surface geometry resulting from plastic defor-
diate intervals during the test. mation of an elastic‐perfectly plastic material subjected to
rolling contact. More recently Olver,32 in his exhaustive
review about the mechanisms of rolling contact fatigue,
5 | DISC USS I ON mentioned the major role covered by surface modifica-
tions occurring during running‐in. In the present work,
The higher durability of specimens with respect to gears changes to the surface of the discs were observed as a
observed in the present work could be attributed to the consequence of wear and micropitting. Figure 14 shows
changes in the contact geometry during the tests. One of the deviation from the original surface on the un‐peened
the most comprehensive attempts to explain the influence cylindrical specimen #4 after 11.36 · 106 cycles with a
of surface modifications in rolling contact fatigue tests load of 2.8 kN, which was intended to generate a contact
dates back to 1991 when, following the work of Hearle,51 pressure of 3000 MPa between the surfaces. The measure-
ment was performed using a Form Talysurf 50 (Taylor
30 Hobson Co Ltd) tactile surface profiler. At the end of

25 6
Maximum depth [µm]

5
20
Volume loss [mm3]

4
15
3
10
Un-peened
Shotpeened 2 Un-peened
5
Shotpeened
1
0
0.0E+00 5.0E+06 1.0E+07 1.5E+07 2.0E+07
0
Number of cycles 0.0E+00 5.0E+06 1.0E+07 1.5E+07 2.0E+07
Number of cycles
FIGURE 19 Changes in the groove depth during tests #14 and
#15 performed on an un‐peened and a shot‐peened disc pair, FIGURE 20 Volume of the material removed by the surface of
respectively, with a load of 2814 N (initial pressure = 3000 MPa) shot‐peened and un‐peened cylindrical discs during tests #14 and #15
14 TERRIN AND MENEGHETTI

FIGURE 21 Modification of the surface geometry of the sun gear reported in Figure 1, showing the deviation of the surface from its
nominal profile in the normal direction before and after the test [Colour figure can be viewed at wileyonlinelibrary.com]

the test, a groove was formed on the surface of the flat whole test. Tests #14 and #15 were performed to investi-
disc while the surface of the crowned disc underwent gate the variations of the contact pressure due to wear by
only mild flattening, which is not appreciable in the monitoring the contact geometry modifications on the
chart. A re‐evaluation of the contact pressure due to the surfaces of shot‐peened and un‐peened disc pairs. The
modified geometry of the cylindrical disc was possible irregular shape assumed by the wear groove that formed
by considering the width w of the groove 10 μm from its on the disc surface made it impossible to reliably calcu-
bottom and by finding the radius of the arc having a late the contact stress. Nevertheless, Figure 18 highlights
sagitta of 10 μm and a span equal to w. how the conformity of the surfaces of the cylindrical and
crowned specimens considerably increased after 1.5 · 106
w2 0:01 cycles. This experimental outcome suggests that the main
R¼ þ (2) cause of the higher durability of discs compared with
8·0:01 2
gears is related to the geometry modifications due to
In the case represented in Figure 14, the arc defined wear, which decreases the contact pressure and delays
by the points A, B, and C has a span of w = 1.2 mm, the onset of pitting. Although the groove depth increased
and the corresponding radius equals 18 mm. Combined more drastically within the initial 2 million cycles in the
with the 15‐mm crown radius of the mating disc, the con- un‐peened cylindrical specimen, as reported in
tact pressure was 1770 MPa, in contrast with the contact Figure 19, similar trends were noted in terms of material
pressure of 3000 MPa imposed at the beginning of the loss on both the un‐peened and shot‐peened discs, as
test. shown in Figure 20. The volume of material removed by
The profile measurements performed on the cylindri- the surfaces of the cylindrical discs during the tests was
cal disc after the test are reported in Figure 15 for the un‐ estimated by considering the product between the area
peened specimens of tests #3, #4, #6, and #10, and in of the groove and the circumference of the disc.
Figure 16 for the shot‐peened specimens used in tests It is important to note that the removal of material
#9, #11, and #12. from the surface due to wear and micropitting is
The tests performed at lower pressures typically show common also in sun gears. Nevertheless, as shown in
deeper grooves because lower loads mean longer life and Figure 21 (a picture of the damaged sun gear is reported
prolonged wear. In the case of tests #3 and #12, where in Figure 1), the geometry modifications in gears are
the wear track showed a regular shape and could be usually more pronounced in the dedendum, where
approximated quite well by a circumference arc, it was the SRR is negative, and therefore a groove forms
possible to estimate the pressure value at the end of the in the orthogonal direction with respect to the relative
test, as previously done in Figure 14, for test #4. The motion of the surfaces. Corresponding with the upper
results are reported in Table 3, and Figure 17 shows the edge of the groove, which is repeatedly overpassed by
contact pressure relevant to the modified geometry of the teeth of the mating gear, the contact pressure
the cylindrical discs, assuming that the curvature radius increases considerably due to the local reduction of the
of the crowned disc remains unchanged. curvature radius. Hence, the pitch line often constitutes
Figure 17 shows that the disc data are more consistent the site of macropit formation.
with the gear test data when the modification of the con- Hence, while wear modifies the surface geometry and
tact geometry is accounted for. However, the groove leads to increased fatigue performance in disc specimens,
geometry changed progressively during the tests, and the same phenomenon could drastically reduce the
therefore, the final pressure could not represent the lifetime of gears.
TERRIN AND MENEGHETTI 15

