See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/276844639

Effects of misalignment on surface wear of spur gears

Article in Proceedings of the Institution of Mechanical Engineers Part J Journal of Engineering Tribology · March 2015
DOI: 10.1177/1350650115574732

CITATIONS READS

41 1,853

2 authors:

Jun Zhang Xianzeng Liu

Fuzhou University Tianjin University
144 PUBLICATIONS 1,525 CITATIONS 7 PUBLICATIONS 227 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Jun Zhang on 24 July 2015.

The user has requested enhancement of the downloaded file.

XML Template (2015) [4.3.2015–5:20pm] [1–14]
//blrnas3.glyph.com/cenpro/ApplicationFiles/Journals/SAGE/3B2/PIJJ/Vol00000/150026/APPFile/SG-PIJJ150026.3d (PIJ) [PREPRINTER stage]

Original Article

Proc IMechE Part J:

J Engineering Tribology
Effects of misalignment on surface 0(0) 1–14
! IMechE 2015
wear of spur gears Reprints and permissions:
sagepub.co.uk/journalsPermissions.nav
DOI: 10.1177/1350650115574732
pij.sagepub.com

Jun Zhang and Xianzeng Liu

Abstract
A surface wear prediction methodology for spur gears is proposed based on which the wear characteristics of the gear
pair can be computed efficiently. The methodology combines the finite element method and Archard’s formula to predict
the surface wear of the spur gear pair. The influences of load levels and wear cycles are investigated to reveal a quantitive
relationship between wear and operating parameters. The misalignment of the gear pair is defined and counted in the
quasi-static model to investigate its effects on gear wear. The results indicate that the introduction of misalignment may
aggravate the load distribution status and accelerate the surface wear. To compensate the negative effects of misalign-
ment, an involute crown modification strategy is proposed and its effectiveness on contact distribution improvement and
wear deceleration are validated through wear depths predictions. The numerical simulation finds that the proposed
modification strategy can help to decelerate surface wear effectively when the summation of involute crown values C
remains within a certain range. Furthermore, a larger C will be required to decrease the wear depth of the gear pair
with larger misalignment values.

Keywords
Spur gears, surface wear, misalignment, tooth modification, Archard’s formula

Date received: 29 July 2014; accepted: 19 January 2015

Introduction involved.24 Herein, k denotes the wear coefficient

Surface wear is one of the four major failure modes in which can be determined with empirical formulas; s
gear systems.1 As a long-term material removal represents the sliding distance between the mating sur-
behavior, surface wear not only alters the tooth pro- faces at the point of interest and can be obtained
file that leads to reduction in gear transmission accur- through kinematic derivations; p is the contact pres-
acy and efficiency but also aggravates system’s sure and is determined either by Hertzian contact4–14
characteristics of vibration and noise.2,3 Therefore, it or by numerical method such as boundary elem-
is of great importance to investigate surface wear in ent15–22 or finite elment23 formulations. As far as the
gear systems, especially for those function in dirty contact pressure is concerned, most of the previous
enviroments with heavy loads such as mining and researches applied a quasi-static p to Archard’s equa-
transportation machinaries. A thorough understand- tion4–10,17–23 while only a few authors used a dynamic
ing of wear mechanism in gear systems helps to reveal p during integral computations.11–16
the effects of design parameters on wear rate. The According to the simulation contents, the focuses
knowledge of this may provide useful information to of wear prediction in aforementioned researches can
gear designers who attempt to decrease wear rate and be roughly classified into two groups: one group
prolong service life. focused on the interaction between wear and tooth
There has been an increasing interest on the surface flank shape4–11,18–21 and the other group emphasised
wear of gearings since the 1990s. Some efforts in wear the interaction between wear and dynamics.12–17
modeling and prediciton have been conducted in the Though a few of these studies took account of the
past two decades, which provide a solid foundation
for further study of gear wear.4–23 There is a common
thread in these efforts in that they all combined School of Mechanical Engineering, Anhui University of Technology,
Archard’s wear model with a gear contact model Ma’anshan, China
and relative sliding distance (or velocity) calculations.
Corresponding author:
The wear depth of a local point on the contacting gear Jun Zhang, School of Mechanical Engineering, Anhui University of
surface can be calculated through Archard’s formula, Technology, Ma’anshan 243032, China.
in which three parameters of k, s, and p are Email: [email protected]

Downloaded from pij.sagepub.com at Anhui University of Techno. on March 9, 2015

XML Template (2015) [4.3.2015–5:20pm] [1–14]
//blrnas3.glyph.com/cenpro/ApplicationFiles/Journals/SAGE/3B2/PIJJ/Vol00000/150026/APPFile/SG-PIJJ150026.3d (PIJ) [PREPRINTER stage]

