Measurement 76 (2015) 3244

Contents lists available at ScienceDirect

Measurement
journal homepage: www.elsevier.com/locate/measurement

Assessment of surface contact fatigue failure in a spur geared

system based on the tribological and vibration parameter
analysis
M. Amarnath a, Sang-Kwon Lee b,
a

Department of Mechanical Engineering, PDPM-Indian Institute of Information Technology Design and Manufacturing, Jabalpur, Jabalpur 482001, India
Acoustics & Vibration Signal Processing Laboratory, Department of Mechanical Engineering, Inha University, 253 Yonghyun-Dong, Nam-Gu, Incheon
402-751, Republic of Korea
b

a r t i c l e

i n f o

Article history:
Received 14 January 2015
Received in revised form 13 June 2015
Accepted 12 August 2015
Available online 28 August 2015
Keywords:
Gear
Lubricant
Spectroscopy
Contamination
Vibrations

a b s t r a c t
Gears are one of the most common mechanisms for transmitting power and motion and
their usage can be found in numerous applications. Studies on gear teeth contacts have
been considered as one of the most complicated applications in tribology. The changes
in operating conditions such as increase in temperature, load, reduction in viscosity result
decrease in lubricant film thickness and degradation of lubricating oil thereby triggering
several types of failures on tooth surfaces viz. pitting, scuffing, micro pitting, scoring,
and spalling, these faults influence changes in vibration signals. This paper presents the
results of experimental investigations carried out to assess wear in spur gears of single
stage spur gear box under fatigue test conditions. The studies considered the lubricant film
thickness analysis, wear mechanism studies on gear tooth surfaces, oil degradation analysis using Fourier transform infrared radiation (FTIR) method along with vibration signal
analysis. The results provide a good understanding of tribological and vibration parameters
as measures for effective assessment of wear in spur gears.
2015 Elsevier Ltd. All rights reserved.

1. Introduction
Gears are known to be the essential and most efficient
mechanical components in many machinery applications
for power and motion transmission. Lubricating oil is used
in a geared system to reduce friction and wear by interposing a film of oil between gear teeth, it plays a significant
role in minimizing propagation of wear damage on gear
tooth surfaces. In order to maintain a proper lubrication,
it is important not only to use lubricant with suitable properties but also to monitor and analyze operating conditions
viz. lubricant film thickness, specific film thickness, and
degradation of lubricants in a periodic manner [13]. Oil
Corresponding author. Tel.: +82 032 860 7305.
E-mail address: sangkwon@inha.ac.kr (S.-K. Lee).
http://dx.doi.org/10.1016/j.measurement.2015.08.020
0263-2241/ 2015 Elsevier Ltd. All rights reserved.

film between working surfaces of gear teeth surface

influences strongly the operating performance of a geared
system. Gear failures such as wear, scuffing, micro pitting,
spalling, and scoring are influenced by the lubricant
parameters such as film thickness, viscosity, and temperature. High temperature leads to low viscosities which
causes reduction in lubricant film, if the lubricant film on
gear teeth fails, the aforementioned failures take place
thereby altering noise and vibration levels [46].
Karacay and Akturk [7] conducted fault detection
experiments on ball bearings using conventional vibration
signal analysis. Statistical parameters of vibration signals
such as peak-to-peak, crest factor and kurtosis values were
considered to investigate formation and development of
localized defects in a ball bearing. Spectrum analysis of
vibration signals revealed defect locations on inner race

M. Amarnath, S.-K. Lee / Measurement 76 (2015) 3244

33

Nomenclature
a
E
E0
G
hm
Hm
HRB
L

contact half width

Young modulus
dimensionless
elasticity
h
2
i1
1v 22
0
1 1v 1
E 2 E1  E2

modulus,

dimensionless parameters of material

oil film thickness
dimensionless oil film thickness
Brinell hardness
face width of the gear

and outer race and balls. Microscopic analysis of rolling

contact surfaces showed the details of wear mechanisms
occurred during the test period. Itagaki et al. [8] presented
a study on the influence of grease degradation on the
vibration levels in a ball bearing operated at allowable load
condition. The vibration signal analysis was carried out on
the ball bearings lubricated with three types of greases viz.
Li soap/silicon grease, Na soap/mineral oil grease and Li
soap/mineral oil grease. Results obtained from the experiments showed that the abnormal vibrations occurred in
the ball bearings lubricated with all types of greases. Li
soap/silicon oil grease and Na soap/mineral oil grease
resulted in abnormal vibrations when the grease acts on
the vibratory system as a negative damping moment
during the normal vibration. However in ball bearing
lubricated with Li soap/mineral oil grease, the abnormal
vibration occurs due to a positive damping moment of
the grease which gradually decreases over a time period
and it acts on the vibratory system.
Peng and Kessissoglou [9] have considered wear particle and vibration spectrum analysis techniques to detect
faults in a worm geared system. Experimental investigations carried out in this work include three operating
conditions viz. lack of proper lubrication, normal operating
condition and contaminant particles added into the lubricating oil. Integration of wear particle and vibration spectrum analysis in gear fault detection resulted in more
reliable assessment of wear on gear tooth surfaces. Wear
debris analysis revealed wear rate and wear mechanism
on the gear wheel, whereas vibration signal analysis provided quick and reliable diagnostic information on bearing
faults in the worm gearbox.
Tan et al. [10] carried out experiments to diagnose
faults in a back-to-back power recirculation type spur
gearbox. Pitting fault was intentionally simulated on a spur
gear tooth, this fault was detected using fast Fourier
transform (FFT) of acoustic emission signals. Authors have
considered operating conditions such as temperature,
viscosity and lubricant film thickness to assess fault
propagation on gear teeth. Results obtained from acoustic
signal analysis in conjunction with lubricant film thickness
analysis appeared to be promising diagnostic tools to
detect developing faults in spur gears. Elforjani and Mba
[11] described results of fault detection investigations in
a spur geared system in which natural pitting was allowed
to occur. Throughout the test period, acoustic emission,

