Mechanism and Machine Theory 59 (2013) 125–137

Contents lists available at SciVerse ScienceDirect

Mechanism and Machine Theory

journal homepage: www.elsevier.com/locate/mechmt

Studying the load carrying capacity of spur gear tooth flanks

M. Ristivojević a, T. Lazović a,⁎, A. Vencl b
a
Machine Design Department, University of Belgrade, Faculty of Mechanical Engineering, Kraljice Marije 16, 11120 Belgrade 35, Serbia
b
Tribology Laboratory, University of Belgrade, Faculty of Mechanical Engineering, Kraljice Marije 16, 11120 Belgrade 35, Serbia

a r t i c l e i n f o a b s t r a c t

Article history: The impact of load distribution in meshed teeth, teeth geometry and manufacturing accuracy
Received 19 October 2011 on wear of the spur gear tooth flanks is studied in this paper. The original geometry of teeth is
Received in revised form 20 August 2012 impaired wearing. Due to that, the load distribution is uneven, dynamic forces are increased,
Accepted 11 September 2012
and energy efficiency is decreased. The aim of the theoretical and experimental studies, carried
Available online 6 October 2012
out in this paper, is to reach more accurate model for the analysis of tooth flanks load carrying
capacity, taking into account a larger number of impacts on the tooth flanks stress state. A
Keywords: mathematical model for the contact stress during contact period is developed, depending on
Spur gear
the value and sign of base pitch difference of meshed teeth. The impact of pitch point's position
Wear
during the meshed teeth contact period was observed. In order to establish a correlation
Load distribution
Pitch error between the tooth flank failures and developed mathematical model, the appropriate
Contact stress experimental studies were carried out. The cemented spur gear pairs were examined on
back-to-back gear test rig under conditions of maximum operational load. The results are
presented in this paper.
© 2012 Elsevier Ltd. All rights reserved.

1. Introduction

Train of gears generates the noise and vibrations and requires a great manufacturing accuracy. Due to the compact construction,
high load carrying capacity and reliability, light weight per unit of transmitted power and small energy losses, gears have a wide range
of application in various industries. Increasingly difficult economic, energy and environmental conditions in the world, require harsh
operational conditions from the train of gears. In these operational conditions, the number of tooth flank failure forms has been
significantly increased. There are more than twenty different forms of tooth flank failures according to ISO 10825 [1]. The dominant
failures, i.e. failures that most commonly appear on the tooth flanks are adhesive wear, pitting and scoring. For that reason, these
failures are the subject of the largest number of studies [1–14]. The influence of the pitch point on the emergence of scoring has been
analysed by Imrek and Unuvar [2]. They found that conditions for scoring are the most unfavourable when pitch point is in the middle
of contact period of teeth. The most favourable condition for scoring is when pitch point is maximally far away from the middle of
contact period of teeth, i.e. when it is in the beginning or the end of tooth contact period. Based on the theoretical and experimental
investigation by Vereš et al. [3], it is shown that scoring resistance of tooth flanks can be significantly improved using a special
convex–concave profile in the plain tooth gears. Ristivojevic and Stefanovic analysed the influence of the pitch point's position of
teeth on the emergence of pitting during contact period [4]. It was demonstrated that parts of tooth flanks below the pitch surface are
much more susceptible to pitting than the parts of tooth flanks that are above the pitch surface. The impact of the intensity of spur
gear tooth flanks wear on the change of contact stress was analysed by Flodin and Andersson [5] on the base of the developed
numerical model. With the increase of tooth wear depth, the contact stress was changed due to the changes in the teeth profile
geometry. Hence, this change is different in different points of teeth profile. The same authors have developed a model for monitoring
the impact of tooth wear on tooth flank contact stress in cylindrical helical gear [6]. Larsson investigated the impact of oil type on the
thickness of oil layer, contact stress and coefficient of friction [7]. It was shown that the oil type has significantly higher impact on the

⁎ Corresponding author. Tel.: +381 62 295962; fax: +381 11 3370364.

E-mail address: tlazovic@mas.bg.ac.rs (T. Lazović).

0094-114X/$ – see front matter © 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.mechmachtheory.2012.09.006
126 M. Ristivojević et al. / Mechanism and Machine Theory 59 (2013) 125–137