2 Proc IMechE Part J: J Engineering Tribology 0(0)

effects of surface deviation and misalign- pressure P, and the sliding distance s. The following
ments,1,11,14,19,22 none has established quantitive will breifly introduce the determination of the three
relationships between wear and imperfections, not to parameters k, P, s.
mention the coupling effects of misalignments, modi-
fication, and operating parameters.
Determination of wear coefficient k
The present work has a different perspective in that
it aims to explore the effect of misalignments on spur Wear coefficient k is affected by several factors such as
gear wear with the consideration of operating condi- material characteristics, surface roughness, lubrica-
tions and gear micro-geometry modifications. To tions and operating conditions. Due to the complexity
compensate the negative effects of misalignment, an of gear contact, the wear coefficient k in gear wear
involute crown modification strategy is proposed. For analysis is often determined either by experimental
this purpose, a new quasi-static wear model for spur tests or through statistical regressions. Janakiraman
gears has been developed. The computation of contact et al.25 statistically analyzed the parametric effects
pressure p is realized in the commercial software of load, speed, lubrication, and surface roughness
Kisssoft to ensure maximum portability and reusabil- on surface wear and proposed an approximate for-
ity. The obtained contact pressure p is then combined mula of wear coefficient k as
with the sliding distance s and the wear coefficient k to
integrally compute the wear depth. 3:981  1029 1:219 7:377 1:589
k¼ L G S ð3Þ
E0
Computational methodology of where L is the dimensionless load, G is the dimension-
gear surface wear less lubricant pressure–viscosity coefficient, S is the
dimensionless composite roughness amplitude, and
Wear formulation
E0 is the equivalent Young’s modulus. And there exist
Gear meshing includes combined rolling and sliding
 
actions of two mating surfaces that have varying geo- W0 0 Rc 1 1 1  21 1  22
metries and load conditions, whose wear depth can be L¼ , G ¼ E , S ¼ , ¼ þ
E0 R0 R0 E0 2 E1 E2
calculated through Archard’s equation as shown in ð4Þ
equation (1). The Archard’s equation assumes that
the wear depth is proportional to the arithmetic prod- where W0 is the load per unit length, R0 is the equiva-
uct of the contact pressure and the sliding distance, lent radius of contact surface,  is the pressure–vis-
with a constant proportionality, known as the wear cosity coefficient, Rc is the composite roughness
coefficient, representing the influences of other param- amplitude, Ei and ti (i ¼ 1,2) are the Young’s modulus
eters such as surface morphology, lubrication condi- and Poisson’s ratios of the mating surfaces, respect-
tion, and material characteristics. ively. And R0 and Rc can be calculated as

V W 1 1 1 c
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
¼K ð1Þ ¼ þ , R  ¼ R1 þ R2 ð5Þ
s H R0 R1 R2

Herein, V is the volume of the worn away material, s where Ri (i = 1, 2) are the radii of contact point of the
is the sliding distance, K is the dimensionless wear two mating surfaces, Ri (i ¼ 1,2) are the roughness
coefficient, W is the applied normal load, and H is amplitudes of the two mating surfaces.
the hardness of the contact surface. The hardness of It is clear from equation (3) that wear coefficient
the contact surface is assumed to be constant during k increases with the load and the roughness amplitude
meshing. Therefore, the wear equation of a local point while decreases with lubricant pressure–viscosity coef-
on one of the mating surfaces can be further ficient. Among all these three parameters, G is the
expressed as most influentail parameter followeed by a less influ-
ential parameter of S and L is the least influential one.
Z s
h¼ kpds ð2Þ
0 Computation of contact pressure P
In this subsection, the computation of contact pres-
where h is the wear depth, s is the sliding dis- sure of engaging surfaces in one wear cycle is intro-
tance, p is the contact pressure, and k is the wear duced. A complete wear cycle of the gear pair is
coefficient. defined as from the point where the tooth enters the
From equation (2), it can be clearly observed that contact zone for the first time to the point where it
the calculation of wear depth of gear surface requires exits the contact zone completely. To obtain the dis-
the acquirement of wear coefficient k, the contact tribution of surface pressure, a three-dimensional

Downloaded from pij.sagepub.com at Anhui University of Techno. on March 9, 2015

XML Template (2015) [4.3.2015–5:20pm] [1–14]
//blrnas3.glyph.com/cenpro/ApplicationFiles/Journals/SAGE/3B2/PIJJ/Vol00000/150026/APPFile/SG-PIJJ150026.3d (PIJ) [PREPRINTER stage]

Zhang and Liu 3

node ij is assigned as zero when it is not contained in

−M ξ M the contact zone at given rotation positions.
N
η Since the time when a point on gear surface within
−N the contact zone is very short when compared with
ij one wear cycle, this makes it reasonable to use the
i=I Tip mean pressure of a contacting point to substitute for
its instanenous pressure when predict the wear depth.
Involute Direction