P0
U
W
W

a
l0
t

Ra
Rz

Hertzian pressure
velocity dimensionless parameter
load normal to tooth
dimension less load parameter
pressure viscosity coefficient
dynamic viscosity
Poissons coefficient
average surface roughness
peak-to-valley surface roughness

vibration and metal contents in oil samples were

monitored continuously to correlate and compare these
techniques with the natural degradation of gears. The
authors observed that the rate of change of three parameters i.e. Fe concentration, acoustic emission and vibration
root mean square (RMS) values with respect to gearbox
operating time increased with increasing the applied torque. Serrato et al. [12] conducted experimental studies to
characterize the vibration behavior of roller bearing and
lubricant film thickness. The bearings were operated under
oil lubrication with radial load condition, the load considered in this experiment was about 10% of the allowable
load of the bearing. Results obtained from the experiments
highlighted the relationship between RMS values of vibration signals and specific film thickness, the trend plot of
RMS versus specific film thickness showed a good correlation with Stribeck curve.
Amarnath and Praveen Krishna [13] carried out experiments to analyze the effects of reduction in film thickness
on the vibration signals in a spur gearbox. Vibration signals
were acquired under full film (FL), elastohydrodynamic
lubrication (EHL) and boundary lubrication (BL) conditions
were processed using conventional signal processing techniques, kurtosis parameters were extracted from vibration
signals to detect surface fatigue wear on gear tooth surfaces. Further, fault diagnostic information obtained by
these results was enhanced by using ensemble empirical
mode decomposition (EEMD) method. Akagaki et al. [14]
carried out fault diagnosis experiments on deep groove ball
bearing using wear debris and vibration signal analysis
methods. Wear debris morphology viz. thread like debris,
cutting chip debris and plate like particles reveals wear
mechanism developed on the rolling contact surfaces.
Authors have highlighted the advantages of vibration and
wear particle analysis methods to detect surface fatigue
wear damage in deep groove ball bearings.
Hamzah and Mba [15] conducted experiments to assess
wear damage in a helical geared system by considering the
influence of different speeds and loads using acoustic
emission (AE) signals. Specific lubricant film thickness
analysis was also considered in conjunction with AE signal
analysis. The RMS values of AE signals increase in nine- and
fourfold changes which provide the exact lubrication condition in the operating geared system. Zakharich et al. [16]
carried out experimental investigations to determine
service deterioration of transformer oil using Fourier

34

M. Amarnath, S.-K. Lee / Measurement 76 (2015) 3244

transform infrared (FTIR) spectroscopy method. Absorption bands in the region of 17002000 cm1 and
33004000 cm1 were considered to obtain the details of
oxidation reactions in the transformer oil. Kumar et al.
[17] carried out experimental studies to analyze the depletion of additives in engine oil. Authors have considered
various oil analysis methods viz. pH measurement, FTIR
analysis and ultraviolet and visible ray spectrum analysis.
Results obtained from the above mentioned methods
provide useful information related to acidic region of oil,
depletion of corrosion inhabitance and antiwear agents.
Al-Ghouti and Al-Atoum [18] made experimental observations to investigate the differences occurring between
the original and recycled oils. Atomic absorption spectrometry, inductive coupled plasma (ICP) and FTIR oil analysis
techniques were used to analyze depletion in physical
and chemical properties of lubricants. The results obtained
were found to be useful to identify the original and
recycled oils. Analysis of the metal concentration and concentration of oxidation products revealed appropriate features for differentiating particular oil samples from the
other oil samples.
Amarnath et al. [19] discussed the experimental results
on the measurements of reduction in the gear teeth stiffness in conjunction with statistical parameter analysis of
vibration signals. Experimental measurement of stiffness
was carried out using modal analysis along with a theoretical model. Authors have concluded that the stiffness
measurement exhibits a direct relationship with the
propagation of surface fatigue wear and the increase in
vibration amplitude of the spur geared transmission
system.
Amarnath et al. [20] conducted experimental studies to
assess wear in a spur geared system using lubricant film
thickness analysis, wear particle analysis based on their
morphology and metal composition analysis of lubricants
using energy dispersive spectrometry (EDS). Authors have
correlated these results with wear propagation on gear
tooth surfaces. Sari et al. [21] carried out experimental
investigations to study the effect of presence of solid particles in the lubricants and their effects on gear teeth damage in a spur geared system. Results highlighted the
development of abrasive wear mechanism on gear tooth
surfaces with high rate of sliding. Measurement of gear
tooth thickness at the root, pitch and tip zones was also
included in the investigations. High reduction in tooth
thickness was found at the root due to high sliding velocity
and the presence of silica particles. Further, as a scope of
future work, the authors suggested to consider the analysis
of vibration origins in a geared system.
In the present work, experimental investigations on
gear damage detection and assessment in line with the
work of Sari et al. [21] have been carried out. The work
comprises of fatigue tests conducted on a spur geared system over a period of 1200 h. The study focuses on wear
assessment using lubricant film thickness analysis along
with lubricant degradation and vibration signal analysis
of a spur geared system. Physical observation of surface
wear propagation and corresponding tooth thickness
reduction measurements were also considered to obtain
additional fault diagnostic information. Results obtained

