Shifeng Wu

Sliding Wear Calculation in Spur

Applied Research,
John Crane Inc., Gears
Morton Grove, III.
Assoc. Mem. ASME In gear applications where precipitous tooth failure mode such as scoring or scuffing
has been avoided, "normal" wear becomes a life-determining factor. In this paper,
sliding wear in spur gears, including the considerations of gear dynamics and rough-
H. S. Cheng elastohydrodynamic lubrication, is analyzed. Formulas for equivalent wear rate and
Center for Engineering Tribology, tooth wear profile along the line of action are derived. Results show that most
Department of Mechanical Engineering, materials are removed from both the addendum and dedendum tooth surfaces, and
Northwestern University,
Evanston, III.
that the highest wear occurs at the beginning of an engagement. This high wear
Fellow ASME
region corresponds to the root of the driving (pinion) teeth and the tip of the driven
(gear) teeth. These analytical results correlate well with the practical evidences in
AGMA documentation.

Introduction
The importance of being able to predict wear in gearing limited dealing with rubbing wear under sliding-rolling con-
systems has long been acknowledged by engineers considering ditions (Wu and Cheng, 1991b), much less under the transient
that gears are the most commonly used elements in power conditions of gear-tooth action. However, there does seem to
transmission and catastrophic wear leads to equipment and be a general consensus in the classification of major lubrica-
eventually to system failure. But owing to the complex nature tion-related gear tooth failure modes, which can be categorized
of wear processes, no reliable and simple quantitative law conveniently as scoring (scuffing), surface pitting, and material
comparable to that for friction is presently available. The fact removal by wear, in addition to the strength-related failure
that most gears run in the mixed- or partial-EHL regime further modes such as plastic flow and tooth breakage (Ku, 1975;
complicates the wear behavior. In this regime, the speed of Dudley, 1980; AGMA, 1980).
gears is high enough to create a partial-EHL film, but the film Scoring or scuffing is a severe form of adhesive wear and
is inadequate to separate the contacting surfaces from direct is precipitous (Blok, 1958; Ku, 1974; AGMA, 1980). In con-
asperity contacts which inevitably produce wear particles. trast, surface pitting and sliding wear generally take time for
Nevertheless, there have been many investigations on gear wear their action to reach destructive magnitude (Ku, 1974; Litt-
in the past decades. For example, in trying to lay some ground- mann, 1970; Archard, 1980). In applications where precipitous
work on gear lubrication research, Ku (1975) offered a broad tooth failure mode such as scoring (scuffing) has been avoided,
perspective on the importance of lubrication and mechanics sliding wear and surface pitting become the major factors
on gear failure modes. From the viewpoint of design practice, affecting gear life (Ku, 1975). Pitting is mainly the result of
Dudley (1980) analyzed the wear failure modes which occur surface fatigue (Littmann, 1970), upon which different lubri-
frequently in gear operations, and provided the information cation regimes have little influence; sliding wear, on the other
on the various things that lead to wear. He also suggested ways hand, is the consequence of direct asperity contacts between
to avoid or reduce the severity of wear in contacting gear teeth. mating surfaces with relative motion, and is strongly affected
A rather complete list of the analytical tools and methods used by the performance of the lubrication process (Rabinowicz,
to diagnose various wear mechanisms was compiled by Godfrey 1965; Archard, 1980; Rowe, 1980).
(1980). In this paper, an attempt will be made to analyze the sliding
In spite of the numerous efforts on gear wear research in wear in spur gears including the considerations of gear dy-
the past, there is still little understanding of the basic failure namics and rough-elastohydrodynamic lubrication. The dis-
mechanisms in gear applications. Although extensive investi- tributions of various sliding wear parameters such as wear rate
gations have been carried out on rubbing wear under steady- and wear depth, will be calculated by utilizing a sliding wear
state simple sliding conditions such as those with pin-on-disk model developed before by Wu and Cheng (1991b) for partial-
machines (Ku, 1968; Rabinowicz, 1965; Archard, 1953, 1980; EHL contacts.
Bowden and Tabor, 1954, 1964; Ling, 1973); there is rather
Sliding Wear Model for Partial-EHL Contacts
Contributed by the Tribology Division of THE AMERICAN SOCIETY OF M E - A necessary condition for sliding wear to occur under lu-
CHANICAL ENGINEERS and presented at the ASME/STLE Tribology Conference, bricated contacts is the existence of asperity rubbing between
San Diego, CA, October 18-21, 1992. Manuscript received by the Tribology
Division February 18, 1992; revised manuscript received July 15, 1992. Paper
surfaces in relative motion. Namely, the contacts must be in
No. 92-Trib-33. Associate Technical Editor: K. Komvopoulos. the boundary or partial-elastohydrodynamic lubrication re-
gime. In the partial-EHL regime, both hydrodynamic and sur-