By replacing the instanenous pressure with mean pres-

sure of a contacting point, the computation efficiency
i can be improved dramatically while keeping satisfac-
tory accuracy. The mean value of contact pressure at
node ij is obtained by averaging all contact values at
each discretized rotation positions during one mesh
cycle and stored for wear depth computation.
i=0 Root
It is worthy to point out that the misalignment of
j=0 j j=J
the gear pair can be included in the contact model at
Lead Direction the very beginning stage of 3D parametric FE model-
ing. The angle of misalignment is defined as one of the
Figure 1. Moving contact grids on fixed surface grids. initial parameters of the gear pair, which are further
used to constitute the 3D contact model. For this pur-
pose, two FE models are firstly developed for the
pinion and the gear, respectively; then the axis of
the pinion is set as the base line according to which
(3D) parametric finite element (FE) model is devel- the axis of the gear is determined; finally, the two
oped with Kisssoft. With the developed FE model, a gears are assembled to form a gear pair according to
loaded tooth contact analysis (LTCA) is conducted to kinetic and geometric relationships. With the
compute the contact pressure at different rotaion pos- proposed FE contact model, the contact pressures of
itions of the gears in mesh. Noticing that the contact an example gear pair shown later in section
pressures vary with the rotation positions of the gear ‘‘Numerical simulation and discussions’’ with and
pair, one may discretize the loaded flanks of spur gear without misalignment are computed and demon-
pair into (I þ 1) equally spaced lead lines along the strated in Figure 2.
profile direction and (J þ 1) equally spaced profile As can be seen from the figures, the introduction of
lines along the lead direction as shown in Figure 1. misalignment alters the contact status of the engaging
Hence, any point on the gear surfaces can be denoted gear pair. The contact pressures of gear pair with mis-
by the grid node ij, (i ¼ 0, 1, 2 . . . , I; j ¼ 0, 1, 2 . . . , J). alignment are no longer evenly distributed as those of
To make sure the accuracy of pressure distribution gear pair without misalignment along the lead direc-
prediction, the contacting grids need to be refined. tion; meanwhile, the contact pressures along the pro-
To achieve this, the instantaneous contact zone is file direction vary from those of gear pair under idea
equally divided into 2N slices along the profile direc- condition. The variation of contact pressures will alter
tion and 2M slices along the lead direction as shown the wear depth of the gear pair in turn.
in Figure 1.
With the knowledge of gear kinetics, it is clear that
the refined contacting grids move on the fixed grids of
Calculation of sliding distance s
surface flanks with the rotation of gear pair. Becasuse The sliding distance is defined as the distance by
of the moving contact zone in a meshing cycle, the which a point represented by node on one gear
pressure value of node n acquired from the LTCA slides with respect to its corresponding point on the
needs to be converted to the pressure value of node ij mating gear as the gear rotates from one position to
of the fixed surface grid. The contact pressure on a another one. The sliding distance can be acquired by
moving grid cell n and its coordinates in the inertial using the integral for the relative speeds relate with
gear frame can be calculated through the FE model. time. The relative sliding distances of nodes ij and uv
The pressure value of node n can be represented by on the fixed surfaces of gears p and g during one wear
the average pressure value of the contacting cell n cycle after the qth pressure update can be obtained as
predicted by the FE contact model if only the size followings
of each grid cell is sufficiently small. If a fixed grid
node ij is contained in any of the loaded moving grid
 p Z tpO Z tpO  
cells n at a given rotation position, then the contact vij ðtÞ  vuv ðtÞdt
sqij ¼ vðtÞdt ¼ ð6Þ
pressure at node ij is equal to the pressure value of  tpI tpI
moving node n. Otherwise, the contact pressure at

Downloaded from pij.sagepub.com at Anhui University of Techno. on March 9, 2015

XML Template (2015) [4.3.2015–5:20pm] [1–14]
//blrnas3.glyph.com/cenpro/ApplicationFiles/Journals/SAGE/3B2/PIJJ/Vol00000/150026/APPFile/SG-PIJJ150026.3d (PIJ) [PREPRINTER stage]

4 Proc IMechE Part J: J Engineering Tribology 0(0)

Figure 2. Contact pressures of an example gear pair (Tp = 500 Nm): (a) without misalignment; (b) with misalignment of
 ¼ 210–5 rad.

Z tgO Z tgO
q g  
suv  ¼ vðtÞdt ¼ vij ðtÞvuv ðtÞdt ð7Þ Compuation of surface wear depth h
tgI tgI
The wear depth occurred at each node ij of the fixed
where v(t) is the relative sliding speed, vij(t) and vuv(t) surface grid during one wear cycle can be calculated
are the speeds of the contact points on gears p and g in by using equation (2) with the mean pressure value
the tangential direction with respect to tooth flanks and the sliding distance obtained above. If the wear
during one wear cycle after the qth pressure update; tpI depth of any node on the contact surfaces reaches the
and tgI are the instants when the nodes ij and uv first predetermined wear threshold "q after the qth pressure
come into contact zone with its mating counterparts; update, the geometry of gear surface needs to be
in addition, tpO and tgO are the instants when the nodes reconstituted to perform contact analysis afresh for
cease contact, respectively. renewed surface pressure. The iterative process is

Downloaded from pij.sagepub.com at Anhui University of Techno. on March 9, 2015

XML Template (2015) [4.3.2015–5:20pm] [1–14]
//blrnas3.glyph.com/cenpro/ApplicationFiles/Journals/SAGE/3B2/PIJJ/Vol00000/150026/APPFile/SG-PIJJ150026.3d (PIJ) [PREPRINTER stage]

Zhang and Liu 5

repeated until the maximum total wear depth on

either of the two gears reaches the maximum allow-
able wear threshold "t .
The accumulated wear depth at any node ij during
one wear cycle is given as
 p,g  p,g  p,g
hqij ¼ k pqij sqij ð8Þ
  

 p,g
where p qij is the mean pressure value at node ij

 p,g
during one wear cycle, and sqij is the sliding dis-

tance at node ij during one wear cycle after the qth
pressure update.
Equation (8) is applied continuously q times until
the wear depth accumulated at any node of either one
of the contacting surfaces after the qth geometry
update reaches the threshhold of "q . Then, the wear
amounts at nodes ij accumulated since the last geom-
etry update are given as Figure 3. Computational methodology of gear wear.
 p,g  p,g
hqij ¼ q hqij ð9Þ