thickness of oil film and value of friction coefficient than the contact stress. Experimental research on the impact of oil amount in the
lubrication system on the load tooth flanks carrying capacity is discussed by Höhn et al. [8]. In the field of small velocities, the oil
amount does not have an expressed influence on the intensity of tooth wear. With the velocity increase, the impact of oil amount on
tooth flanks wear is increased also. In addition, the impact gradient of oil amount on tooth wear is reduced when the velocity is larger
than 0.57 m/s. In the study conducted by Pedrero et al. [9], a model of non-uniform load distribution along the line of contact of spur
and helical gear teeth, obtained from the minimum elastic potential energy criterion, has been applied in the determination of the
highest contact stress. Also a new design of cylindrical spur gears is suggested by Imrek [10] in which the tooth is wider in the single
mesh zone than in the double mesh zone. By this design solution, the uniform load distribution along the teeth pair contact line during
the contact period is obtained.
In this paper, the impact of the accuracy of teeth manufacturing and teeth geometry on load carrying capacity of tooth flanks is
examined. The impact of manufacturing accuracy on the load carrying capacity of tooth flanks is discussed with respect to the
value and the sign of base pitch difference of meshed teeth. The three cases were analysed: 1) the base pitches of simultaneously
meshed teeth are equal, 2) the base pitch of driving gear is larger than base pitch of driven gear, and 3) the base pitch of the
driven gear is larger than the base pitch of driving gear. This analysis is conducted through the developed model of load
distribution in simultaneously meshed teeth. The impact of teeth geometry is observed through the tooth flank curve factor and
through the load distribution factor. In the case of the observed cylindrical tooth pairs, by choosing the appropriate coefficients of
profile shift, the position of pitch point during contact period was changed. In addition, the two cases were observed. The first case
is standard, i.e. the pitch point is in the zone of single mesh, and in the second case the pitch point is located in the zone of double
mesh. For this analysis, the appropriate mathematical models for the determination of the stress distribution on active surfaces of
meshed teeth were developed. In the numerical example, the impact of geometrical and kinematical values of gear pair and load
distribution on the tooth flank contact stress was studied. It was done by the application of the developed models. The
appropriate experimental studies in the laboratory conditions were also conducted. The two groups of gear pairs were examined
on a back-to-back gear test rig, in order to monitor the behaviour of tooth flanks under the conditions of maximum operation
load. The examined groups of gear pairs had the same geometric and kinematic values like the gear pairs used in numerical
example. The failures of tooth flanks were analysed with scanning electron microscope (SEM).
By comparative analysis of theoretical and experimental research, the correlation between the contact stress, load distribution,
sliding speed and failures of tooth flank contact surfaces is established.

2. Theoretical considerations

2.1. Load distribution in simultaneously meshed teeth

The first step in the analysis of gear pair teeth operational ability from the aspect of bending and surface strength, noise,
vibrations, efficiency, reliability and heating, is to determine the load distribution in simultaneously meshed teeth pairs. By
engaging a larger number of tooth pairs in load transfer, the operational ability of gear pair teeth is increased. In general, the load
distribution in simultaneously meshed teeth pairs is uneven. It means, that the simultaneously meshed teeth are differently
engaged in load transfer of the gear pair.
This variation occurs due to inevitable form and dimension deviation in the manufacturing of teeth and due to elastic deformations
of the teeth, body and crown of gears, shafts and supports. In addition, during the exploitation, due to the wear process, micro and
macro geometries of tooth flanks are changed. For that reason, it is very difficult to provide a high degree of uniformity of load
distribution in simultaneously meshed teeth pairs as well as immutability of distribution during the life cycle of gear pair.
Quantitative analysis of the participation degree of simultaneously meshed teeth pairs in transferring the total load of gear
pair is conducted here through the factors of load distribution (Fig. 1):

Fy
K αy ¼ ð1Þ
F

where Fy is the load transferred by the observed tooth pair and F is the total load of gear pair.
The interval of load distribution factor values is 0 ≤ Kαy ≤ 1.
For describing the load distribution, the set of simultaneously meshed teeth pairs is observed as statically indeterminate system.
From the conditions of elastic deformations and base pitch differences of simultaneously meshed teeth pairs, the appropriate
mathematical models of load distribution, at characteristic contact points of meshed teeth (Fig. 1b), are formed. The first moment of
the meshed teeth contact is between the dedendum of driving gear tooth (point E1) and the top of driven gear point A2 (Fig. 1a). In
this moment of contact, the participation level of meshed teeth pair can be determined by the expression:
!
1 cB1D2 ðpb1 −pb2 Þ
K αE1A2 ¼ 1− ð2Þ
1 þ ccB1D2 Ft
b
E1A2

where pb1 and pb2 are the real values of the base pitch of simultaneously meshed teeth pairs, b is the face width, cB1D2 and cE1A2
are the equivalent specific stiffness of the meshed teeth at contact points B1D2 and E1A2, respectively and Ft is the tangential force
in reference circle.
 M. Ristivojević et al. / Mechanism and Machine Theory 59 (2013) 125–137 127

The participation of simultaneously meshed teeth pair in load transmission at point B1D2 is determined by the expression:

K αB1D2 ¼ 1−K αE1A2 : ð3Þ

In the last moment of meshed teeth's contact, the top of driving gear touches at the point A1 the dedendum of driven gear at the
point E2. Load distribution factor in this moment of contact is determined according to the expression:
!
1 cD1B2 ðpb1 −pb2 Þ
K αA1E2 ¼ 1 þ ð4Þ
1 þ ccD1B2 Ft
b
A1E2

where cD1B2 and cA1E2 are the equivalent specific stiffness of meshed teeth at contact points D1B2 and A1E2, respectively.
Load distribution factor in simultaneously meshed teeth pair at point D1B2 is:

K αD1B2 ¼ 1−K αA1E2 :

According to the conventional calculation procedures [11], for teeth stiffness at characteristic contact points, the following
assumptions are introduced [14]:

cB1D2 ≈cD1B2
ð5Þ
cA1E2 ≈cE1A2 :

According to this assumption, when 1 b εα b 2, stiffness can be determined using the following expressions [11,14]:
4
cBD ¼ c
9 γ
4
cAE ¼ c
11 γ

where cγ is the medium equivalent specific stiffness of simultaneously meshed teeth pairs.