Journal of Tribology JULY 1993, Vol. 115/493

Copyright © 1993 by ASME
Downloaded From: http://tribology.asmedigitalcollection.asme.org/ on 01/27/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use
 face films exist. The hydrodynamic film acts to separate the area associated with an adsorbed lubricant molecule; U, the
mating bodies and the surface films such as adsorbed and sliding velocity; t„, the fundamental time of vibration of mol-
oxidative films prevent the asperities from direct metallic con- ecule in adsorbed state; £ , the heat of adsorption of mineral
tacts. Although both oil and surface films are the controlling oil molecules on steel surfaces; A0, the Arrhenius constant for
factors in influencing the tribological performance such as the oxidation; Q0, the activation energy for linear oxidation; Ach
loading capacity and traction characteristics of a partial-EHL the local asperity contact area for asperity i; A„, the nominal
contact, it is the latter that affects most the wear behavior. Hertzian contact area; Th the local asperity contact temper-
The interaction among opposing asperities generates more ature for asperity /; R, the molar gas constant; and Cx, the
heat in the vicinities of these asperities because of the higher oxide constant. It has three different values for three different
local pressures and fluid shear rates. Consequently, the tem- oxides and can be written as
peratures at these asperities are higher. This action may first M,
cause micro-EHL films at asperities to collapse, resulting in Cx = X'pVe (2)
direct metallic contacts among opposing asperities which have M Fe
surface film defects. The breaking of these metal-metal junc- where p Fe is the density of iron; M 0 is the molecular weight
tions is producing adhesive wear debris. This is the main ar- of oxygen; MFe is the molecular weijht of iron; and x can be
gument of the thermal desorption wear mechanism in partial- 3/4, 2/3, or 1/2 for rhombohedral ( a - F e 2 0 3 ) , spinel (Fe 3 0 4 ),
EHL contacts (Rowe, 1966). If the heat generated at the vi- or wiistite (FeO) oxides, respectively, depending on the contact
cinities of some asperities is excessively high, the flash tem- temperature. It has been shown that, when the contact tem-
peratures at these asperities can be readily above the desorption perature is in the ranges of 200°C to 350°C, 350°C to 570°C,
temperature of the lubricant species. Then, at these "hot or higher than 570°C, the primary oxide wear debris will be
spots," oxide films can be formed. When the oxide reaches a rhombohedral, spinel, or wiistite (Kubaschewski and Hopkins,
critical thickness, it spalls off and appears as wear particle 1962; Quinn, 1967, 1983) and x takes the values of 3/4, 2/3,
(Quinn, 1967, 1983). This manifests itself as the oxidative wear or 1/2. The total wear rate equation of a partial-EHL contact
mechanism. can be obtained by taking the summation of all the local wear
Based on these arguments, an elaborate sliding wear model volumes over the whole contact area, that is,
was developed by Wu and Cheng (1991b) for partial-EHL
contacts. In this model the thermal desorption wear mechanism
at low asperity contact temperature and the oxidative wear (3)
mechanism at elevated asperity contact temperature were con- *•(?')
sidered. This wear model was subsequently verified experi- where W is the total wear rate. It is a function of the sliding
mentally by using a gear simulator—a two-disk machine. velocity, the distributions of asperity contact area and its cor-
According to this wear model, the local wear volume of one responding asperity contact temperature within a partial-EHL
asperity, Vh accounting for both thermal desorption (7} < contact, and therefore, implicitly also a function of load.
200°C) and oxidation (7} > 200°C), can be expressed as: Among all the constants embedded in the sliding wear equa-
X tion, there are two parameters of special significance. They
j- = km-A„)l-exp exp are respectively the wear coefficient parameter specific to the
U'tc R-Tj characteristics of the metal-metal contact, km\ and the Ar-
rhenius constant for oxidation, A0. The parameter km is of
'6a Ti<200°C importance in the region where the asperity contact temper-
A, ature is relatively low and the thermal desorption wear mech-
A0'An anism predominates and the parameter A0 is pivotal in the
_Qo_
exp r(>2oo°c (i) elevated asperity contact temperature region where the oxi-
' Cx-U dative wear mechanism predominates. The values used for all
where L is the sliding distance; km, the wear coefficient pa- the parameters in this sliding wear equation are listed in Table
rameter specific to contacting asperities; X, the diameter of 1.

Nomenclature
Vj = local wear volume for asper-
A, a = addenda of gear and pinion, subscript representing indi- ity ;' (m3)
respectively (m) vidual asperity contact Vm = wear volume for one single-
b = face width of gear teeth (m) L sliding distance (m) engagement (m3)
Ac = total asperity contact area Lg, Lp lengths of recession and ap- W total wear rate (mVm)
(m2) proach of tooth mesh (m) X diameter of area associated
A„ = nominal Hertzian contact MFe molecular weight of iron with an adsorbed lubricant
area (m2) (kg/kmole) molecule (m)
A0 = Arrhenius constant for oxi- M0 molecular weight of oxygen x = a fraction number denoting
dation (kg/m 2 »s) 2
(kg/mole) the density of different fer-
Aci = local asperity contact area Qo activation energy for oxida- rous oxides
for asperity i (m2) tion (J/mole) A = wear depth along the line of
= oxide constants R molar gas constant (J/ contact (m)
c
\d x = pitch diameters of gear and mole-K) 5Vm = infinitesimal sliding wear
pinion respectively (m) T, local asperity contact tem- volume in duration 8T (m3)
E = heat of adsorption of min- perature for asperity / (K) 8T = infinitesimal time (s)
eral oil molecules on steel to fundamental time of vibra- £ = a dummy variable for coor-
surfaces (J/mole) tion of molecule in adsorbed dinate along the line of ac-
km = wear coefficient parameter state (s) tion
specific to contacting asperi- U sliding velocity (m/s) p Fe = density of iron (kg/m 3 )
ties UP pitch line velocity (m/s) \p = pressure angle (°)