The total number of wear cycles q is given as Once the wear depth of any point on mating surfaces
during one wear cycle is obtained, the above integral
"q process is repeated until the wear depth of any point
q ¼  p  q ð10Þ
on the contact surfaces reaches a predetermined wear
max hqij , hqij
  threshold, under which the contact pressures change
accordingly. Then the gear surface needs to be recon-
The wear amounts accumulated after each geom- stituted to perform another contact analysis for
etry updated are summed to get the final wear depth renewed surface pressure. The iterative process is
of node ij as repeated until the maximum total wear depth on
either of the two gears reaches a maximum allowable
Q  p,g
X wear threshold. The wear depth of every point on the
hp,g
ij ¼ hqij ð11Þ surface is obtained by summing up wear depth of
q¼1 every point accumulated in all processes of different
pressure update.
The total number of wear cycles required to reach
to this wear depth is
Numerical simulation and discussions
X
Q With the proposed computational methodology in
t ¼ q ð12Þ section ‘‘Computational methodology of gear surface
q¼1 wear’’, a numerical example is used to demonstrate
the prediction of wear depth of engaging gear sur-
faces. Then the misalignment and micro-geometry
modification are introduced into the wear model
Compuation methodology of spur gear pair and their effects on gear wear are investigated to pro-
As aforementioned, gear wear is a long-term material vide useful information for gear designers. The design
removal process, which indicates that the numerical parameters are listed in Table 1.
method is the only possible solution for wear depth According to the LTCA simulations, the influence
calculation. The computational methodology of surface profile deviation on contact pressures is
employed here to predict the wear of contacting quite trivial when the deviation value is less than
gear surfaces is shown in Figure 3. 2 lm. This means that a wear threshold value below
As shown in Figure 3, a 3D FE model is developed 2 lm results in very slight changes, and a wear thresh-
to compute the mean value of contact pressures, old value of 2 lm or larger will be a practical value.
which are then combined with the sliding distance For the sake of accuracy, the wear threshold value for
and the wear coefficient to calculate the wear depth. surface profile update is set to 2 lm in this study.

Downloaded from pij.sagepub.com at Anhui University of Techno. on March 9, 2015

XML Template (2015) [4.3.2015–5:20pm] [1–14]
//blrnas3.glyph.com/cenpro/ApplicationFiles/Journals/SAGE/3B2/PIJJ/Vol00000/150026/APPFile/SG-PIJJ150026.3d (PIJ) [PREPRINTER stage]

6 Proc IMechE Part J: J Engineering Tribology 0(0)

Table 1. Parameters of the example system.

Parameter Pinion p Gear g Parameter Pinion p Gear g

Number of teeth, z 29 57 Module, m (mm) 4.0 4.0

Pressure angle, an/(º) 20 20 Tooth width, F (mm) 48 48
Pitch diameter, dw (mm) 116.000 228.000 Base diameter, db (mm) 109.004 214.250
Addendum diameter, da (mm) 124.000 236.000 Dedendum diameter, df (mm) 106.000 218.000
Normal circular tooth thickness, sd (mm) 6.283 6.283 ISO precision grade 5 5
Material 45# steel 45# steel Density,  (kg/m3) 7800 7800
Young’s modulus, E (GPa) 207 207 Poisson’s ratio, t 0.3 0.3
Roughness, Ra (lm) 0.6 0.6 Wear coefficient, k (m2/N) 5  1019 5  1019
Input torque Tp (Nm) 500
Center distance, a (mm) 172.000
Load per unit width, W (N/um) 0.18
Pressure–viscosity coefficient, a (GPa) 6.8

During this interval, it is assumed that the micro-geo- On the contrary, the maximum wear occurs near the
metry and pressure are constant. As shown in the dedendum region of each gear followed by a second-
flowchart in Figure 3, once the accumulated wear highest value occurring near the addendum region.
depth of the gear surface reaches 2 lm, the contact Around these regions, the sliding velocities are much
model for the gear pair will be reconstructed by count- higher than those at other areas, which in turn leads
ing in the worn-off 2 lm profile. With the recon- to the acceleration of surface wear. By comparing the
structed contact model, a new LTCA is performed wear depths of two gears, it can be found that the
to obtain the renewed pressure distributions, with maximum and second-highest wear depths of the
which a new wear integration starts. For the present pinion are much higher than those of the gear, with
case study, the geometry and pressure updates are values of 16.00 lm and 2.80 lm with respect to
carried out in 2 lm and the simulation is terminated 3.83 lm and 1.10 lm. These observations agree well
after eight geometry updates. Meanwhile, the wear with the previous experimental data and engineering
coefficient is determined according to empirical for- wear failure mode, which in turn, manifests that the
mulas and applied to the pinion and the gear with quasi-static wear model proposed in this paper can
the same value of k. The sliding distance s between correctly predict the wear behavior thus can be used
the mating surfaces at the point of interest can be for further investigations on effects of design param-
obtained through kinematic derivations. eters such as operating conditions, misalignments and
modifications.
The following will explore the effects of operating
Wear of an ideal spur gear pair
parameters of load levels and wear cycles on surface
The simulated wear of the pinion and the gear after wear. Firstly, the wear depths along tooth profile of
136 million wear cycles are presented in Figure 4, in the pinion and the gear after a series of numbers of
which the rotations of the gears are denoted as the wear cycles (106 is the cycles’ unit), e.g. 4, 20, and 60
radial values of corresponding meshing points on under a given load of Tp¼500 Nm are computed and
mating surfaces. plotted in Figure 5.
It can be seen from Figure 4 that the wear depths As can be seen from Figure 5, the wear cycles
are uniformly distributed along the lead direction affects the wear depths of spur gears significantly.
both in the pinion and the gear. This is coincident The wear depths increase with the increments of
with the contact states of an ideal spur gear pair in wear cycles dramatically. For example, the maximum
that the pressure distributions are even along the wear depths of the pinion reach 0.47 lm, 2.36 lm, and
tooth width. Nevertheless, the wear depths of both 7.09 lm at 4, 20, and 60 million cycles, respectively.
the pinion and the gear vary dramatically along the Compared with the pinion, the wear depths of the
profile direction with the minimum wear at the pitch gear are much smaller at each kind of wear cycles,
points and the maximum wear at the roots of the gear ranging from 0.11 lm to 1.69 lm. Though the values
pairs. To be specific, the wear depth approaches zero of wear depth in both gears are different, the curves of
at radial values of 116 mm on the pinion p and wear depth after different wear cycles share the same
228 mm on the gear g, which are exactly the pitch tendency for both the pinion and the gear as indicated
circles of the two gears. Though the contact pressures in Figure 4.
reach their maximus at pitch points, the sliding dis- Figure 6 demonstrates the influences of wear cycles
tance is zero at this position thus leads to no wear. on the maximum wear depths of gear pair under