2.2. Influence of manufacturing accuracy and load magnitude on the teeth load distribution

Impact of the accuracy of teeth manufacturing on the load distribution in simultaneously meshed teeth pairs is analysed using
the value and the sign of base pitch difference of the meshed teeth. Manufacturing process of gears is exposed to the various
impact factors. Due to that, the variations of the real value of base pitch can be the consequence of systematic or random
deviations. For the analysis of teeth operational ability, it is very important to know the probability level at which the
implementation of the real value of base pitch can be expected. Statistical studies have shown that in a large-scale manufacturing,

a b

Fig. 1. a) Load distribution in simultaneously meshed teeth pairs; b) characteristic contact points in tooth profile.
128 M. Ristivojević et al. / Mechanism and Machine Theory 59 (2013) 125–137

the real values of machine parts obey the law of normal, Gauss's distribution. The base pitch differences are symmetrically
distributed with respect to the zero line, which limits nominal value pbn (Fig. 2).
In the case of meshed gears, the nominal values of base pitches are the same, but the tolerance, i.e. the deviation is different. It
depends on the value of reference circle and on the accuracy of teeth manufacturing. The base pitch difference of simultaneously
meshed teeth pairs varies in the following interval (Fig. 2):

0≤Δpb ≤Δpb max

where

Δpb max ¼ pb2 max −pb1 min :

According to this condition, the mean value of base pitch difference is:

0 þ Δpb max pb2 max −pb1 min

Δpbm ¼ ¼ :
2 2

The base pitch difference in meshed teeth does not affect the load distribution of meshed teeth pairs, only. The sign of that
difference is also very important, i.e. which pitch is higher. Quantitative analysis of the impact of load magnitude, value and sign
of base pitch difference of meshed teeth on the load distribution in simultaneously meshed teeth pairs is conducted on the
appropriate numerical example according to Eq. (2). The two groups of gear pairs (I and II) with real teeth are also considered.
Their geometrical and kinematical values are given in Table 1.
The three cases of the base pitch difference of meshed teeth are considered:
a) Δpb = 0
b) Δpb = Δpbmax
c) Δpb = Δpbm in accordance with [11].
The results of this analysis are shown by the diagrams in Fig. 3.
The value and the sign of base pitch difference of meshed teeth pairs have expressed impacts on load distribution at the small
load magnitudes (Fig. 3). This impact is reduced and tends to a constant value with increase of the load. This constant value
corresponds to the case when the real values of the base pitch of simultaneously meshed teeth pairs are equal. Its value does not
depend on the operational load magnitude, but only on the stiffness of meshed teeth.
In Fig. 3 for Ft/b = 600 N/mm and Δpb = 22 μm, it could be noticed that the pair of teeth in the first moment of contact will
transfer approximately twice higher load when pb2 > pb1 than in the case when pb1 > pb2. When Ft/b = 1600 N/mm, the load is
higher approximately by 25%.
By this analysis, the frequent occurrence of intensive tooth flank failure can be explained only on a limited number of teeth.
The value and the sign of base pitch difference of meshed teeth have the greatest impact on the tooth flank failure in case of gear
pairs with integer values of transmission ratio. In addition, the least favourable case occurs when the transmission ratio equals
one. Then, during the entire life cycle, the same pairs of teeth are always meshed. Accordingly, the greatest contact surface failures
will appear in teeth pairs with the greatest base pitch difference and the most unfavourable sign of that difference.

3. Mathematical value of the tooth flank contact stress

The teeth contact surfaces are not ideally smooth. They consist of asperities with different shapes and dimensions and the load
is transferred from one to another surface over the tops of asperities (Fig. 4a). The real contact surface is significantly different

Fig. 2. Distribution of real value of the base pitch of simultaneously meshed teeth pairs.
 M. Ristivojević et al. / Mechanism and Machine Theory 59 (2013) 125–137 129

Table 1
Geometrical and kinematical values of the analysed gear pairs.