494 /Vol. 115, JULY 1993 Transactions of the AS ME

Downloaded From: http://tribology.asmedigitalcollection.asme.org/ on 01/27/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 Table 1 Parameter values used in sliding wear model
Parameters Values Units GEAR
km 5xl(TK
A0 4xlO (kg/m «s)
E 49 (kJ/mole)
Qo 193 (kJ/mole)
R 8.31 (J/mole-K)
X 3x10"' (m)
to 3x10"' (s) 3,
PFc 7800 (kg/m )
M0 32 (kg/kmole)
. 56 (kg/kmole)
MFr2

Sliding Wear in Spur Gear Operations

For years, gear operations have been used as examples of
partial-elastohydrodynamic lubrication because the EHL film
thickness between mating gear teeth is in the same order of
magnitude as their surface roughness. This is a clear indication
of the non-existence of an intact EHL film between the surfaces
of mating gear teeth. Therefore, the occurrence of sliding wear
in most gears is inevitable. Among the many gear types, the
most commonly used and yet the simplest one is the involute
spur gear.
The involute spur gears during tooth engagement can be
modeled as two constantly changing cylinders in contact. The Fig. 1 Gear tooth contact configuration
mesh cycle can then be considered as a series of changing
elastohydrodynamic contacts as shown in Fig. 1, in which the
two instantaneous cylinders on contact can be replaced by an where Lg and Lp denote the lengths of approach and recession,
equivalent cylinder and a plane. For involute spur gears, the respectively (Fig. 1) and can be written for involute spur gears
tooth contact always lies on a straight line tangential to both
base circles and passing through the pitch point, P. This straight
line is called line of contact or line of action and is used in d
this study as the basic coordinate system with the origin at the • X Jd A d
•, (6a)
pitch point. The sliding wear model can be applied locally *>-§ - sin^ l + 41 - + 1 I - - sint/-
along this line of action during one single-engagement for its
sliding wear rate as a function of the local sliding speed, load
and radius of equivalent cylinder. Since the sliding velocity,
load and radius of equivalent cylinder are changing along the D
line of action, the sliding wear rate can be expressed ultimately L. = - sinxj/ A D A D • , (6b)
+ 41—+1 - — smi/-
as a function of the contacting position, £. The infinitesimal \A A
wear volume on one gear tooth, SVm, due to sliding at contact
position, £, along the line of action during any infinitesimal in which a and A are the addenda of pinion and gear-; re-
time, <5T, can be written as: spectively.
SVm=W-\U\'dT. (4a) The kernel of the above integral (Eq. (5)):

By converting the infinitesimal time, 8T, into spatial interval, — =2.sec^(- + - ) . ^ l (7)
d%, along the line of action; and expressing the sliding velocity,
U, as a function of contact position, £:
can be defined as the equivalent sliding wear rate which, instead
of expressing the wear volume in unit distance slid as the
8T = — •sec\p-dS; traditional wear rate does, depicts the changes in wear volume
Up
in a constant interval (or duration) along the line of action.
Then, the sliding wear depth, A, at any contact position, £,
l/=2I/,.l*|.A + i can be calculated by the following:
6Vm 1 1 1*1
Equation (4a) can be rewritten as: >W> (8)
= — = 2.secM- + -
+ (4b) where b is the face width of gear teeth. The above formulations
5F m = 2.secV" \^ ^)'W'l^d^ provide the basis for predictions of sliding wear in spur gears.
where d and D are the diameters of pinion and gear, respec-
Other Considerations in Gear Wear Calculations
tively; Up is the pitch line velocity; and \p is the pressure angle,
the angle between the common normal of the two contacting Gear operations, even in their simplest form—the engage-
teeth and the common tangent to the pitch circles. ment of involute spur gears—are of great complexity because
By integration, the sliding wear volume, Vm, for one gear of their dynamic and transient nature. It is particularly true
tooth during a single-engagement can be expressed as (Wu, when dealing with lubrication-related phenomena in the par-
1990): tial-EHL regime such as sliding wear which can be affected
by an array of factors in many different ways. The following
1 l\ [LP
K,„ = 2«seci/" + IV ld€, (5) is a brief description of the major.factors considered in this
d D) L study.