Downloaded from pij.sagepub.com at Anhui University of Techno. on March 9, 2015

XML Template (2015) [4.3.2015–5:20pm] [1–14]
//blrnas3.glyph.com/cenpro/ApplicationFiles/Journals/SAGE/3B2/PIJJ/Vol00000/150026/APPFile/SG-PIJJ150026.3d (PIJ) [PREPRINTER stage]

Zhang and Liu 7

Figure 4. Wear distributions of an ideal gear pair: (a) the pinion; (b) the gear.

different load conditions, in which the solid red lines increases monotonically with the wear cycles under
denote the wear of pinion and the dotted blue ones each kind of load level. However, this increment
represent the wear of gear. trend declined slightly when wear cycles getting
It can be observed from Figure 6 that the load larger. The explanation for this declination is that the
affects the wear depths of gears noticeably. Take the surface roughness decreases as wear accumulates,
pinion for instance, the maximum wear at Tp¼800 Nm resulting in a larger contact area and lower contact
after 60 million wear cycles is about 1.9 times of the pressures which in turn slows down the wear ratio.
one at Tp¼500 Nm and 16.7 times of the one at The simulation results drawn from Figures 5 and 6
Tp¼200 Nm. Meanwhile, the maximum wear depth show that the coupling effects of operating parameters

Downloaded from pij.sagepub.com at Anhui University of Techno. on March 9, 2015

XML Template (2015) [4.3.2015–5:20pm] [1–14]
//blrnas3.glyph.com/cenpro/ApplicationFiles/Journals/SAGE/3B2/PIJJ/Vol00000/150026/APPFile/SG-PIJJ150026.3d (PIJ) [PREPRINTER stage]

8 Proc IMechE Part J: J Engineering Tribology 0(0)

Figure 6. Effects of load and wear cycles on maximum wear

depths.

0.5
g

z
x y
Figure 5. Wear depths with respect to wear cycles: (a) the
pinion; (b) the gear.
p

of load level and wear cycles on gear wear must be

Figure 7. Misalignment between the pinion and the gear.
considered during a gear design aimed at slow wear
and long life-span.

through a LTCA. The pressure distributions are

Wear of spur gear with misalignment
then combined with the quasi-static wear model of
In this subsection, the effects of misalignment on the gears to predict the surface wear distributions of
surface wear of the spur gears are investigated. The gears p and g. For the sake of content limitation,
introduction of misalignment into the contacting the contact pressures vary along the contact cycles
gears is to represent the assembling and manufactur- for both no-misalignment and misalignment condi-
ing errors occur in actual system. Due to the content tions will not be demonstrated though they can be
limitation, only the angular misalignment of the gear easily computed. The following only demonstrate
pair is involved. The angular misalignment between the wear distributions of gears p and g with a mis-
the gears p and g is denoted as  as shown in Figure 7. alignment of  ¼ 2105 rad at Tp¼500 Nm after 136
As aforementioned, the introduction of misalign- million wear cycles.
ment into the gear pair changes the contact pressures By comparing Figure 8 with Figure 4, it can be
of the mating surfaces thus affects the surface wear of easily found that the wear distributions of both
gears. As can be seen from Figure 2(b), when  exists, gears are no longer uniform along their lead direc-
the contact pressures of tooth surfaces are no longer tions. The maximum wear depths of gears p and g
uniform along the lead direction, leading to stress increase 12.6% and 4.4% with the introduced mis-
concentrations at one edge of the teeth while the alignment, respectively. Hence, the misalignment not
other edges almost unloaded. only changes the distribution status of the gear wear
By introducing the misalignment into the 3D FE but accelerates the wear rates noticeably.
contact model of the spur gear pair, the contact pres- To reveal the effects of the misalignment  on sur-
sures of the engaging surfaces can be obtained face wear, a quantative relationship between the

Downloaded from pij.sagepub.com at Anhui University of Techno. on March 9, 2015

XML Template (2015) [4.3.2015–5:20pm] [1–14]
//blrnas3.glyph.com/cenpro/ApplicationFiles/Journals/SAGE/3B2/PIJJ/Vol00000/150026/APPFile/SG-PIJJ150026.3d (PIJ) [PREPRINTER stage]