Notation I II

Number of driving gear teeth z1 25 25

Number of teeth of driven gear z2 25 25
Coefficients of profile shift x1 0 0.6
Coefficients of profile shift x2 0 −0.6
Helix angle β 0º 0º
Nominal pressure angle αn 20º 20º
Pressure angle at pitch point αw 20º 20º
Face width, mm b 20/10 20/10
Centre distance, mm a 125 125
Module, mm m 5 5
Contact ratio εα 1.680 1.623
Manufacturing accuracy IT7 IT7
Sliding velocity, m/s vE1A2 3.74 1.78
vD1B2 0.71 1.00
vB1D2 0.71 2.68
vA1E2 3.74 5.45

from theoretical (ideal) contact surface. Surface pressure has a different value at each contact point (Fig. 4a). During the
operation, the micro geometry of contact surfaces, which are in relative motion, is constantly changing. The newly formed
asperities can be bigger or smaller than the asperities formed in the manufacturing process, depending on tribological conditions
in contact zone. Each new contact occurs on contact surfaces of another micro geometry that is directly reflected on the

Fig. 3. Load distribution factor vs. load, sign of base pitch difference and coefficient of profile shift.
130 M. Ristivojević et al. / Mechanism and Machine Theory 59 (2013) 125–137

distribution of surface pressure on meshed teeth contact surfaces. The deformation of the contact surfaces, lubricant viscosity changes
and hydrodynamic effect, according to the EHD theory, forms the pressure distribution in contact zone of meshed teeth (Fig. 4b).
This complex nature of the phenomenon of tooth flank surface and a large number of impact parameters have not provided
the formation of mathematical model that would include all the relevant aspects of surface failure. Based on numerous studies,
the appropriate standard procedures for the evaluation of operational ability of tooth flanks were formed [11,12]. In the case of
these procedures, the Hertzian stress was used to evaluate surface strength of tooth flanks. The other impacts were taken into
account through the appropriate correction factors.
In developed standard procedures, the impact of load distribution with simultaneously meshed teeth pairs on tooth flank
contact stress is analysed by the application of the corresponding approximate mathematical models. In order to monitor the
simultaneous impact of load distribution of both geometrical and kinematical values of gear pair on the load tooth flanks carrying
capacity, the appropriate mathematical models were developed.
According to the Hertz theory [12,13], the stress on the meshed tooth flanks in a moment of contact can be written in the
following form:
!1
F ny 2
σ Hy ¼ Z E ð6Þ
ρy by

where ZE is the elasticity coefficient of meshed gear material, Fny is the load transferred by the observed teeth pair and whose line
of action matches the teeth profile contact line, ρy is the mean radius of the imaginary cylinder curve in the observed moment of
contact of meshed teeth pair, by is the length of the contact line of the observed teeth pair (with spur gear pairs, when by = b).
After appropriate transformations, expression (6) can be written in the following form:
 1
K αy 2
σ Hy ¼ σ HC Z ρy ð7Þ
K αC

where σHC is the stress on tooth flanks at pitch point and Zρy is the curve radius factor.
A change of stress on the meshed tooth flanks during the contact period was monitored by the stress ratio between the stress
at contact point and the stress at pitch point:

 1
σ Hy K αy 2
¼ Z ρy : ð8Þ
σ HC K αC

Mathematical models of tooth flank contact stress in certain intervals of contact period (y/pb) are shown in Table 2, depending
on the kinematic pitch point (yC/pb).

3.1. Numerical example

Quantitative analysis of the load distribution impact and the teeth manufacturing accuracy on the stress condition of tooth
flanks was conducted on the appropriate numerical example, applying mathematical models shown in Table 2. The two groups of
spur gear pair (Table 1) were considered as in the previous example (Section 2.2).
This analysis is conducted for the corresponding load intensities of the considered gear pairs. The load represented as the
tangential force to the kinematic circle is 13,600 N in the first group of gears, and 10,000 N in the second one. These load

a b
Hertz

Real
Mean

Fig. 4. a) Impact of the asperities on pressure distribution in contact zone; b) pressure distribution in contact zone.
 M. Ristivojević et al. / Mechanism and Machine Theory 59 (2013) 125–137 131

Table 2
Mathematical models of contact stress in certain intervals of teeth mesh.

yC/pb − ymin/pb y/pb − ymin/pb σHy/σHC

1…(εα − 1) 0… (εα − 1) (Kαy/KαC)1/2Zρy

(εα − 1)…1 (1/KαC)1/2Zρy
1… εα (Kαy/KαC)1/2Zρy
(εα − 1)…1 0… (εα − 1) (Kαy)1/2Zρy
(εα − 1)…1 Zρy
1… εα (Kαy)1/2Zρy
1… εα 0… (εα − 1) (Kαy/KαC)1/2Zρy
(εα − 1)…1 (1/KαC)1/2Zρy
1… εα (Kαy/KαC)1/2Zρy