6 | CONCLUSIONS ORCID
A. Terrin http://orcid.org/0000-0003-0185-924X
Rolling contact fatigue tests were performed on both
G. Meneghetti http://orcid.org/0000-0002-4212-2618
shot‐peened and un‐peened 17NiCrMo6‐4 case‐hardened
steel specimens by means of a two‐disc test rig. Although
the disc geometry and test parameters were intended to RE FER EN CES
recreate the operating conditions of the sun gears in off‐ 1. Fajdiga G, Glodež S, Kramar J. Pitting formation due to surface
highway axles, the disc specimens showed considerably and subsurface initiated fatigue crack growth in contacting
higher durability compared with data relevant to gear mechanical elements. Wear. 2007;262(9‐10):1217‐1224.
tests. Surface modifications due to the removal of mate- 2. Errichello R. Morphology of micropitting. In American Gear
rial caused by micropitting and wear during the test made Manufacturers Association Fall Technical Meeting Cincinnati,
it difficult to interpret the results and to extrapolate an Ohio, 2011.
endurable contact pressure suitable for use in the context 3. Olver AV, Tiew LK, Medina S, Choo JW. Direct observations
of gear design. Nevertheless, some conclusions may be of a micropit in an elastohydrodynamic contact. Wear.
drawn on the basis of the outcomes of the tests performed 2004;256(1‐2):168‐175.
within the scope of the present work. 4. Davis JR. Gear Materials, Properties, and Manufacture. ASM
International, Materials park, OH44073‐0002; 2005.
• Surface distress constitutes the initial phase of the 5. Li S, Kahraman A. Micro‐pitting fatigue lives of lubricated point
damaging process, leading to the formation of cracks contacts: experiments and model validation. Int J Fatigue.
that propagate inside the material at small angles to 2013;48:9‐18.
the surface and in the direction against the sliding 6. AGMA 925‐A03, Effect of lubrication on gear surface distress.
direction of the counterface. This process is followed American Gear Manufacturing Association (2003).
by the formation of micropits with consequent mate- 7. Murakami Y, Sakae C, Ichimaru K. Three‐dimensional fracture
rial removal from the surface. Macropitting may even- mechanics analysis of pit formation mechanism under
tually form from subsequent crack propagation that lubricated rolling‐sliding contact loading. Tribol Trans.
occurs preferentially on the surface with negative SRR. 1994;37(3):445‐454.
• Substantial agreement both in terms of crater mor- 8. ISO 6336–1:2006, Calculation of load capacity of spur and
phology and the path of the crack beneath the surface helical gears—Part 1: Basic principles, introduction and general
was found between the macropitting observed in the influence factors.
gear and disc specimens. 9. Wilkinson CMR, Olver AV. The durability of gear and disc spec-
• The material removal due to wear and micropitting imens—Part I: The effect of some novel materials and surface
was deemed as the main reason for the higher durabil- treatments. Tribol Trans. 1999;42(3):503‐510.
ity of the discs compared with the gears. It was 10. Wilkinson CMR, Olver AV. The durability of gear and disc
estimated that the related enhancement of the confor- specimens—Part II: Post failure examination and gear‐disc
mity between the mating surfaces caused a decrease correlation. Tribol Trans. 1999;42(3):610‐618.
in the contact pressure during the test, resulting in a 11. Niemann G, Winter H. Maschinenelemente: Band 2: Getriebe
contact pressure that was comparable with that allgemein, Zahnradgetriebe ‐ Grundlagen, Stirnradgetriebe.
observed for gears. Berlin: Springer‐Verlag; 1985.
• Although few specimens were used, no significant 12. Flamand L, Berthe D, Godet M. Simulation of Hertzian contacts
improvement in the contact fatigue strength was found in spur gears with a high performance disk machine. J
observed for the shot‐peened discs compared with the Mech Des. 1981;103(1):204–209.
un‐peened discs. The experimental outcomes suggest 13. Rösch H. Untersuchungen zur Wälzfestigkeit von Rollen:
that additional compressive residual stresses intro- Einfluss von Werkstoff, Wärmebehandlung und Schlupfe
duced by shot peening are ineffective in preventing (Investigation on pitting capacity of rollers—influence of
contact fatigue damages, which develop under cyclic material, heat treatment and slip ratio). Dissertation, T.U.
Munchen, 1976.
compression and shear loading conditions.
14. Kobayashi M, Hasegawa K. Effect of shot peening on the pitting
fatigue strength of gears in International Conference on Shot
Peening ICSP‐4 1990. 465‐476.