from the experimental investigations offer a good potential

for surface fatigue wear assessment in a spur geared
system.
2. Experimental setup
The aim of the experiment was to conduct fatigue tests
in a spur geared system and to develop a reliable condition
monitoring system to detect and diagnose faults in a single
stage spur gearbox operating under various lubrication
regimes. Experimental setup used in this study is shown
in Fig. 1. The arrangement consists of two parallel steel
shafts and two gears: a pinion with 24 teeth and a gear
with 25 teeth, this gear pair had a module of 3 mm and
pressure angle of 20. Table 1 gives the other specifications
of the test rig.
The spur geared pair is driven by a 0.75 kW variable
speed D.C motor, the output of the gear is given to eddy
current braking system. The D.C motor, gearbox and eddy
current brake are connected via a rubber coupling. The
motor rotation is controlled by a speed controller which
allows the geared system to operate at different speeds.
A dip type lubrication has been used to lubricate the gear
pair. Experiments are performed under a constant load of
12 N m. The experimental setup with equipment and
sensors is shown in Fig. 2.
2.1. Experimental procedure
The spur gearbox was kept run in for over 30 h, and
then it was considered to conduct the experiments. Fatigue
test experiments were carried out for 8 h a day, the tests
were continued over a period of 1200 h, the vibration signals were acquired during regular intervals at a load of
12 N m, the speed was set to 450 rpm. The vibration
signals were post processed in a MATLAB 2008 to obtain
vibration spectra and fault related statistical features. On
the other hand, experiments were paused after every
200 h for the following purposes:
(i) To collect lubricant samples from the gearbox casing, these samples were used to analyze oil degradation using FTIR method.
(ii) To measure reduction in chordal gear tooth thickness occurred due to surface fatigue wear.
(iii) To measure surface roughness on gear tooth
surfaces.
(iv) To observe the development of wear mechanism on
gear tooth surfaces using USB micro scope.
2.2. Instruments and sensors
The vibration signals were acquired using a Bruel &
Kjaer 4506 tri-axial accelerometer. The accelerometer
was mounted on the input shaft bearing. A commercial
data acquisition system LMS SCADASIII was used to
acquire the vibration signals, these signals were sampled
at 8.2 kHz. This frequency range is sufficient to reveal
frequency content of vibration signals approximately up
to the fifth tooth mesh harmonic frequency. Each time

35

M. Amarnath, S.-K. Lee / Measurement 76 (2015) 3244

Power supply

Gear

Driven sha

D.C Speed
controller

D.C load
controller

Eddy
current
brake

D.C motor

Accelerometer
Driving sha

Pinion

Data Acquision

Fig. 1. Experimental setup.

Table 1
Gear test rig specifications.
Parameter

Pinion

Gear

Number of teeth
Deport angle
Pressure angle
Height
Module
Face width
Pitch diameter
Diameter of base
Diameter of head
Pinion speed
Torque on pinion shaft

24
0.00
20
6.53
3
30
72.72
68.34
78.78
450 rpm
012 N m

25
0.00
20
6.53
3
30
75.75
71.18
81.81

Material properties of gears

Material
Brinell hardness number hardness
(HRB)
Poissons ratio
Youngs modulus
Viscosity

SM 45 C
167
0.3
2  105 N/mm2
Lubricant oil SAE
40

domain signal consists of the average of 16 time history

plots which seemed to be sufficient to reduce the noise
in the vibration signals. The vibration data were acquired
and stored in a personal computer for post processing.
The oil bath temperature is measured at every half
hour, as temperature rises the viscosity decreases, the
decrease in viscosity values has been obtained from ASME
viscosity temperature chart and these values were used to
estimate lubricant film thickness. A USB port micro scope
was used to obtain the images of wear mechanisms developed on the gear tooth surfaces. A HommelEtamic (W5)

roughness measuring instrument (Perthometer) was used

to measure surface roughness on the gear tooth. The
reduction in tooth thickness was measured using a digital
vernier caliper which has a resolution of 0.01 mm. Fatigue
test experiments were carried out for over 1200 h. The
vibration signals along with tribological parameters such
as minimum lubricant film thickness, lubricant degradation and reduction in tooth thickness were considered to
diagnose fault propagation on the gear tooth surfaces.
3. Lubricant film thickness and assessment of oil
degradation
3.1. Lubricant film thickness analysis
Fig. 3 shows a gear pair in meshing position in a dip
type lubrication. The primary function of liquid lubricant
is to minimize friction, wear and surface damage over
the intended life of a mechanical system that contains
machine elements such as gears, bearings, and cams. Secondary function is to prevent corrosion and for scavenging
heat, dirt and wear debris [1]. Gear teeth contacts have
complex combination of sliding and rolling, which vary
along the profile of each tooth as depicted in Fig. 3. In
general, gears are very often operated under elastohydrodynamic lubrication (EHL) in full, mixed or boundary film
regimes, where the viscosity of the lubricant plays an
important role in terms of both film thickness and coefficient of friction.
In the addendum, the direction of rolling and sliding is
same which results positive sliding condition in the
addendum region, where as in the dedendum the direction
of rolling is opposite to that of sliding thereby giving rise to

36

M. Amarnath, S.-K. Lee / Measurement 76 (2015) 3244

K- type thermocouple

Eddy current brake

and indicator
D.C motor

Accelerometer
Spur gearbox

Fig. 2. View of the test rig showing sensors and equipment.

Fig. 3. Combination of sliding and rolling in gear teeth and worthy contacting points on gear and pinion teeth [21,22].

negative sliding condition. Hence, the contact fatigue is

more likely to initiate in the dedendum. The surface fatigue
wear in the dedendum region is more severe which causes
gear failures viz. pitting, scuffing, spalling, and scoring.
[22,23].

In the elastohydrodynamic lubrication regime a very

high contact pressure occurring on rolling contact surfaces
causes elastic deformation on the gear tooth surface which
causes origin to small elliptical contact areas. The repetitive
formation of the elastically deformed contacts eventually

37

M. Amarnath, S.-K. Lee / Measurement 76 (2015) 3244

leads to surface fatigue [2,6]. Alternative methods to predict gear failure are often related to film thickness and
specific film thickness, which provide an indication of contact severity or the possibility of lubricant breakdown. The
analysis of film thickness between lubricated contacts is
very complex. It involves two rough surfaces in relative
motion separated by a thin lubricant film. The lubricant
film between gear teeth contact is subjected to high contact
pressure and sliding, due to difference in surface velocities
and increase in temperature, lubricant undergoes a change
in its physical properties [1,2,5,11,12]. Under typical operating conditions, the lubricant film separating the contact
surfaces is very thin, usually of the same order of magnitude as the surface roughness, which may cause breakdown
of the lubricant film. Hence, this drastic reduction in lubricant film thickness is responsible for gear failure modes
such as micro pitting, macro pitting, spalling, scoring, and
scuffing. The thickness of lubricant film depends on the
lubricant properties, the geometry of bearing surfaces and
operating conditions [2,4,5].
The oil film thickness hm is calculated using Dowson
Higginson formula which is given by [6,21]:

Fig. 3 shows a gear pair in meshing position in a dip

type lubrication, lubricant film thickness can be estimated
by considering geometry and operating conditions of the
gearbox. Worthy contacting points are indicated on the
line of action T1 T2.
The following code has been considered to represent
the location of tooth profile.