magnitudes are determined according to the teeth load carrying capacity (from the aspect of their bending strength, i.e. fatigue
fracture). The results of the analysis for both gear pair groups are shown in Table 3 and in Figs. 5 and 6.
Table 3 shows the numerical value of the load distribution factors and stress ratio on the tooth flanks, at characteristic contact
points and pitch point, for various values and signs of base pitch difference, and for an appropriate load magnitude.
Fig. 5 shows the changes of load distribution factor and contact stress during the contact period of meshed teeth. When Δpb b 0,
teeth pair in the domain 0 ≤ y/pb b (εα − 1) transfers a larger part of load than simultaneously meshed teeth pair in the domain
1 b y/pb b εα. In the case when Δpb > 0, teeth pair in the domain 0 ≤ y/pb b (εα − 1) transfers less load than simultaneously meshed
teeth pair in the domain 1 b y/pb b εα. When Δpb = 0, at the beginning of the domain 0 ≤ y/pb b (εα − 1), the teeth transfer smaller
load than simultaneously meshed teeth in the domain 1 b y/pb b εα. Afterwards, the engagement of all meshed teeth pairs becomes
equal at a certain moment of contact period (Kα = 0.5). This is the case of the uniform load distribution. After this distribution,
teeth pair in the domain 0 ≤ y/pb b (εα − 1) is more engaged in load transfer than simultaneously meshed teeth pairs in the
domain 1 b y/pb b εα. This various engagement of simultaneously meshed teeth pairs in load transfer causes various stress values
on tooth flanks.
In the first group of gear pairs (x1 = x2 = 0), the stress on tooth flanks is the highest in the single mesh (Fig. 5). In the field of
double mesh, it is the highest in the first (E1A2) and the last (E2A1) moments of meshed teeth contact. In this domain, the contact
stress depends on sign and value of the base pitch difference. With the reduction of base pitch difference, the impact of base pitch
difference sign of meshed teeth on the contact stress is also reduced (Fig. 5). When Δpb = 0, the contact stress does not depend on

Table 3
Results of the numerical analysis.

I group of gears (x1 = 0; x2 = 0; εα = 1.680; cγ = 18.9 N/(mm/μm); F = 13,600 N)

yi yi/pb Δpb Kαy σHy/σHC
mm Δpb = 10 μm Δpb = 22 μm Δpb = 10 μm Δpb = 22 μm
ymin = 8.978 0.608 =0 0.450 0.823
E1A2 >0 0.422 0.389 0.797 0.766
b0 0.478 0.511 0.849 0.878
ymin + (εα − 1)pb = 19.015 1.288 =0 0.550/1 0.746/1.006
D1B2 >0 0.522/1 0.489/1 0.727/1.0006 0.704/1.006
b0 0.578/1 0.611/1 0.765/1.006 0.786/1.006
ymin + pb = 23.739 1.608 =0 1/0.550 1.006/0.746
B1D2 >0 1/0.578 1/0.611 1.006/0.765 1.006/0.786
b0 1/0.522 1/0.489 1.006/0.727 1.006/0.704
ymin + εαpb = 33.776 2.288 =0 0.450 0.823
A1E2 >0 0.478 0.511 0.849 0.878
b0 0.422 0.389 0.797 0.766

II group of gears (x1 = 0.6; x2 = −0.6; εα = 1.623; cγ = 17.6 N/(mm/μm); F = 10,000 N)

yi yi/pb Δpb Kαy σHy/σHC
mm Δpb = 10 μm Δpb = 22 μm Δpb = 10 μm Δpb = 22 μm
ymin = 15.489 1.049 =0 0.450 0.973
E1A2 >0 0.415 0.373 0.968 0.961
b0 0.485 0.527 0.978 0.982
ymin + (εα − 1)pb = 24.685 1.672 =0 0.550/1 1.047/1.412
D1B2 >0 0.515/1 0.473/1 1.050/1.462 1.053/1.531
b0 0.585/1 0.627/1 1.045/1.366 1.043/1.317
ymin + pb = 30.250 2.049 =0 1/0.550 1.533/1.137
B1D2 >0 1/0.585 1/0.627 1.588/1.215 1.663/1.317
b0 1/0.515 1/0.473 1.483/1.065 1.430/0.983
ymin + εαpb = 39.446 2.672 =0 0.450 1.751
A1E2 >0 0.485 0.527 1.883 2.055
b0 0.415 0.373 1.627 1.487
132 M. Ristivojević et al. / Mechanism and Machine Theory 59 (2013) 125–137

Δpb= 10 μm
a
1,0 i=1 pb1<pb2 (Δpb<0)
Kα
x1 = 0; x2 = 0 pb1=pb2 (Δpb=0)
0,8
Ft = 1360 N pb1>pb2 (Δpb>0)
0,6

0,4

0,2

0,0 y/pb
2,0 10
v
σHy/σHC 8 m/s
1,5
6
1,0
4
0,5
2
v = f(y/pb)
C
0,0 0

y/pb
b Δpb= 22 μm
Kα 1,0 i=1 pb1<pb2 (Δpb<0)
x1 = 0; x2 = 0 pb1=pb2 (Δpb=0)
0,8
Ft = 1360 N
pb1>pb2 (Δpb>0)
0,6

0,4

0,2

0,0 y/pb
2,0 10
σHy/σHC v
8 m/s
1,5
6
1,0
4
0,5
v = f(y/pb) 2
C
0,0 0

y/pb

Fig. 5. Change of load distribution factor and contact stress during teeth mesh (I group of gears).

the sign of base pitch difference. In this case, the stress image of tooth flanks is symmetric with respect to the middle of teeth
contact period in which the pitch point C is located (Fig. 5).
In the second groups of gear pairs (x1 = 0.6; x2 = −0.6), tooth flank contact stress is the highest in the last contact moment of
meshed teeth E2A1 (Fig. 6). This highest stress is the result of unfavourable geometry of meshed teeth profile in the domain of contact
period and expressed impact of sign of base pitch difference. Due to favourable geometry of meshed teeth profile, the sign of base
pitch difference has a slight impact on stress condition of tooth flanks in the domain 0 ≤y/pb b (εα − 1).