ACK NO WLE DGE MEN TS 15. Graham RC, Olver A, Macpherson PB. An investigation into the
mechanisms of pitting in high‐hardness carburized steels. In
The authors would like to acknowledge Carraro S.p.A. for ASME Paper No. 80‐C2/DET‐118 Contribution of the American
technical and financial support of the present work. Society of Mechanical Engineers, 1980.
16 TERRIN AND MENEGHETTI

16. Simon M. Messung von elasto‐hidrodynamischen parameter, repeated lubricated rolling‐sliding contact: effects of reversal of
und hire Auswirkung auf die Grübchentragfähigkeit vergüteter rotation and change of the driving roller. J Tribol.
Scheiben und Zahnräder. (Measurement of EHD parameters 1997;119(4):788.
and their influence on pitting capacity of through hardened 34. Fujita K, Yoshida A. The effect of changing the rolling direction
discs and gears). Dissertation, T.U. Munchen 1984. on the rolling contact fatigue lives of annealed and case‐hard-
17. Hoehn B‐R, Michaelis K, Doleschel A. Limitations of bench ened steel rollers. Wear. 1977;43(3):315‐327.
testing for gear lubricants. In Bench Testing of Industrial Fluid 35. Tyfour WR, Beynon JH. The effect of rolling direction reversal
Lubrication and Wear Properties Used in Machinery on fatigue crack morphology and propagation. Tribol Int.
Applications ASTM STP 1404, American Society for Testing 1994;27(4):273‐282.
and Materials, West Conshohocken, PA, 2001.
36. Grabulov A, Petrov R, Zandbergen HW. EBSD investigation of
18. Kleemola J, Lehtovaara A. Experimental simulation of gear con- the crack initiation and TEM/FIB analyses of the microstruc-
tact along the line of action. Tribol Int. 2009;42(10):1453‐1459. tural changes around the cracks formed under rolling contact
19. BS 436–3:1986 ‐ Spur and helical gears. Method for calculation fatigue (RCF). Int J Fatigue. 2010;32(3):576‐583.
of contact and root bending stress limitations for metallic 37. Evans M‐H. White structure flaking (WSF) in wind turbine gear-
involute gears, Standard withdrawn in december 2005. ISBN box bearings: effects of ‘butterflies’ and white etching cracks
0580152278. (WECs). Mater Sci Technol. 2012;28(1):3‐22.
20. Way S. Pitting due to rolling contact. J Appl Mech. 1935;
38. Errichello R, Budny R, Eckert R. Investigations of bearing fail-
A49‐A58.
ures associated with white etching areas (WEAs) in wind
21. Littmann WE, Widner RL. Propagation of contact fatigue from turbine gearboxes. Tribol Trans. 2013;56(6):1069‐1076.
surface and subsurface origins. J Basic Eng. 1966;88(3):624.
39. Bruce T, Rounding E, Long H, Dwyer‐Joice RS. Characterisation
22. Bower AF. Influence of crack face friction and trapped fluid on of white etching crack damage in wind turbine gearbox
surface initiated rolling contact fatigue cracks. J Tribol. bearings. Wear. 2015;338–339:164‐177.
1988;110:704‐711.
40. Evans M‐H. An updated review: white etching cracks (WECs)
23. Bogdanski S. A rolling contact fatigue crack driven by squeeze and axial cracks in wind turbine gearbox bearings. Mater Sci
fluid film. Fatigue Fract Eng Mater Struct. Technol. 2016;32(11):1133‐1169.
2002;25(11):1061‐1071.
41. ISO 12107:2003—Metallic materials—Fatigue testing—Statisti-
24. Höhn BR, Michaelis K, Otto HP. Pitting load carrying capacity cal planning and analysis of data.
under increased thermal conditions. In 3rd International
42. AGMA 2101: Fundamental rating factors and calculation
Conference on Integrity, Reliability and Failure, Porto/Portugal