Zhang and Liu 9

Figure 8. Wear distributions with misalignment: (a) the pinion; (b) the gear.

maximum wear depths of gears p and g and the values cycles. The variations of the maximum wear depths
of misalignment at Tp¼500 Nm and 136 million wear with respect to misalignments and input torques are
cycles is established through numerical simulations plotted as in Figure 10.
and shown in Figure 9. It is very clear that the interactions between the
Obviously, the surface wear of gear pair is strongly surface wear and the misalignment are much more
affected by gear misalignment . As  increases, the complicated when the effects of load (i.e. the input
maximum wear depths of gears p and g increase grad- torque) are considered. It can be observed from the
ually, with a sharper slope for the pinion p rather than above figure that the effects of input torque on the
the gear g. maximum wear depth are comparably remarkable
The following will explore the compound effects of than those of the misalignment. The reason may lie
misalignment and load on surface wear of spur gears. in that the increasing torque improves the contact
For clarity, the wear cycles are set as 136 million pressure and brings extra misalignments to the gear

Downloaded from pij.sagepub.com at Anhui University of Techno. on March 9, 2015

XML Template (2015) [4.3.2015–5:20pm] [1–14]
//blrnas3.glyph.com/cenpro/ApplicationFiles/Journals/SAGE/3B2/PIJJ/Vol00000/150026/APPFile/SG-PIJJ150026.3d (PIJ) [PREPRINTER stage]

10 Proc IMechE Part J: J Engineering Tribology 0(0)

(a)

20

/ µm
10

d m
(h )
p
0
2
0
20 4
15 6 m
10 /µ
Cp / µ 8 C

g
m 5
0 10

(b)

Figure 9. Effects of misalignment on maximum wear depths.

/ µm
3
d m
(h ) 2 0
g

1 2
20 4
15 m
6 /µ
10 C

g
Cp / µ 8
5
m
0 10

Figure 11. Effects of modification amounts on maximum

wear depths of the gear pair: (a) the pinion; (b) the gear.

thus to avoid uneven load distributions and to decel-

erate the wear of gears.

Modifications for spur gears with misalignment

Since the misalignment degenerates the contact
pressure distributions and then accelerates the surface
wear, one may naturally think about how to improve
the contact status. Among the methods for contact
pressure improvement, gear micro-geometry modifi-
cation is the most effective one. The involute crowning
modification is one of the most effective methods
applied in spur gears. Therefore, an involute crowning
modification strategy is adopted here to decelerate the
surface wear of the gear pair with a given misalign-
ment of  ¼ 2105 rad.
The following will explore the effects of modifica-
tion amounts on gear wear, based on which an opti-
mal modification amount can be determined. For this
purpose, two parameters of Cp and Cg are defined to
Figure 10. Effects of misalignment and load on maximum
denote the modification amounts of the pinion and
wear depths: (a) the pinion; (b) the gear.
the gear, respectively. The variations of the maximum
wear depths of the pinion with respect to the involute
system which in turn accelerates the surface wear. crowning modification  amounts  are depicted in
From this point of view, the misalignment between Figure 11, in which hpd m and hgd m represents the
the pinion and the gear must be strictly controlled maximum wear depths of the pinion and the gear at
during gear manufacturing and assembling stages the dedendum, respectively.

Downloaded from pij.sagepub.com at Anhui University of Techno. on March 9, 2015

XML Template (2015) [4.3.2015–5:20pm] [1–14]
//blrnas3.glyph.com/cenpro/ApplicationFiles/Journals/SAGE/3B2/PIJJ/Vol00000/150026/APPFile/SG-PIJJ150026.3d (PIJ) [PREPRINTER stage]

Zhang and Liu 11

Figure 12. Wear distributions after modifications: (a) the pinion; (b) the gear.

From Figure 11(a) it can be easily found that the negligible. Similarly, the maximum wear depth of
crowning modifications of the gear pair effectively the gear at the dedendum can be reduced to a very
decrease the wear depth of the pinion at the deden- small value with the proposed modification strategy.
dum in that the maximum value decreases from And the maximum wear depth of the gear is much
18.0 lm of unmodified gears to 4.61 lm of modified more sensitive to the modificaiton amounts when
gears with Cp¼5 lm and Cg¼6 lm. The maximum Cp<5 lm. After this value, the effects of crowning
wear depth of the pinion at the dedendum is very modification becomes very tirvial.
sensitive to the variations of Cp and Cg during the With the above discussions, it seems that an opti-
intervals of Cp¼[0, 5] lm and Cg¼[0, 6] lm. When mal modification amounts of Cp¼5 lm and Cg¼6 lm
the modification amounts exceed these intervals, the for the gear pair can be made to achieve a minimal
effectiveness of crowning strategy becomes very wear depths of gear surfaces. To validate the

Downloaded from pij.sagepub.com at Anhui University of Techno. on March 9, 2015

XML Template (2015) [4.3.2015–5:20pm] [1–14]
//blrnas3.glyph.com/cenpro/ApplicationFiles/Journals/SAGE/3B2/PIJJ/Vol00000/150026/APPFile/SG-PIJJ150026.3d (PIJ) [PREPRINTER stage]