h m H m  Rx

0:6
0:11
Hm 0:985  U 0:7
  G  W 

In many industrial applications, lubricant oils are subjected to high operating temperature, surface fatigue and
severe stresses. Over extended operation, the lubricant oil
undergoes mechanical and thermal degradation which
alters the properties of the lubricant. Degradation of lubricant causes formation of new gases (CO, CO2), volatile
hydrocarbons, liquids (aldehydes, ketones, alcohols, acids,
eather, resins), solids (asphaltes and carbenes) and chemical products, these changes led to deplete the properties of
antioxidation additives, and formation of oxidation products. The presence of impurities and oxidation products
result a negative impact on the normal working regimes
of machine elements [17,27,28].
FTIR analysis provides the diagnostic information of the
chemical processes taking place in the lubricant which renders possibilities for both qualitative and quantitative
analyses with the sensitivity of the method between 0.5%
and 1% [25]. FTIR spectrum analysis reveals information
about specific chemical bonds and functional groups present in the investigated sample.
Covalent chemical bonds in the molecules absorb infrared radiation at the characteristic wavelengths, which
depend on the atomic composition of the molecule and
the strength of the chemical bond between atoms. Lubricant contamination such as water and glycol can be analyzed in the infrared by detection of the AOH stretch for
water and the ACAO stretch for glycol. Degradation due
to the oxidation is detected as the carbonyl, i.e. C@O
stretch from the formation of organic acids and conjugated

where hm is oil film thickness in micrometer, Hm is dimensionless oil film thickness. G*, U*, W* represent dimensionless parameters of material, velocity and load respectively,
which are given by:

G a  E0

U  l0  u1 u2 =E0  Rx

W  W=E0  Rx  L

where l0 , E , Rx represent dynamic viscosity, Youngs modulus and radius of relative curvature and u1, u2 represent
peripheral velocity of gear and pinion respectively.
The maximum Hertzian pressure P0, for a linear contact
is calculated by:
0

P0


a

2W

paL
8  W  Rx
p  L  E0


7

where a, L represent contact half-width and face width of

gear respectively.
In meshing of gears, lubricant film thickness decreases
as a result of quick change in normal tooth load, which
arises with decrease in viscosity due to increase in temperature of lubricant film, discontinuous meshing or pitch
error eventually trigger failure in gear tooth contacting
surfaces. Hence continuous wear failures can occur under
thin separating lubricant film conditions, typically at slow
pitch line velocities, where asperity interaction occurs
which leads to material removal from mating surfaces
thereby influencing strongly the operating performance
with regard to the noise and vibration [2,6,10,24].

P: Start of the active profile

Q: Lowest point of single tooth contact
O: Pitch point
R: Highest point of single tooth contact
S: End of active tooth profile
For better understanding of EHL contact operation it is
necessary to evaluate functional parameters of the contact.
Table 2 gives the values of oil film thickness hm and Hertzian pressure P0 at worthy points of gear teeth
engagement.
3.2. Assessment of oil degradation using FTIR spectrometry

Table 2
Functional parameters of gear tooth contact.
Points

Rx (mm)

P0 (N/mm2)

hmin (lm)

P
Q
O
R
S

6.194
6.263
6.291
6.378
6.485

32.85
46.21
46.09
45.81
32.11

0.0171
0.0214
0.0270
0.0354
0.0390

M. Amarnath, S.-K. Lee / Measurement 76 (2015) 3244

carbonyls. Lubricant oxidation also results in the depletion

of antioxidant additives in the lubricant, usually Zinc
dialkyldithiophosphate (ZDDP) or phenolic compounds.
The ZDDP concentration can be monitored via the PAOAC
stretch. Several other additives can be detected in FTIR
analysis which include detergents, dispersants, and occasionally polymethacrylate pour-point depressants, etc.
[16,18,25,26].

(a)
Absorbance

38

1.2

(a)

Healthy oil
Aged oil (200 hours)

1460 cm-1

0.9
1380 cm

-1

0.6
720 cm-1

0.3
0.0
600

4. Results and discussions

800

1000

1200

1400

1600

1800

Wavenumber (cm-1)

4.1. Lubricant degradation analysis using FTIR analysis

(b)

1.2

Healthy oil
Aged oil (400 hours)

1460 cm-1

Absorbance

0.9
1380 cm-1

0.6
720 cm-1

0.3
0.0
600

800

1000

1200

1400

1600

1800

Wavenumber (cm-1)

(c)

Healthy oil
Aged oil (after 600 hours)

1.2

1460 cm-1

Absorbance

The lubricant oil used in rotating machine elements

lubrication may undergo both physical and chemical degradation due to various operating conditions viz. speed variations, load fluctuations, and working environment. The
changes in the quality of lubricant in the process of their
usage mainly depend on the chemical compounds formed.
Under service conditions, machinery lubricants undergo
various profound chemical alterations: oxidations, polymerization, alkation, decomposition and so on. The formation of all these undesired compounds produces a negative
impact on the performance of rolling/sliding contact
machine elements in the rotating machinery [18,27,28].
In the present work, FTIR analysis method has been used
in conjunction with the lubricant film thickness analysis.
IR spectra were measured using Parkin Elmer Spectrum
One spectrometer within the range of 4003600 cm1,
however a wave number range of 6001800 cm1 has been
to considered to analyze absorbance bands related to thickener, base oil and oxidation products. The oil samples were
kept in Thorium bromide absorption cell of 3 mm thickness
and examined in the transmission mode.
The gear fatigue tests were carried out over a period
was 1200 h, the lubricant samples were collected from
the gearbox at interval of 200 h, Fig. 4(a)(c) depict the
IR spectra of the samples of gear box oil. Fig. 4(a) shows
the IR spectrum of healthy oil and oil sample collected
after 200 h. The absorption peaks shown at 1460 cm1
and 1380 cm1 correspond to deformation vibration of
methyl (CH2)/methylene (CH3). Fig. 5(c) shows a significant
increase in the absorbance peaks at 1460 cm1, this is due
to deformation vibration occurred in the lubricant after
600 h. A broadband between 800 cm1 and 1400 cm1 is
shown in Fig. 5(a)(c), the absorbance peaks present in this
region correspond to CAO bonds associated with various
oxygenated organic compounds, and sulfur species formed
due to oxidation of sulfur compounds present in the gear
oil.
Lubricant oil consists of sulfur compounds which consist of anti-wear agents and oxidation inhibitors. Lubricant
degradation and loss of lubrication ability are strongly
linked to oxidation. Fig. 5(b) and (c) show absorbance spectra obtained after 1000 h and 1200 h, a slight increase in
the height of the band around 1600 cm1 can be seen in
the plots. This band represents the presence of aromatics
in the aged oil sample, on the other hand a slight increase
in 1720 cm1 band shown in Fig. 5(c) corresponds to C@O
vibration of oxidation products viz. ketones, acids and