Journal of Tribology JULY 1993, Vol. 115/495

Downloaded From: http://tribology.asmedigitalcollection.asme.org/ on 01/27/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 Table 2 Data used in sliding wear calculations asperity friction. The coefficient of asperity friction itself is
Gear Data determined by the characteristics of surface film formed by
Pitch diameters: cm (in) boundary lubricants. This friction model is used to calculate
Pinion 15.2(6.0) the hydrodynamic coefficient of friction accounting for the
Gear 25.4 (10.) fluid shearing. Then, in conjunction with Lee and Cheng's
Number of Teeth:
Pinion 48 rough-EHL model (1991 and 1992) the total coefficient of
Gear 80 friction accounting for both fluid and asperity shearing is cal-
Diametral pitch: 1/in 8.0 culated by assuming the coefficient of asperity friction as a
Pressure angle: 20° constant.
Face width: cm 4.0
Contact ratio: 1.79 Elastic Tooth Deflection and Dynamic Loading. For most
Lubricant data (SAE 80W-90 GL-5) spur gear applications, the contact ratio is usually between 1
Ambient viscosity: Pa-s (30°C) .2238 and 2. Namely, during tooth engagement there will be, at times,
Thermal conductivity: W/mK .125 only one pair of teeth in contact, and at other times, two pairs
Pressure-viscosity coefficient: 1/Pa 1.15X 1 0 "
Temperature-viscosity coefficients (/30, /3j): K 1040., - 1 7 5 . of teeth in contact. This condition is statically indeterminate
because of the elastic deflection of gear teeth. The static load
Material properties of gears exerted on two pairs of teeth in contact simultaneously may
Young's modulus: GPa 200.
Combined Young's modulus: GPa 218. not be equal, since at any instance the combined stiffness of
Poisson ratio: 3 .2875 one pair may not be equal to the combined stiffness of the
Density: kg/m 7800. other. Therefore, it is necessary to consider the elastic tooth
Specific heat: J/kgK 544. deflection under load for each pair of teeth in order to deter-
Thermal conductivity: W/mK 46.
mine the load sharing characteristics between the pairs of teeth
Surface topographic properties (nominal) in contact.
Composite RMS Roughness: jam (^in) .34(14.) .40(16.) At any point along the line of action, the actual load usually
Coefficient of asperity friction: .10
does not follow the static loading pattern. The variation in
Operating conditions dynamic load may result from a number of situations. For
Power transmitted: kW 100. instances, the transitions from single to double and from dou-
Angular velocity of pinion: rpm 1000.
Bulk temperature: °C 50.
ble to single pair of teeth in contact, the variation in combined
Pitch line velocity: m/s 8.0 bending stiffness of a pair of contacting teeth along the line
Load per unit width: kN/m 334. of action, or the deviation of tooth profile from true involute
(Single tooth contact) profile may all be the causes of dynamic loading. In this study,
only the first two, that is, those due to the inertia forces of
the gear wheel will be considered by superimposing an addi-
tional oscillation term on the static load. The approach em-
Rough-Elastohydrodynamic Lubrication. Spur gears have ployed in this study follows closely that developed by Wang
counterformal surfaces and are therefore subject to high con- and Cheng (1980) for calculating the gear tooth stiffness and
tact stresses and strains under load. These can result in local dynamic load pattern.
elastic deformation which gives rise to elastohydrodynamic
lubrication when the counterformal surfaces are in relative Results and Discussion
motion (Grubin, 1949). If the elastohydrodynamic film thick-
ness parameter which is defined as the ratio of average film First, tabulated in Table 2 are the gear data and operating
thickness to composite RMS surface roughness, is less than 3, condition as well as lubricant and surface topographical data
the consideration of rough elastohydrodynamic lubrication used in the sliding wear calculations. The pinion-gear com-
should be included (Ku, 1974; Akin, 1974; Nemlekar and bination is typical in that the pinion is the driving member and
Cheng, 1980). In this study, the rough-EHL model developed the gear is the driven member. One gear lubricant with the
by Lee and Cheng (1991 and 1992) for rough surface contact specification of SAE 80W-90 GL-5 is used. It is a mineral oil
with pure longitudinal roughness profiles is used. In this model with known viscosity as a function of temperature and pres-
the concept of average gap rather than compliance was em- sure. The material property data of 52100 steel are utilized for
ployed, and this average gap can be physically interpreted as the contacting teeth of pinion and gear with an assumed surface
the average film thickness. The crucial point is to find the load hardness of Rc56 ~ 62. As for the operating conditions, the
and average gap relationship which can be accomplished transmitted power through this pinion-gear combination is 100
through numerically simulating the contacts between rough kW and the angular velocity of pinion is 1000 rpm; this gen-
surfaces using actual digitized roughness profiles of the mating erates a static load of 334 kN/m for single tooth contact.
bodies. The results of contact simulation can be used to carry Since the bulk temperature of tooth surfaces affects appre-
out a series of rough-EHL calculations, and the ultimate out- ciably the film thickness which in turn has a profound influence
come is the cumulative asperity contact area and asperity con- on the distributions of asperity contact area and asperity con-
tact temperature distributions. These distributions are essential tact temperature, a good estimate of the bulk temperature is
in applying the sliding wear Eq. (3) to the calculation of sliding deemed necessary. Wang and Cheng (1980) calculated the equi-
wear in spur gears (Eq. (4)). librium surface temperature distribution of gear teeth by using
finite element techniques and showed that the equilibrium tem-
Friction in Partial-EHL Contact. In partial-EHL contacts, perature varies from about 45 °C to 75 °C depending on a
the friction is contributed by two components: the shearing of number of factors. For the cases in this study, a 50°C bulk
fluid and the sliding of asperity contacts. This friction is pri- temperature is considered a good estimate.
marily responsible for flash temperature rise, which affects the Two longitudinal surface roughness profiles, as shown in
sliding wear of an EHL contact. Therefore, to characterize Fig. 2, are used in the calculation. These profiles were taken
sliding wear in partial-EHL contacts, it is necessary to analyze in a direction perpendicular to the longitudinal lays which
the friction first. Wu and Cheng (1991a) developed a friction coincide with the sliding direction. They have composite RMS
model for partial-EHL contacts. In this model, it was assumed surface roughness parameters of 0.34 ywn and 0.40 /xm, re-
that the contribution from the shearing of fluid film is con- spectively. Each profile shown here is actually the statistical
trolled by the rheology of lubricants and the contribution from average of twelve individual traces from two ground disks
sliding of asperity contacts is dictated by the coefficient of during different intermittent stops of a two-disk test. Of the