4. Experimental research and wear mechanism analysis

Experimental research was conducted on the real samples — gear pairs, in the laboratory conditions. The cemented steel spur
gear pairs were examined on the back-to-back gear test rig. Geometrical and kinematical values of gear pairs are given in Table 1.
The tested gear material was steel 18CrNi8 (DIN EN 10027–1), depth of cemented layer was (0.8…1.0) mm, and hardness was
(60…62) HRC. During the research, gear pairs were lubricated with spray lubrication, in accordance with used operating speed,
using the monograde gear oil SAE 90.
 M. Ristivojević et al. / Mechanism and Machine Theory 59 (2013) 125–137 133

a Δpb= 10 μm
Kα 1,0 i=1
pb1<pb2 (Δpb<0)
x1 = 0.6; x2 = -0.6
0,8 pb1=pb2 (Δpb=0)
Ft = 1000 N
0,6 pb1>pb2 (Δpb>0)

0,4

0,2

0,0 y/pb
2,0 10
σHy/σHC v
8 m/s
1,5
6
1,0
4
0,5
2
v = f(y/pb)
C
0,0 0

y/pb

b Δpb= 22 μm

Kα 1,0 i=1 pb1<pb2 (Δpb<0)

x1 = 0.6; x2 = -0.6 pb1=pb2 (Δpb=0)
0,8
Ft = 1000 N pb1>pb2 (Δpb>0)
0,6

0,4

0,2

0,0 y/pb
2,0 10
v
σHy/σHC 8 m/s
1,5
6
1,0
4
0,5
v = f(y/pb) 2
C
0,0 0

y/pb

Fig. 6. Change of load distribution factor and contact stress during teeth mesh (II group of gears).

It is already known that in gear transmissions with transmission ratio larger than one, the surface distractions of driving gear tooth
flanks are more intensive than that of driven gear tooth flanks. This phenomenon is usually explained by the fact that teeth of small gear
in relation to the teeth of big gear have larger number of load changes. In order to eliminate these impacts on tooth flank failures, the
gear pairs with transmission ratio 1 (i=1) were examined. This transmission ratio is also the most unfavourable for teeth load carrying
capacity, in terms of their manufacturing accuracy. Test load of gear pairs was determined from conditions of teeth bending strength. In
order to make test conditions harsher from the aspect of tooth flanks load carrying capacity, gears were phase moved in axial direction
for the half of tooth width. In this way, the operational conditions of highly loaded power transmission were simulated in laboratory
conditions. At the same time, the conditions for comparative analysis of active and inactive parts of meshed tooth flanks were created.
In order to observe the impact of pitch point on tooth flank failure of the examined gear pairs, their position on meshed teeth
profiles was changed (Fig. 7). This was achieved by changing the form of meshed teeth profile, i.e. the selection of corresponding
coefficients of profile shift (Fig. 7). In the first group (I) of examined gear pairs, these coefficients were the same and equal to zero
(x1 = x2 = 0) (Fig. 7a). Fig. 7a shows the pitch point during the contact period in the middle of contact period, i.e. in the domain of
single mesh. In the second group (II) of the examined gear pairs, the coefficients of profile shift were different (x1 = 0.6; x2 =−0.6)
(Fig. 7b). Fig. 7b shows the pitch point during contact period in the domain of double teeth mesh (on the part of the driving gear tooth
dedendum, and on the part of the driven gear tooth addendum).
134 M. Ristivojević et al. / Mechanism and Machine Theory 59 (2013) 125–137

The contact of gear teeth between point E and the pitch point C, and between point C and point A is a mixture of rolling and
sliding. At the pitch point C there is pure rolling.
The contact length CA on the driving gear teeth flank is longer than the contact length CE on the driven teeth. In addition, the
contact length CE on the flank of the driving gear teeth is shorter than the contact length CA on the flank of the driven teeth. For
this reason, the wear is higher below than above the pitch point.
Typical failures on tooth flanks of the first group of gear pairs are shown in Fig. 8. These failures appeared after 50 working
hours with the load Ft = 10,000. The damages on tooth flanks of the second group of gear pairs are shown in Fig. 9. These failures
appeared after 300 working hours with the load Ft = 10,000 N. These operational conditions are determined based on tooth
bending strength.