methods for involute spur and helical gear teeth.
2009.
43. Townsend DP, Zaretsky EV. Effect of shot peening on surface
25. Meneghetti G, Terrin A, Giacometti S. A twin disc test rig for
fatigue life of carburized and hardened AISI 9310 spur gears.
contact fatigue characterization of gear materials. Struct Integr
Gear Technol. 1986;29‐48.
Procedia. 2016;2:3185‐3193.
44. Widmark M, Melander A. Effect of material, heat treatment,
26. Niemann G, Winter H, Höhn B‐R. Maschinenelemente/Band 1,
grinding and shot peening on contact fatigue life of carburised
Konstruktion und Berechnung von Verbindungen, Lagern, Wellen.
steels. Int J Fatigue. 1999;21(4):309‐327.
Springer; 2005.
45. Batista D, Lebrun LF, Inglebert G. Contact fatigue of automotive
27. Li S, Kahraman A. A fatigue model for contacts under mixed
gears: evolution and effects of residual stresses introduced by
elastohydrodynamic lubrication condition. Int J Fatigue.
surface treatments. Fatigue Fract Eng Mater Struct.
2011;33(3):427‐436.
2000;23(3):217‐228.
28. Dengo C, Terrin A, Meneghetti G. Design curves against contact
46. Jia‐Xiang C, Bing‐ju A. Effect of shot peening on contact fatigue
fatigue for case hardened gears of off‐highway drivelines. In 44°
behavior life of carburized steel. In Proceedings of the I Interna-
Convegno AIAS, Messina 2015.
tional Conference on Shot Peening 1981, 333–340.
29. Terrin A, Dengo C, Meneghetti G. Experimental analysis of
47. Zhu J, Guo P. The research to the contact fatigue behaviour
contact fatigue damage in case hardened gears for off‐highway
influenced by the residual stresses on the surface of the shot
axles. Eng Fail Anal. 2017;76:10‐26.
peened specimen. In Proceedings 3rd International Conference
30. Terrin A. Analysis and design against pitting of gears in power on Shot Peening, Garmisch‐Partenkirchen, Germany 1987,
transmissions for off‐highway vehicles. Ph.D. dissertation, 525–532.
Department of Industrial Engineering, University of Padova,
48. Peyrac C, Flavenot J. Optimization of carburising and shot
2018.
peening, in order to improve both bending and contact fatigue
31. ISO 6336‐5:2016—Calculation of load capacity of spur and heli- behaviour for gear applications. In Proceedings of the 9th
cal gears—Part 5: Strength and quality of materials. International Conference on Shot Peening, Paris 2005.
32. Olver A. V. The mechanism of rolling contact fatigue: an update. 49. Vrbka M, Křupka I, Svoboda P, et al. Effect of shot peening on
Proc Inst Mech Eng Part J J Eng Tribol. 2005;219:313‐330. rolling contact fatigue and lubricant film thickness within mixed
33. Murakami Y, Sakae C, Ichimaru K, Morita T. Experimental and lubricated non‐conformal rolling/sliding contacts. Tribol Int.
fracture mechanics study of the pit formation mechanism under 2011;44(12):1726‐1735.
TERRIN AND MENEGHETTI 17

50. Fernández Pariente I, Guagliano M. Contact fatigue damage

analysis of shot peened gears by means of X‐ray measurements. How to cite this article: Terrin A, Meneghetti G.
Eng Fail Anal. 2009;16(3):964‐971. A comparison of rolling contact fatigue behaviour
51. Hearle AD. Deformation, shakedown and fatigue in rolling of 17NiCrMo6‐4 case‐hardened disc specimens and
contact. Ph.D. Dissertation, University of Cambridge, 1985. gears. Fatigue Fract Eng Mater Struct. 2018;1–17.
52. Kapoor A, Johnson KL. Effect of changes in contact geometry on https://doi.org/10.1111/ffe.12867
shakedown of surfaces in rolling/sliding contact. Int J Mech Sci.
1992;34(3):223‐239.