1380 cm-1

0.9
0.6

720 cm-1

0.3
0.0
600

800

1000

1200

1400

1600

1800

Wavenumber (cm-1)
Fig. 4. Infrared spectra of oil samples obtained during 0600 h.

aldehydes. Further, Fig. 5(b) and (c) highlighted a significant increase in absorption peak at 720 cm1 which indicates degradation of anti-wear agents of the lubricant
samples obtained after 1000 h and 1200 h respectively.

4.2. Fe content analysis

In the present study, the lubricant was collected for
wear particle contamination analysis after every 50 h of
operation. Experimentally measured Fe concentration level
in the lubricant from atomic absorption spectroscope (PerkinElmer AAS-5000) is plotted in Fig. 6. From this figure it
can be seen that during the initial stage of wear propagation the absolute values of Fe concentration level are comparatively low. Increase in fatigue load cycles causes
increase in combined surface roughness values, which led
to decrease in lubricant film thickness thereby causing
increase in asperity contacts. Prolonged operation under
increase in asperity contact condition causes material
removal from the tooth surface which causes increase in

39

M. Amarnath, S.-K. Lee / Measurement 76 (2015) 3244

(a) 1.2

Healthy oil
Aged oil (800 hours)

Absorbance

0.9
0.6

Fig. 6, which shows a gradual increase in Fe concentration

level over a period of 1200 h. Oxidation occurs more
rapidly in the presence of metallic debris, degradation of
oil due to oxidation process is already discussed in the previous section.

1460 cm-1
1380 cm-1
1600 cm-1

720 cm-1

0.3

4.3. Reduction in tooth thickness and increase in surface

roughness

0.0
600

800

1000

1200

1400

1600

1800

Wavenumber (cm-1)

Absorbance

(b)

Healthy oil
Aged oil (1000 hours)

1.2

1380 cm

0.9
0.6

1460 cm-1
-1

1600 cm-1
-1

720 cm

-1

1290 cm

870 cm-1

0.3

1720 cm-1

0.0
600

800

1000

1200

1400

1600

1800

Wavenumber (cm-1)

(c)

1380 cm-1



720 cm-1



1600 cm-1

1290 cm-1

-1

870 cm

1720 cm-1



600

750

900

1050 1200 1350 1500 1650 1800

Wavenumber (cm-1)
Fig. 5. Infrared spectra of oil samples obtained during 6001200 h.

Fe content (PPM)

150

100
at the root
at the tip
at the pitch circle
fitting curve (polynomial)

0.6

50

0
200

400

600

800

1000

1200

Operating time (Hours)

Fig. 6. Fe concentration versus operating time.

Fe contamination level in the lubricant. Therefore the concentration level increases with operational time and reduction in lubricant film thickness. This can be observed from

Reduction in tooth thickness (mm)

Absorbance

1460 cm-1

Healthy oil
Aged oil (1200 hours)



The severity of surface fatigue wear on the gear tooth

surfaces results a considerable changes in gear tooth profile which causes a reduction in gear tooth thickness. The
gear tooth thickness measurements were carried out after
run-in wear test for 30 h, tooth thickness measured after
run-in wear test was considered as the initial thickness
of tooth of the test gear. Four teeth on the test gear separated by 90 were considered to measure tooth thickness.
Thickness measurement was carried out at tooth tip, pitch
diameter and root regions of four teeth, however the average value was considered to assess severity of tooth wear
occurred during fatigue tests. The propagation of surface
fatigue wear results a reduction in tooth thickness on
above said tooth regions. Prolonged running time, increase
in fatigue load cycles on gear tooth contacting surfaces and
loss of physical and chemical properties of lubricant led to
increase in sliding action on gear tooth contacting surfaces,
this sliding action is maximum at the root. On the other
hand, at the tooth root, rigidity to deflection is comparatively higher than at tip and pitch regions [21], hence sliding action in conjunction with higher rigidity to deflection
led higher wear at the root. Theoretical and experimental
studies on gear tooth deflection, stiffness modeling and
measurements were published in our previous publication
[19]. Pure rolling and no sliding at the pitch diameter
causes a relatively lesser amount of reduction in tooth
thickness compared to that of root and tip regions. Fig. 7
shows a reduction in tooth thickness versus operating
time. It can be seen from the figure that a reduction in
thickness at the tooth root is comparatively higher than
that of tip and pitch regions. The mechanical energy lost

0.5
0.4
0.3
0.2
0.1
0.0
0

200

400

600

800

1000

Operating time (Hours)

Fig. 7. Reduction in tooth thickness versus operating time.