4.1. Analysis of the wear mechanism and correlation with the mathematic model

After the tests, the condition of driving and driven gear contact surfaces was examined by visual inspection. The presence of
teeth breakage was not noticed, while on a certain number of teeth (app. 15–20%) pitting wear occurred on the dedendum
surfaces of both driving and driven gear teeth (Fig. 8). Surface failures of tooth flanks in the form of pitting were more intensive on
tooth flanks of driving gear (Fig. 8a) than on tooth flanks of driven gear (Fig. 8b). It is well known that this kind of damage appears
on the tooth dedendum, i.e. under the kinematic surface, due to the primary influence of the friction force direction, but not of the
contact stress [4]. For better wear mechanism analysis, the appearance of teeth without visually noticed pitting occurrence was
analysed at the end of the tests by SEM.
The SEM micrographs were taken on the worn surfaces, as well as on the unworn surfaces (before the test and after
machining) of both driving and driven gears. Only worn surfaces of the second group of gears were shown. The second group of
gear pairs was chosen for analysis because of less favourable operational conditions (due to specific geometry of meshed gear
teeth shown in Fig. 7b). Positions where the micrographs are taken on the tooth were: app. 1 mm from the A point (Fig. 7b), app.
1 mm from E point (Fig. 7b) and at top edge.
Driving gear addendum SEM micrographs show normal (running in) wear (Fig. 9a) with mild abrasion and mild adhesion
(Fig. 9b). Small holes that could be noticed are pores and could be noticed on unworn surfaces as well (Fig. 10). Unworn surfaces
of both gears (driving and driven) were the same, since the dimensions and machining were the same, as well. Worn surface of
the driving gear top edge shows adhesive wear appearance (Fig. 9c), and present pores and imperfections are more open. On the

a
Driving gear Driven gear

A A
Fµ Fµ
C C
Fµ Fµ

E E

x1 = 0 x2 = 0

b Driving gear Driven gear

A A Fµ
Fµ C

Fµ
Fµ
C
E E

x1 = 0.6 x2 = -0.6
Fig. 7. Friction force directions on the meshed tooth flanks and pitch point position during contact period in the case of: a) the first group and b) the second group
of examined gear pairs.
 M. Ristivojević et al. / Mechanism and Machine Theory 59 (2013) 125–137 135

Fig. 8. The appearance of the contact surface of tested materials — pitting (t = 50 h; Fn = 10,000 N): a) driving gear x1 = 0.6 and b) driven gear x2 = −0.6.

other hand, driving gear dedendum dominant type of wear was abrasive wear (Fig. 9d) which was more intensive than on the driving
gear addendum.
The SEM micrographs of driven gear addendum worn surfaces did not show the presence of abrasive wear (Fig. 11a). Dominant
type of wear was adhesive one with adhesive plates of deformed material and presence of wear debris caused by fracture,
accumulated into the adhesive wear pits (Fig. 11b). Driven gear top edge worn surfaces show similar appearance as in driving gear
(adhesive wear), but the worn surface zone was wider on driving gear compared to the driven gear (Figs. 9c and 11c). Driven gear
dedendum surface was more worn out, compared to both gears (driving and driven) addendum and dedendum worn surfaces. The
presence of the very intensive adhesive wear, with adhesive wear pits was noticeable (Fig. 11d).
Comparing the appearance of worn surfaces in the mesh points of driving and driven gears to their operating parameters (load
and sliding speed), the following conclusions can be reached:

• Wear of gear teeth dedendum (both driving and driven) was more intensive than wear of teeth addendum, as expected.
• Wear of the driven gear teeth was more expressed in relation to the wear of driving gear teeth, which is in accordance with the
results of theoretical and numerical analyses in this paper. The main reason for this is the extremely unfavourable geometry of
driven gear teeth (x2 = − 0.6). The unfavourable geometry has also affected the type of the present wear, so in the case of
driving gear teeth, the abrasion wear was dominant, and in the case of the driven one, the adhesive wear was present only.
• The width of worn areas on the gear top edges of both driving (Fig. 9c) and driven (Fig. 11c) gears is different. At the driven gear
teeth, this area is narrower due to smaller sliding path — pitch point is closer to the top edge area of the gear teeth (Fig. 7b).

Fig. 9. SEM micrographs of the driving gear worn surface (t=300 h; Fn =10,000 N): a) and b) addendum, c) top edge and d) dedendum (friction force directions are
denoted with arrows).
136 M. Ristivojević et al. / Mechanism and Machine Theory 59 (2013) 125–137

Fig. 10. SEM micrographs of the driving gear unworn surfaces: a) addendum and b) dedendum.

• The unfavourable geometry of the driven gear tooth is mostly pronounced and could be noticed compared to the appearance of
worn surface of driving gear tooth addendum (point A1 in Fig. 7b) and worn surfaces of driven gear tooth dedendum (point E2
in Fig. 7b). Although the specific loads (contact stresses) were in the same domain (1981…2737) N/m 2 (Fig. 7b), wear at the
point E2 (Fig. 11d) was more intensive than at the point A1 (Fig. 9b). Specific sliding speed at the point E2 (13.9) was higher than
at the point A1 (0.93), as well as in relation to the other two points. Such high specific sliding speed is the main reason for the
presence of very intensive adhesive wear with adhesive wear pits at the point E2 (Fig. 11d).