4 9 6 / V o l . 115, JULY 1993 Transactions of the ASME

Downloaded From: http://tribology.asmedigitalcollection.asme.org/ on 01/27/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 F Composile RMS Surface Roughness: 0.34 um
Q
— 0.34 urn (RMS)
3 - .
s 1.25 - . ,
1
rofi
0.00-
0.
m
-0.25 -

.
VA^wv^vvviy ^/^
Composite RMS Surface Roughness: 0.40 Mm

0.25 1
P
A A
J \ l\ A ,
A
Ai/ \
H\
/
M ;J\
„/ \A/ /
0.00 •
r U
I /V
v/i/A
W *f \A
-
LL -0.25 - / V ^ A f J Ur" u\r W M /
l A/
(0
V t-
3 E
CO , 1 , . , 1 , , ,

0 200 400 600 800 1000

Traverse Distance (jim) -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6

Fig. 2 Digitized composite surface roughness profiles measured in a Normalized Line of Action
direction perpendicular to the longitudinal lays Fig. 5 Distributions of film thickness parameters, the ratio of minimum
film thickness to composite RMS roughness

Jf*Y
o
O 0.02

-0.6 -0.4 -0.2 0.0 0.2 0.4 0.6

-0.6 -0.4 -0.2 0.0 0.2 0.4 0.6
Normalized Line of Action
Normalized Line of Action
Fig. 3 Comparison between dynamic and static load patterns
Fig. 6 Distributions of coefficients of friction (coefficient of asperity
friction: 0.10)

Sliding Velocity the left to the origin is the approach where the coordinates are
Roiling Velocity
Equivalent Radius negative and the segment on the right is the recession where
the coordinates are positive.
The distributions of sliding and rolling velocities along the
line of action are presented in Fig. 4, in which the rolling
velocity shows a continuous increase as the tooth engagement
progresses but the sliding velocity shows a decrease in mag-
nitude in the approach and an increase in the recession. The
rolling velocity is important in influencing the film thickness
and the sliding velocity is important in determining the flash
-0.4 -0.2 0.0 0.2 temperature rises in the EHL contacts. Therefore, both are of
Normalized Line of Action significance in controlling the wear rate distribution along the
Fig. A Distributions of sliding and rolling velocities as well as equiv- line of action. For later reference, the changes in the contacting
alent radius instantaneous cylinders is also plotted in this figure in the form
of their equivalent radius.
Shown in Fig. 5 are the distributions of film thickness pa-
twelve traces, there are six from each contacting body. It is rameters, which are defined as the ratio of minimum film
those twelve digitized roughness profiles that are used in con- thickness to the composite RMS surface roughness, along the
tact simulation to obtain the asperity pressure as a function line of action for two different rough surfaces. The changes
of average gap between two rough surfaces; and subsequently, in both curves are the direct results of the changes in the
in rough-EHL calculations to obtain the distributions of as- dynamic load exerted on the gear teeth and the changes in the
perity contact area and its corresponding contact temperature. radius of equivalent cylinder, as shown in Figs. 3 and 4. The
To illustrate the significance of dynamic loading, Fig. 3 radius of the equivalent cylinder increases gradually and the
shows a comparison between dynamic and static load distri- dynamic load fluctuates as the gear tooth contact progresses.
butions along the line of action. The static load variation shows •The distributions of coefficient of friction along the line of
the load sharing characteristics between two pairs of teeth in action are illustrated in Fig. 6. The hydrodynamic coefficient
contact due to elastic deflection. The fluctuation in dynamic of friction corresponds to a smooth elastohydrodynamic con-
load reflects the varying tooth stiffness due to elasticity and tact where only shearing of fluid contributes to the total fric-
alternating engagement of single and double pairs of teeth, tion. Rough surfaces increase considerably the overall
and the oscillation due to the inertia forces of the gear wheels. coefficients of friction in partial-EHL contacts as shown here
The abscissas in this figure and those to follow have all been in Fig. 6, where the coefficient of asperity friction is assumed
normalized along the line of action. The sum of the lengths to be 0.10. The total friction together with the distributions
of approach and recession, Lg + Lp, is used as the reference of sliding velocity and pressure will.be used to calculate the
length. The origin is set at the pitch point. The segment on distributions of asperity contact area and the corresponding

Journal of Tribology JULY 1993, Vol. 115/497

Downloaded From: http://tribology.asmedigitalcollection.asme.org/ on 01/27/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 7000
0.40 jam RMS Roughness
6000- n
\
- J 5000