1200

M. Amarnath, S.-K. Lee / Measurement 76 (2015) 3244

by the friction is generally inevitable and sometimes very

difficult to overcome, the thermal aspects of friction may
result in unexpected consequences in gear transmission
system [3,21]. Apparently, lubricant temperature measured from the gear mesh position showed increase in
temperature values during first six hours of the test; however the lubricant temperature stabilizes at the end of the
test. The viscosity of the lubricant is dependent on operating temperature, hence increase in lubricant temperature
results in decrease in viscosity and lubricant film thickness. Table 3 gives increase in temperature and decrease
in lubricant film thickness obtained under three load conditions. The wear on gear tooth surfaces is a continuous
failure under thin film conditions. On the other hand, typically higher loads, speeds, temperatures and increase in
fatigue load cycles, result increase in asperity contacts
[2,9]. The wear propagation leads to material removal from
the mating surfaces with increase in load cycles, which
causes increase in friction and surface roughness. Fig. 8
shows increase in average surface roughness (Ra) and
peak-to-valley surface roughness (Rz) values as function
of time.
4.4. Wear mechanism developed on gear tooth surfaces
Gear tooth damage is caused by a variety of factors
including inadequate lubrication, inappropriate operating
conditions or specifications, material insufficiencies and
manufacturing installations. When gears operate at their
maximum load, very high contact pressure occurs at their
mesh interface. This may lead to partial breakdown of
the lubricant film at the contact surfaces thereby resulting
in two commonly encountered regimes, the elastohydrodynamic and boundary lubrication conditions [2,6,30].
Thus the physical property of gear lubricant oil i.e.
viscosity is the main parameter affecting the lubricant.
High temperatures lead to low viscosities which cause a
reduction in lubricant film thickness and thus thin lubricant film in the gear mesh has detrimental influence on
the failure performance of a geared system. At the initial
stage of the gear fatigue test, a small amount of micro pitting was observed after 200 h as shown in Fig. 9(a). The oil
additives viz. sulpher, phosphorous, and polysulfides
present in the SAE grade oil cause formation of chemical
reaction layer on the contacting surfaces, this prevents

Table 3
Reduction in lubricant film thickness.
Operating

Lubricant temperature

9
8

Roughness of Ra, Rz (m)

40

Ra
Rz

7
6
5
4
3
2
1
0
0

200

400

600

800

1000

1200

Operating time (Hours)

Fig. 8. Gear tooth surface roughness as function of time.

metal to metal contacts thereby increasing scoring resistance during the initial stage of operation. Over a period
of time, degradation in physical and chemical properties
of lubricating oil accelerates wear damage on the tooth
surfaces. Increase in pit size and the distribution of pits
on the adjacent tooth can be observed in Fig. 9(b), this
damage was observed after 400 h.
Scuffing is a severe form of adhesive wear, which results
in sudden damage to one or both surfaces in relative
motion. In contrast to other types of lubrication related
failure modes, scuffing generally takes time to develop or
reach destructive magnitude. However, scuffing on the
gear tooth surfaces cannot occur if an oil film of sufficient
thickness separates the contacting surfaces [3,22]. The
reduction in oil film thickness causes increase in metal to
metal contact which causes micro welding between asperity contacts, increase in fatigue load cycles under such
operating conditions causes scuffing defect. Fig. 9(c) shows
scuffing defect appeared after 800 h. Gear teeth contacts
have complex combination of sliding and rolling which
vary along the profile of each tooth. In the dedendum,
the direction of rolling is opposite to that of sliding, hence
negative sliding condition exists in the dedendum [22,23].
The contact fatigue is more likely to initiate in the dedendum region. Fig. 9(d) shows severe scoring damage
occurred on the dedendum, this damage appeared after
1200 h of fatigue test.
4.5. Vibration signal analysis

Dimension less lubricant

film thickness values

Time (h)

0Nm

6Nm

12 N m

0Nm

6Nm

12 N m

0
1
2
3
4
5
6
7
8

24
25
26
27
29
33
35
38
39

23
26
28
30
31
34
37
40
41

24
26
29
31
33
36
40
41
43

6.985
6.196
5.553
4.899
3.827
3.213
2.937
2.456
2.221

6.421
5.852
5.281
4.223
3.281
2.556
2.383
2.222
2.011

6.102
5.482
4.909
3.692
2.472
2.164
2.111
2.091
1.825

Gear mesh stiffness plays a key role in gear dynamics,

magnitude of noise and vibrations generated as well as load
carrying capacity of geared systems. Surface fatigue is the
root cause which accelerates fault propagation on gear
tooth surfaces; pitting is one type of surface fatigue and
often occurs in the early stages of gear failure. When pitting
is initiated, the load is applied only on a finite region of the
gear tooth surface. At the same time contact stress
increases enormously in the gear teeth contact area. Hertzian compliance is the most significant part of the total
compliance around the pitch line. The presence surface

M. Amarnath, S.-K. Lee / Measurement 76 (2015) 3244

(a) Pitting on dedendum appeared

aer 200 hours

(c) Scung damage appeared aer 800

hours

41

(b) Micro ping on adjacent teeth

appeared aer 400 hours

(d) Scoring on dedendum appeared

aer 1000 hours

Fig. 9. Surface fatigue wear on the pinion tooth surfaces.

RMS values of vibration signals

14
12
10
8
6
4
2
0

200

400

600

800

1000

1200

Operating time (Hours)

Fig. 10. RMS values versus operating time.

fault in the gear teeth contact region results a loss in tooth

contact thereby causing a reduction in tooth stiffness which
leads to increase in vibration levels of gear transmission
system [19,29,30]. Further, alternative phenomena to diagnose the wear propagation on gear tooth, for instance pitting, are often related to the specific film thickness, which
is a hint for the possibility of lubricant film breakdown.
The physical properties of the lubricant viz. viscosity and

temperature play an important role, these properties have

a strong influence to maintain sufficient oil film thickness
between gear teeth contacts. Obviously the connection
between pitting, stiffness and vibration signal amplitude
provides a means for monitoring gear defects. Wear mechanisms viz. macro pitting, scuffing, and scoring faults were
observed on gear teeth as shown in Fig. 9. During the gear
fatigue tests, the pinion speed was set to 450 rpm giving
raise to fundamental gear mesh frequency of 180 Hz. The
vibration signals were measured using the tri-axial
accelerometer mounted on the pinion bearing housing.
The vibration signals were sampled at 8.2 kHz, this frequency range is enough to show up to fifth gear mesh frequency harmonics i.e. 1fm, 2fm, . . . 5fm.
The statistical parameters viz. root mean square (RMS),
crest factor and kurtosis values have been considered to
indicate wear progression on gear tooth surfaces. Fig. 10
depicts a plot of vibration RMS values against operating
time. The reduction in oil film thickness between gear
teeth contacts causes increase in vibration levels, this can
be seen in Fig. 10. At the initial stage of operation, hydrodynamic lubrication regime resulted in lesser wear on
tooth surfaces thereby resulting in lower vibration levels
during 0400 h, further reduction in film thickness led to
increase in surface fatigue wear which causes increase in
the vibration levels, a steeper raise in RMS values extracted
from the vibration signals is shown in Fig. 10. A similar
trend is observed in the crest factor values as shown in
Fig. 11. The kurtosis values are known to be fourth
statistical moment and most commonly used to detect