The experimental results show that there is a correlation between tooth flank failures and developed mathematical model for
contact stress during contact period, with respect to the value and sign of base pitch difference, and pitch point position.

5. Conclusion

Adhesive wear is always present on the tooth flanks, mostly in the addendum and dedendum areas. Geometrical and kinematical
values of the gear pair teeth are unfavourable in these areas from the aspect of contact stress magnitude and sliding speed values
(Figs. 5 and 6). Simultaneously, when the teeth feet and head are in contact, the total load of gear pair is distributed on the larger
number of meshed teeth. In addition, their level of engagement depends on the accuracy of manufacturing (Fig. 3), stiffness of teeth
and magnitude of total load. On the teeth addendum and dedendum, apart from the reduced load magnitude, the adhesive wear is
mostly pronounced. Its intensity primarily depends on the tribological conditions and contact stress. The original geometry and

Fig. 11. SEM micrographs of the driven gear worn surface (t=300 h; Fn =10,000 N; x2 =−0.6): a) and b) addendum, c) top edge and d) dedendum (friction force
directions are denoted with arrows).
 M. Ristivojević et al. / Mechanism and Machine Theory 59 (2013) 125–137 137

proper teeth mesh are impaired by adhesive wear. These phenomena are reflected in the lower efficiency, larger variation of load
distribution and higher dynamic forces. Uniformity and intensity of tooth flanks wear depend on pitch point position on the profiles of
meshed teeth (Fig. 7).
Due to the nature of teeth mesh, parts of tooth flanks below pitch surface are more susceptible to surface failures than parts of
tooth flanks above pitch surface. These parts can be protected from intensive failure by proper harmonisation of geometrical and
kinematical values of teeth, load magnitude and accuracy level of teeth manufacturing. Mathematical models developed in this paper
can establish the optimal correlation between loads, manufacturing accuracy and teeth geometry in order to achieve higher tooth
flanks load carrying capacity. The theoretical studies show that the sign of base pitch difference and the pitch point position during
contact period have strong impacts on the tooth flanks load carrying capacity. It is shown, by the experimental research of gear pairs,
that in the highly loaded gear power transmissions, dominant wear types are pitting and adhesive wear, while abrasion wear is less
pronounced.

Acknowledgement

This work has been performed within the projects TR 35029, TR 35043, TR 35011, TR 34028, and TR 35021. These projects are
supported by the Republic of Serbia, Ministry of Education, Science and Technological Development, which financial help is
gratefully acknowledged.

References

[1] ISO 10825, Gears – Wear and Damage to Gear Teeth – Terminology, 1995.
[2] H. Imrek, A. Unuvar, Investigation of influence of load and velocity on scoring of addendum modified gear tooth profiles, Mechanism and Machine Theory 44
(2009) 938–948.
[3] M. Vereš, M. Nemčeková, A. Marinković, Tooth flanks scoring resistance of noninvolute teeth profiles in plane toothed cylindrical gears, FME Transactions 37
(2009) 103–106.
[4] M. Ristivojevic, N. Stefanovic, Research of influence teeth geometry and position of the pitch point on the fatigue the tooth flanks of cylindrical gears, in:
Proceedings of 7th Yugoslav Tribology Conference, Belgrade, 2001, pp. 4-13–4-18.
[5] A. Flodin, S. Andersson, Simulation of mild wear in spur gears, Wear 207 (1997) 16–23.
[6] A. Flodin, S. Andersson, A simplified model for wear prediction in helical gears, Wear 249 (2001) 285–292.
[7] R. Larsson, Transient non-Newtonian elastohydrodynamic lubrication analysis of an involute spur gear, Wear 207 (1997) 67–73.
[8] B.-R. Höhn, K. Michaelis, H.-P. Otto, Minimised gear lubrication by a minimum oil/air flow rate, Wear 266 (2009) 461–467.
[9] J.I. Pedrero, M. Pleguezuelos, M. Munoz, Critical stress and load conditions for pitting calculations of involute spur and helical gear teeth, Mechanism and
Machine Theory 46 (2011) 425–437.
[10] H. Imrek, Width modification for gears with low contact ratio, Meccanica 44 (2009) 613–621.
[11] ISO 6336‐1, Calculation of Load Capacity of Spur and Helical Gears — Part 1: Basic Principles, Introduction and General Influence Factors, 2007.
[12] ISO 6336‐2, Calculation of Load Capacity of Spur and Helical Gears — Part 2: Calculation of Surface Durability (Pitting), 2007.
[13] M. Huirong, On problems of surface pressure in gear load-carrying capacity calculation, Proceedings of the International Symposium on Gearing & Power
Transmissions (1981) 407–412 (Japan Society of Mechanical Engineers, Tokyo).
[14] M. Ristivojevic, R. Mitrovic, Load Distribution — Gears and Rolling Bearings, Faculty of Mechanical Engineering, Belgrade, 2002.