4000
I
1 t
-

i .
3000

2000
• I
1000
\J^J -0.4 -0.2
1
0.0
1 ^ff^^TT^^^Hiiii^i^i
0.2 0.4
< 390
>< Normalized Line of Action
-0.6 -0.4 -0,2 0.0 0.2 0.4 0.6
m Fig. 9 Distributions of wear rate and equivalent wear rate with rough
S Normalized Line of Action
surfaces of 0.40 urn RMS value
Fig. 7 Distributions of maximum average-asperity-contact temperature

0.34 pm RMS roughness

0.40 fim RMS roughness

-0.4 -0.2 0.0 0.2 0.4

Normalized Line of Action
Normalized Line of Action Fig. 10 Wear depth distributions or tooth wear profiles after one million
Fig. 8 Distributions of wear rate and equivalent wear rate with rough cycles
surfaces of 0.34 pm RMS value

among the interacting asperities. In this case, a transition could

average asperity contact temperature. As an illustration, the occur to promote excessive wear, rendering a drastic rise in
distributions of maximum average-asperity-contact-tempera- wear rate (Wu and Cheng, 1991b). In the cases concerned here,
ture rises along the line of action are plotted in Fig. 7, which this limit was never exceeded.
shows the temperature pattern as well as how high the asperity To describe the sliding wear phenomena in spur gears in
temperature can reach. The variations displayed in these curves terms of wear rate may not be adequate and may even be
are the combined results of the changes in sliding velocity (Fig. confusing to a large degree because of the way the wear rate
4) and in dynamic load (Fig. 3). is defined. The wear rate historically is defined as the ratio of
Presented in Fig. 8 are the distributions of wear rate (Eq. wear volume to the distance slid for simple sliding conditions
(3)) and equivalent wear rate (Eq. (7)) in one single engagement such as that with the pin-on-disk machine (Archard, 1953). It
along the line of action for rough surfaces of 0.34 nm RMS does not take into account the effect of the changes in tooth
value. The same distributions for rough surfaces of 0.40 nm surface velocities in gear applications. For example, to slide a
RMS value are shown in Fig. 9. At pitch point where the sliding unit distance around pitch point takes much longer time than
velocity equals zero, no sliding wear takes place by definition; at the rest of the contacts, because of the smaller sliding velocity
sliding wear rate is not defined at that point. To distinguish around the pitch point. Therefore, a better description of slid-
it from the rest along the line of action, a zero wear rate is ing wear accounting for both sliding and rolling may lie in the
assigned arbitrarily, and consequently, all other related sliding distributions of equivalent wear rate (Eq. (7)) and wear depth
wear parameters have zero values at the pitch point. (Eq. (8)) along the line of action.
From the distribution curves of sliding wear rate, it is seen The equivalent wear rate as defined in Eq. (7) is essentially
that around pitch point where sliding is small, the wear rate the dynamic change in wear in a constant interval (or duration)
has higher values than at the rest of contact where higher sliding at any contacting position along the line of action, and con-
exists. This is in agreement with what has been observed ex- sequently, at the corresponding contacting point along the
perimentally on a sliding-rolling gear simulator (Wu and Cheng, tooth profiles. Therefore, it can be conveniently converted into
1991b). In the region where sliding is relatively small, the the wear depth as shown in Eq. (8). Such wear profiles are
thermal desorption wear mechanism predominates and the wear shown in Fig. 10 for different surface roughnesses.
rate increases as sliding decreases. Physically, this phenomenon It should be noticed that the definition of equivalent wear
was explained by Kingsbury (1958) who argued that, from the rate in spur gears includes the effect of changes in tooth surface
standpoint of thermal desorption, a higher sliding velocity velocity and that the reference distance is an interval, A£,
allows adsorbed lubricant molecules less time to detach from traveled by the tooth contact along the line of action, instead
surfaces, resulting in a smaller wear rate; but in the situation of the unit distance slid along a sliding direction. This interval
where sliding is small, the adsorbed lubricant molecules have can also be conceptually considered as a constant duration the
ample time to take off from the surfaces because of the changes tooth contact travels since it travels at a constant speed along
in surface potential and a higher wear rate can be obtained. the line of action. This leads naturally to the observation that
However, this argument holds only when sliding is within a during the duration for the tooth contact to travel a constant
certain limit beyond which excessive heat can be generated interval, A£, along the line of action, the sliding distance of