42

M. Amarnath, S.-K. Lee / Measurement 76 (2015) 3244

Fig. 13(a) shows the vibration spectrum of a healthy

gear, gear mesh frequency and its higher harmonics along
with side bands have been indicated on the spectrum, it
can seen from the figure that the fundamental gear mesh
frequency harmonic and its corresponding higher harmonics show lower amplitudes. The severity of wear mechanisms on gear tooth surfaces was discussed in the
previous section, the increase in surface fatigue wear on
gear tooth surfaces during 4001200 h causes a gradual
increase in the amplitudes of gear mesh frequency harmonics as shown in Fig. 13(b)(e).

5. Summary and conclusions

7
6

Crest factor

5
4
3

0
0

200

400

600

800

1000

1200

Operating time (Hours)

Fig. 11. Crest factor values versus operating time.

and diagnose the faults in rotating machine elements. Fault

propagation causes increase in spikes in the vibration signature, kurtosis value of vibration signals provides a measure of spikes. Fig. 12 depicts a plot of the variation of
kurtosis values versus operating time, kurtosis values
obtained at regular interval show uneven trend due to
the higher susceptibility to spurious effect of noise and
high frequency vibration signals. Hence, in some cases,
the adverse effect of noise on the values of kurtosis is more
than the benefit gained from the higher sensitivity of kurtosis to detect incipient faults [31], similar trend was
observed in our previous publications [13,32], in which
both surface fatigue and simulated faults were investigated
using vibration and sound signals in spur and helical
geared systems. Fig. 13(a)(e) show vibration spectra for
the healthy and worn gears, the X and Y axes represent frequency in Hz and amplitude in m/s2 respectively.

18
16

Kurtosis values

14
12
10
8
6
4
2
0
0

200

400

600

800

1000

Operating time (Hours)

Fig. 12. Kurtosis values versus operating time.

1200

The work presented in this paper was aimed to assess

the development of surface fatigue wear in a spur geared
system. A single stage spur gear box was used to conduct
fatigue test experiments. In the fatigue test experiments,
the gearbox was allowed to operate over a period of
1200 h. Experimental studies include lubricant film thickness analysis in conjunction with oil degradation measurement and analysis. Further, vibration signal analysis,
observations on the development of wear mechanism,
and measurements on reduction in tooth thickness were
also considered in the experimental investigations. The following conclusions were drawn from the experimental
observations.
1. Degradation of the lubricant oil due to operating conditions was effectively monitored using FTIR technique.
Degradation products viz. ketones, acids, aldehydes,
deformation vibrations of methylene (CH2) and methyl
(CH3) and sulfonate S@O stretching and phinolate
bending OAH were observed in the wavenumber range
of 6001800 cm1.
2. A reduction in the lubricant film thickness was estimated using DowsonHigginsons equation. Reduction
in film thickness causes increase in metal to metal contact in the geared system thereby triggering tooth damage. Increase in wear severity on gear teeth causes
increase in vibration levels in the geared system.
3. Fe contamination levels showed increase in trend as
function of operating time, increase in contamination
levels in the lubricant results in oxidation, the oxidation
process was effectively monitored using FTIR method.
4. The vibration data acquired from the gearbox under
healthy and worm gear revealed the operating conditions of the gearbox. Increase in tooth wear causes
increase in amplitudes of fundamental gear mesh frequency and its corresponding higher harmonics in the
vibration spectra.
5. Furthermore, a commonly used statistical parameters
of vibration signals viz. RMS, crest factor and kurtosis
values were used to extract fault diagnostic information. RMS and crest factors showed increase in trend
as a function of operating time. However, kurtosis values showed irregular trend, this is due to its high susceptibility to the spurious effect of noise and high
frequency vibration signals.

43

M. Amarnath, S.-K. Lee / Measurement 76 (2015) 3244

1.0

0.025

Amplitude (m/s2)

(a)
0.75

0.000
0.5

2f m

1f m

0.25

3f m

4fm

0.00
0

200

400

600

800

1000

1200

Frequency (Hz)
1.0

1.0

Amplitude (m/s2)

Amplitude (m/s2)

(b)
0.75
0.5

1fm

0.25

2fm

3fm

4fm

(c)

0.8
0.6
0.4

2f m

0.2

4fm

0.0

0.0

200

400

600

800

1000

1200

200

400

Frequency (Hz)

800

1000

1200

1.0

(d)

0.8

Amplitude (m/s2)

Amplitude

600

Frequency (Hz)

1.0

(m/s2)

3fm

1fm

3f m

0.6

2f m

0.4

4f m

1f m

0.2

(e)

0.8
0.6

3f m

2f m

0.4

4f m

1f m

0.2
0.0

0.0
0

200

400

600

800

1000

1200

Frequency (Hz)

200

400

600

800

1000

1200

Frequency (Hz)

Fig. 13. FFT spectra of vibration signals at 12 N m gearbox input torque (a) 0 h, (b) 300 h, (c) 600 h, (d) 900 h, and (e) 1200.

6. Tooth thickness measurements showed higher values of

tooth reduction at the tooth root than that of tip and
pitch circle diameter. Apparently, roughness values
were measured at the root, pitch line and tip regions,
a higher roughness values were observed at the tooth
root which is due to negative sliding and higher rigidity
to deflection at the root zone.
It is postulated that aforementioned fault diagnostic
indicators within a monitoring technique must be used
simultaneously in order to establish and explain the wear
phenomena. The most ideal situation is to use various
techniques at the same time to monitor the gearbox health
status.

Acknowledgments
This work was supported by Mid-career Researcher
Program through NRF grant funded by the MEST (No.