12 Proc IMechE Part J: J Engineering Tribology 0(0)

property of the gear pair, not only an individual

(a)
gear. Therefore, it is necessary to consider the relative
values of modification parameters for a gear pair,
25 rather than the individual values for each gear.22
For this purpose, the summation of involute crown
20
values C ¼ CpþCg is introduced to denote the total
/ µm

15 involute crowning modifications.

d m

As known from the above analyses, the misalign-

(h )

10
p

5
ment affects the surface wear depth dramatically. It is
very clear from Figure 13 that the proposed modifi-
0 0
6 cation strategy can help to decelerate surface wear of
5
4 pinion and gear effectively when the summation of
χ ×1 2 10
0 −5
/ rad 0 15 / µm involute crowning values C remain within a certain
CΔ
range of C¼[0, 11] lm. When the modification
(b) amounts exceed this interval, the effectiveness of
crowning strategy becomes very negligible.
Furthermore, a larger C will be required to decrease
5
the wear depth of the gear pair with larger misalign-
4 ment values.
/ µm

From the above analyses, the summation of invo-

3
d m

lute crown values C for the spur gear pair must be

(h )
g

2
considered to determine the optimal modification
scheme when the gear pair has a given misalignment
1 0 amount. For instance, C¼13 lm is the determined
6
4
5 modification scheme for  ¼ 4105 rad. With this
χ ×1 2 10 scheme, the maximum deddendum wear depths of
0 −5
/ rad / µm  p
0 15 CΔ
pinion
 g and gear are h d m ¼ 4:91 mm and
hd m ¼ 1:55 mm, respectively. The modification
Figure 13. Variation of wear depths of the gears with respect scheme improves
 the wear of gears in this way that
to misalignment and modifications: (a) the pinion; (b) the gear. hpd m and hgd m are 75.4% and 65.1% lower than
those before modification.

effectiveness of the optimized modification param-

Conclusions
eters, the wear depths of the gear pair with a misalign- A surface wear prediction methodology for spur gears
ment of  ¼ 2105 rad after modifications are is proposed based on finite element method and
predicted. Figure 12 demonstrates the wear distribu- Archard’s formula to predict the surface wear of the
tions of the gears p and g with misalignment and invo- spur gear pair efficiently. The influences of load levels
lute crown modification after 136 million wear cycles. and wear cycles are investigated to reveal a quantita-
Comparing Figure 12 to Figure 8, it can be tive relationship between wear and operating param-
observed that with micro-geometry modifications, eters. This implies that the effects of load level and
the wear distributions of the gears p and g along the cycle number on gear wear must be considered during
lead direction are more uniform than those without a gear design aimed at slow wear and long life-span.
modifications, especially in the dedendum zone. The misalignment of the gear pair is defined and
Moreover, the crown modification makes the wear counted into the quasi-static model to investigate its
depth change smoothly along the profile direction. effects on gear wear. The results indicate that the
The wear depths of gears p and g are reduced signifi- introduction of misalignment may aggravate the
cantly in that the maximum wear amounts of gears p load distribution status and accelerate the surface
and g are 74.5% and 65.7% lower than those before wear. What’s more, the numerical simulation reveals
modification. Though the wear depths in the adden- that the wear depth distributes inconsistently when
dum zone near the pitch point increase slightly, the misalignment is counted in. Hence, it is highly recom-
amplitudes are much smaller than those in dedendum mended to control the misalignment amounts to
area. Therefore, it can be concluded that the proposed improve the load distribution and reduce the wear.
modification strategy is effective to improve pressure To compensate the negative effects of misalignment,
distribution status on the engaging surfaces and thus an involute crowning modification strategy is pro-
decelerates the gear wear. posed and its effectiveness on contact distribution
The following will explore the optimal modifica- improvement and wear deceleration are validated
tions for the spur gear pair with different misalign- through wear depths predictions. The simulations
ment amounts. Surface wear is affected by the manifest that the proposed modification strategy can

Downloaded from pij.sagepub.com at Anhui University of Techno. on March 9, 2015

XML Template (2015) [4.3.2015–5:20pm] [1–14]
//blrnas3.glyph.com/cenpro/ApplicationFiles/Journals/SAGE/3B2/PIJJ/Vol00000/150026/APPFile/SG-PIJJ150026.3d (PIJ) [PREPRINTER stage]