4 9 8 / V o l . 115, JULY 1993 Transactions of the ASME

Downloaded From: http://tribology.asmedigitalcollection.asme.org/ on 01/27/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 Static Load (Single Pair of Teeth) (kN/m) ness; and this increase in film will reduce not only the per-
446 334 223 centage of asperities in contact but the contacting temperature
as well, and thus decreases wear. In the cases analyzed here
—a 0.34 firm RMS Roughness the effect of the increase in film thickness coupled with a
decrease in load due to the increase in gear speed outweighs
the effect of the increase in sliding velocity. Therefore, a de-
crease in wear can be observed as the gear speed is being
i io3 ^ ^ increased.
Under a given operating condition, the notion that rougher
surfaces produce more wear is very obvious. However, the
^ degree to which rougher surfaces aggravate wear under dif-
ferent operating conditions may not be uniform. In the cases
102
of low speed-high load where interacting asperities undergo
600 800 1000 1200 1400 1600 far more elastic and plastic deformations, the changes in sur-
face roughness alters little the relations among asperity pres-
Pinion Speed (rpm)
sure, asperity contact area and average gap; and as a
Fig. 11 Total wear volume for different pinion speeds
consequence, wear is relatively insensitive to the changes in
surface roughness. In contrast, the effect of surface roughness
on wear becomes more pronounced for cases of high speed-
the contact is shorter around the pitch point than at any other low load, because the changes in surface roughness in the latter
position because of the smaller sliding velocity. Consequently, situations affect appreciably the asperity load and average gap
the equivalent wear rate curves appear smooth around the pitch relation between two contacting rough surfaces.
point as compared to the high values observed in the sliding
wear rate curves (Figs. 8 and 9). Therefore, from the distri-
butions of wear rate and wear profiles one sees that although Conclusion
the sliding wear rate has higher values around the pitch point In this paper, sliding wear in involute spur gears, including
because of the small sliding velocity, it is at the commencement the considerations of gear dynamics and rough-elastohydro-
of an engagement that most wear occurs because of the higher dynamic lubrication in partial-EHL contacts, is analyzed. For-
sliding and higher loading. This point corresponds to the root mulas for equivalent wear rate and tooth wear profile along
of pinion teeth and the tip of the gear teeth. This conclusion the line of action are derived. The gear tooth wear profiles
is in agreement with the assessment that it is usually the de- due to sliding indicate that it is from both the addendum and
dendum section of the driving gear (pinion, in the current case) dedendum tooth surfaces that most materials are removed;
that first shows the signs of surface distress (AGMA, 1980), and that the highest wear occurs at the beginning of an en-
and it is also in this section that most wear occurs. The ag- gagement which corresponds to the dedendum of the driving
gravation of wear by a rougher surface is also observable from gear and the addendum of the driven gear. These analytical
the wear profiles as shown in Fig. 10. results correlate well with the practical evidences documented
It must be pointed out that the squeeze effect on the elas- in AGMA's standards (AGMA, 1980).
tohydrodynamic film formation was ignored in the current
study. As a consequence, smaller film thicknesses are obtained, Acknowledgment
especially in the beginning of tooth engagement. Therefore, The authors are grateful for financial support from the ECUT
the actual wear due to sliding at the beginning of tooth en- Tribology Program of Department of Energy, under NIST
gagements may not be as high as predicted in this study. How- grant No. 60NANB1D1104, administered by the National In-
ever, in light of the current moderate operating conditions, stitute of Standards and Technology.
the effect of squeeze film is not expected to be large, and the
current prediction still sheds light on this difficult subject of
gear wear in partial-EHL contacts. Nevertheless, a better es- References
timate of wear warrants the inclusion of transient-EHL analysis AGMA Standard, 1980, "Nomenclature of Gear Tooth Failure Modes,"
in the future studies. American Gear Manufacturers Association, ANSI/AGMA, 110.04.
Akin, L. S., 1974, " E H D Film Thickness Formulaes for Power Transmission
To demonstrate the effect of operating conditions on wear, Gears," ASME JOURNAL OF LUBRICATION TECHNOLOGY, Vol. 96, pp. 426-431.
three pinion speeds of 750 rpm, 1000 rpm, and 1500 rpm with Archard, J. F., 1953, "Contact and Rubbing of Flat Surfaces," Journal of
the same transmitted power of 100 kW are used to calculate Applied Phys., Vol. 24, pp. 981-988.
the total wear volume of one pair of teeth in one single-en- Archard, J. F., 1980, "Wear Theory and Mechanism," ASME Wear Control
Handbook, ed. M. B. Peterson and W. O. Winer, pp. 35-80.
gagement. The results are plotted in Fig. 11 for two different Blok, H., 1958, "Lubrication as a Gear Design Factor," Proc. Int. Conf. on
surface roughnesses. Since the power transmitted for all three Gearing, IMechE, pp. 144-158.
speeds is the same, the loads carried by the contacting teeth Bowden, F. P., and Tabor, D., 1954, The Friction and Lubrication of Solids,
will be different: a smaller speed yields a higher load; whereas Part I, Oxford University Press, 1954.
Bowden, F. P., and Tabor, D., 1964, The Friction and Lubrication of Solids,
a higher speed gives rise to a lighter load. In this case, the Part II, Oxford University Press.
static loads carried when there is only one pair of teeth in Dudley, D. W., 1980, "Gear Wear," ASME Wear Control Handbook, ed.
contact are, respectively, 446 kN/m for 750 rpm, 334 kN/m M. B. Peterson and W. O. Winer, pp. 755-830.
for 1000 rpm, and 223 kN/m for 1500 rpm. It is demonstrated Godfrey, D., 1980, "Diagnosis of Wear Mechanisms," ASME Wear Control
Handbook, ed. M. B. Peterson and W. O. Winer, pp. 283-311.
in this figure that, there is far less wear for the cases of high Grubin, A. N., 1949, "Fundamentals of the Hydrodynamic Theory of Lu-
speed-light load than for the cases of low speed-high load; and brication of Heavily Loaded Cylindrical Surfaces," Central Scientific Research
the effect of surface roughness on wear is more pronounced Inst, for Tech. andMech. Eng'ng, Book No. 30, Moscow, D.S.I.R. Translation,
in the cases of high speed-light load than in the cases of slower pp. 115-166.
Kingsbury, E. P., 1958, "Some Aspects of the Thermal Desorption of a
speed-heavier load. Boundary Lubricant," J. of Applied Phys., Vol. 29, No. 6, pp. 888-891.
The effect of gear speed on wear is two-fold: The increase Ku, P. M., 1968, ed. Interdisciplinary Approach toFriction and Wear, NASA
in speed, on one hand, increases the sliding velocity which Spec. Publ., 181.
Ku, P. M., 1975, "Gear Failure Modes—Importance of Lubrication and
tends to raise the contacting temperature, and therefore, pro- Mechanics," ASLE Trans., Vol. 19, pp. 239-249.
motes more wear; on the other hand, the increase in gear speed Ku, P. M., 1974, "Tribology of Gears and Splines," Proc. Tribology Work-
increases also the rolling velocity which increases the film thick- shop, NSF, pp. 402-427.