2010-0014260) and this work was supported by INHA

University Research Grant.
References
[1] R. Rakic, The influence of lubricants on cam failure, Tribol. Int. 37
(2004) 365373.
[2] J.W. Bartz, Lubrication of Gearing, Mechanical Engineering
Publication Ltd, London, 1993.
[3] B.R. Hhn, K. Michaelis, Influence of oil temperature on gear failures,
Tribol. Int. 37 (2003) 103109.
[4] C.K. Tan, P. Irving, D. Mba, A comparative study on the diagnostic and
prognostic capabilities of acoustic emission, vibration and
spectrometric oil analysis for spur gears, Mech. Syst. Signal
Process. 21 (1) (2007) 208233.
[5] J. Castro, J. Seabra, Scuffing and lubricant film breakdown in FZG
gears Part 2. New PV scuffing criteria, lubricant and temperature
dependent, Wear 215 (1998) 114122.
[6] D. Dowson, G.R. Higginson, Elasto Hydrodynamic Lubrication,
Pergamon Press Ltd, New York, 1977.
[7] T. Karacay, N. Akturk, Experimental diagnostics of ball bearings using
statistical and spectral methods, Tribol. Int. 42 (2009) 836843.
[8] T. Itagaki, H. Ohta, T. Igarashib, Effect of grease type on abnormal
vibration of ball bearing, J. Sound Vib. 268 (5) (2003) 933946.

44

M. Amarnath, S.-K. Lee / Measurement 76 (2015) 3244

[9] Z. Peng, N. Kessissoglou, An integrated approach to fault diagnosis of

machinery using wear debris and vibration analysis, Wear 225 (7
12) (2003) 12211232.
[10] C.K. Tan, P. Irving, D.A. Mba, Comparative study on the diagnostic
and prognostic capabilities of acoustic emission, vibration and
spectrometric oil analysis for spur gears, Mech. Syst. Signal
Process. 21 (1) (2007) 208233.
[11] M. Elforjani, D. Mba, Monitoring the onset and propagation of
natural degradation process in a slow speed rolling element bearing
with acoustic emissions, ASME, J. Vib. Acoust. 130 (2008) 041013.
[12] R. Serrato, M.M. Maru, L.R. Padovese, Effects of lubricant viscosity
grade on mechanical vibration of roller bearings, Tribol. Int. 40
(2007) 12701275.
[13] M. Amarnath, I.R. Praveen Krishna, Detection and diagnosis of
surface wear failure in a spur gear system using EEMD based
vibration signal analysis, Tribol. Int. 61 (2013) 224234.
[14] T. Akagaki, M. Nakamur, T. Monzen, M. Kawabata, Analysis of the
behavior of rolling bearings in contaminated oil using some
condition monitoring techniques, Proc. IMechE Part J: J. Eng.
Tribol. 220 (5) (2006) 447453.
[15] R.I. Hamzah, D. Mba, The influence of operating condition on
acoustic emission (AE) generation during meshing of helical and
spur gear, Tribol. Int. 42 (1) (2009) 314.
[16] M.P. Zakharich, I.I. Zaitsev, V.P. Komar, F.N. Nikonovich, M.P.
Ryzhkov, I.V. Skornyakov, Analysis of transformer oil using IR
analyzers, J. Appl. Spectrosc. Mol. Struct. 68 (1) (2001) 6165.
[17] S. Kumar, N.M. Mishra, S. Mukherjee, Online condition monitoring of
engine oil, Ind. Lubr. Tribol. 57 (6) (2005) 260267.
[18] M.A. Al-Ghouti, L. Al-Atoum, Virgin and recycled engine oil
differentiation: a spectroscopic study, J. Environ. Manage. 11
(2007) 19.
[19] M. Amarnath, C. Sujatha, S. Swarnamani, Experimental
investigations of surface wear assessment of spur gear teeth, J. Vib.
Control 18 (2012) 10091024.

[20] M. Amarnath, C. Sujatha, S. Swarnamani, Detection and diagnosis of

gear tooth wear through metallurgical and oil analysis, Tribol.
Online 5 (2) (2010) 102110.
[21] M.R. Sari, A. Haiahem, L. Flamand, Effect of lubricant contamination
on gear wear, Tribol. Lett. 27 (1) (2007) 119126.
[22] P.J.L. Fernandes, C. Mc Duling, Surface contact fatigue failures in
gears, Eng. Fail. Anal. 4 (2) (1997) 99107.
[23] P.J.L. Fernandes, Tooth bending failures in gears, Eng. Fail. Anal. 3 (3)
(1996) 215219.
[24] P.D. Mc Fadden, J.D. Smith, A signal processing technique for
detecting local defects in a gear from the signal average of the
vibration, Proc. Inst. Mech. Eng. 199 (C4) (1985) 287292.
[25] V.E. Goleva, M.I. Rundev, V.A. Pisanov, Determination of esters in
motor and industrial oils by IR spectrometry, J. Anal. Chem. 60 (8)
(2005) 719722.
[26] F.E. Lockwood, R. Dalley, Lubricant analysis, in: ASM Handbook
Friction, Lubricant and Wear Technology, vol. 18, 1992, pp. 299312.
[27] P.M. Cann, J.P. Doner, M.N. Webster, V. Wikstrom, Grease
degradation in rolling element bearings, Tribol. Trans. 44 (3)
(2001) 399404.
[28] P.M. Cann, M.N. Webster, J.P. Doner, V. Wikstrom, P. Lugt, Grease
degradation in R0F bearing tests, Tribol. Trans. 50 (2007) 187197.
[29] I. Yesilyurt, Fengshou, A.D. Ball, Gear tooth stiffness measurement
using modal analysis and its use in wear fault severity assessment of
spur gears, NDT&E Int. 36 (2003) 357372.
[30] I. Yesilyurt, Gearbox Fault Detection and Severity Assessment Using
Vibration Analysis, Ph.D Thesis, University of Manchester, 1997.
[31] H. Oack, K.A. Laprao, HMM based fault detection and diagnosis
scheme for roller element bearing, J. Vib. Acoust. 127 (4) (2005)
299306.
[32] M. Amarnath, I.R. Praveen Krishna, Empirical mode decomposition
of acoustic signals for diagnosis of faults in gears and rolling element
bearings, IET Sci., Meas. Technol. 6 (4) (2012) 279287.