Zhang and Liu 13

help to decelerate surface wear effectively when the 16. Kahraman A and Ding H. A methodology to predict
summation of involute crowning values C remains surface wear of planetary gears under dynamic condi-
within a certain range. Furthermore, a larger C will tions. Mech Des Struct Mach 2010; 38(4): 493–515.
be required to decrease the wear depth of the gear pair 17. Yuksel C and Kahraman A. Dynamic tooth loads of
planetary gear sets having tooth profile wear. Mech
with larger misalignment values.
Mach Theory 2004; 39(7): 695–715.
18. Brauer J and Andersson S. Simulation of wear in gears
Funding with flank interference—a mixed FE and analytical
This work was jointly sponsored by National Natural approach. Wear 2003; 254(11): 1216–1232.
Science Foundation of China (Grant No. 51375013, 19. Bajpai P, Kahraman A and Anderson NE. A surface
50905122, 51405003) and Anhui Provincial Natural wear prediction methodology for parallel-axis gear
Science Foundation (Grant No. 1208085ME64). pairs. J Tribol 2004; 126(3): 597–605.
20. Park D and Kahraman A. A surface wear model for
hypoid gear pairs. Wear 2009; 267(9): 1595–1604.
Conflict of interest
21. Park D, Kolivand M and Kahraman A. Prediction of
None declared. surface wear of hypoid gears using a semi-analytical
contact model. Mech Mach Theory 2012; 52: 180–194.
References 22. Kahraman A, Bajpai P and Anderson NE. Influence of
1. Imrek H and Düzcükoğlu H. Relation between wear tooth profile deviations on helical gear wear. J Mech
and tooth width modification in spur gears. Wear Des 2005; 127(4): 656–663.
2007; 262(3): 390–394. 23. Hegadekatte V, Hilgert J, Kraft O, et al. Multi time
2. Choy FK, Polyshchuk V, Zakrajsek JJ, et al. Analysis scale simulations for wear prediction in micro-gears.
of the effects of surface pitting and wear on the vibra- Wear 2010; 268(1): 316–324.
tion of a gear transmission system. Tribol Int 1996; 24. Archard JF. Contact and rubbing of flat surfaces. J
29(1): 77–83. Appl Phys 1953; 24(8): 981–988.
3. Mackaldener M, Flodin A and Andersson S. Robust 25. Janakiraman V, Li S and Kahraman A. An investiga-
noise characteristics of gears due to their applications, tion of the impacts of contact parameters on wear coef-
manufacturing errors and wear. In: JSME International ficient. ASME J Tribol 2014; 136(3): 1–7.
conference on motion and power transmission, MPT
2001, Fukuoka, Japan, pp. 21–26.
4. Põdra P and Andersson S. Wear simulation with the Appendix
Winkler surface model. Wear 1997; 207(1): 79–85.
5. Flodin A and Andersson S. Simulation of mild wear in Notation
spur gears. Wear 1997; 207(1): 16–23.
6. Flodin A and Andersson S. Simulation of mild wear in a center distance
helical gears. Wear 2000; 241(2): 123–128. Cp modification amount of the pinion
7. Flodin A and Andersson S. A simplified model for wear Cg modification amount of the gear
prediction in helical gears. Wear 2001; 249(3): 285–292. C summation of involute crown values
8. Dhanasekaran S and Gnanamoorthy R. Gear tooth dwp pitch diameter of the pinion
wear in sintered spur gears under dry running condi- dwg pitch diameter of the gear
tions. Wear 2008; 265(1): 81–87. dbp base diameter of the pinion
9. Tunalioğlu MŞ and Tuç B. Theoretical and experimen- dbg base diameter of the gear
tal investigation of wear in internal gears. Wear 2014; dap addendum diameter of the pinion
309(1): 208–215.
dag addendum diameter of the gear
10. Park D, Kolivand M and Kahraman A. An approxi-
mate method to predict surface wear of hypoid gears dfp dedendum diameter of the pinion
using surface interpolation. Mech Mach Theory 2014; dfg dedendum diameter of the gear
71: 64–78. Ei (i ¼ 1,2) Young’s modulus of the mating
11. Karpat F and Ekwaro-Osire S. Influence of tip relief surfaces
modification on the wear of spur gears with asymmetric E0 equivalent Young’s modulus
teeth. Tribol Trans 2008; 51(5): 581–588. F tooth width
12. Kuang JH and Lin AD. The effect of tooth wear on the G dimensionless lubricant pressure–vis-
vibration spectrum of a spur gear pair. J Vib Acoust cosity coefficient
2001; 123(3): 311–317. h wear depth
13. Lin AD and Kuang JH. Dynamic interaction between
H hardness of the contact surface
contact loads and tooth wear of engaged plastic gear
k wear coefficient
pairs. Int J Mech Sci 2008; 50(2): 205–213.
14. Osman T and Velex P. Static and dynamic simula- K dimensionless wear coefficient
tions of mild abrasive wear in wide-faced solid spur L dimensionless load
and helical gears. Mech Mach Theory 2010; 45(6): m module of gears
911–924. p contact pressure
15. Ding H and Kahraman A. Interactions between non- q number of pressure update
linear spur gear dynamics and surface wear. J Sound Ri (i ¼ 1,2) radii of contact point of the two mating
Vib 2007; 307(3): 662–679. surfaces

Downloaded from pij.sagepub.com at Anhui University of Techno. on March 9, 2015

XML Template (2015) [4.3.2015–5:20pm] [1–14]
//blrnas3.glyph.com/cenpro/ApplicationFiles/Journals/SAGE/3B2/PIJJ/Vol00000/150026/APPFile/SG-PIJJ150026.3d (PIJ) [PREPRINTER stage]

14 Proc IMechE Part J: J Engineering Tribology 0(0)

Ri (i ¼ 1,2) roughness amplitudes of the two mating V worn volume

surfaces W applied normal load
Rc composite roughness amplitude W0 load per unit length
R0 equivalent radius of contact surface z (i ¼ 1,2) number of teeth
s sliding distance
 pressure–viscosity coefficient
sd normal circular tooth thickness
n pressure angle
S dimensionless composite roughness
 angular misalignment of the gear pair
amplitude
eq predetermined wear threshold
tI instant when the node first comes into
et maximum allowable wear threshold
contact zone
ti (i ¼ 1,2) Poisson’s ratios of the mating surfaces
tO instant when the nodes ceases contact
 density
Tp input torque
f number of wear cycles
v(t) relative sliding speed

Downloaded from pij.sagepub.com at Anhui University of Techno. on March 9, 2015

View publication stats