Journal of Tribology JULY 1993, Vol. 115/499

Downloaded From: http://tribology.asmedigitalcollection.asme.org/ on 01/27/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 Kubaschewski.O., and Hopkins, B.E., 1962, Oxidation of Metals and Alloys, Rabinowicz, E., 1965 Friction and Wear of Materials, Wiley, New York.
2nd ed., Butterworths, London. Rowe, C. N., 1966, "Some Aspects of the Heat of Adsorption in the Function
Lee, S. C , and Cheng, H. S., 1992, "On the Relation of Load to Average of a Boundary Lubricant," ASLE Trans., Vol. 9, pp. 100-111.
Gap in the Contact Between Surfaces With Longitudinal Roughness," STLE Rowe, C. N., 1980, "Lubricated Wear," ASME Wear Control Handbook,
Tribology Transactions, Vol. 35, No. 3, pp. 523-529. eds., M. B. Peterson and W. O. Winer, pp. 143-160.
Lee, S. C , and Cheng, H. S., 1991, "Further Correlation Between Scuffing Wang, K. L., and Cheng, H. S., 1980, "A Numerical Solution to the Dynamic
Experiments With EHL Analysis of Rough Surfaces," ASME JOURNAL OF TRI- Load, Film Thickness, and Surface Temperature in Spur Gears, Part I—Anal-
BOLOGY, Vol. 113, pp. 318-326. ysis," ASME Journal of Mechanical Design, Vol. 103, pp. 177-187.
Ling, F. F., 1973, Surface Mechanics, Wiley, New York, 1973.
Wang, K. L., and Cheng, H. S., 1980, " A Numerical Solution to the Dynamic
Littmann, W. E., 1970, "The Mechanism of Contact Fatigue," Interdisci-
Load, Film Thickness, and Surface Temperature in Spur Gears, Part II—Re-
plinary Approach to the Lubrication of Concentrated Contacts, NASA Special
Publ. 237, pp. 309-368. sults," ASME Journal of Mechanical Design, Vol. 103, pp. 188-194.
Nemlekar, N. E., and Cheng, H. S., "Traction in Rough Elastohydrodynamic Wu, S., and Cheng, H. S., 1991a, "A Friction Model of Partial-EHL Contacts
Contacts," ASME Publ., Proc. of Symposium on Surface Roughness Effects and its Application to Power Loss in Spur Gears," STLE Tribology Transac-
in Hydrodynamic and Mixed Lubrication, pp. 173-191. tions, Vol. 34, No. 3, pp. 398-407.
Quinn, T. F. J., 1983, "Review of Oxidation Wear, Parts I and I I , " Tribology Wu, S., 1990, "Friction and Wear Modeling in Spur Gears," Ph.D. Disser-
International, Vol. 16, Part I pp. 257-205 and Part II pp. 305-315. tation, Northwestern University, Evanston, 111.
Quinn, T. F. J., 1967, "The Effects of 'Hot-Spot' Temperatures on the Wu, S., and Cheng, H. S., 1991b, "A Sliding Wear Model for Partial-EHL
Unlubricated Wear of Steel," ASLE Trans., Vol. 10, pp. 158-168. Contacts," ASME JOURNAL OF TRIBOLOGY, Vol. 113, pp. 134-141.

Subscribe To
Journal of Applied Mechanics
This ASME Transactions Journal covers the areas of: aerodynamics; aeroelasticity; boundary layers;
computational mechanics; constitutive modeling of materials; dynamics; elasticity; flow and
fracture; heat transfer; hydraulics; impact; internal flow; mechanical properties of materials;
micromechanics; plasticity; stress analysis; structures; thermodynamics; turbulence; vibration; and
wave propagation.
Published quarterly: March, June, September, December
$165 Non-Members/$40 ASME Members

To order write ASME Order Department, 22 Law Drive, Box 2300, Fairfield, NJ 07007-2300
or call 800-THE-ASME (843-2763) or fax 201-882-1717.

500/Vol. 115, JULY 1993 Transactions of the ASME

Downloaded From: http://tribology.asmedigitalcollection.asme.org/ on 01/27/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use