Abstract

Rolling element bearing is a most critical part of any rotating machinery. The
machine could be unsafe if any fault occurs in the bearing elements due to fatigue,
manufacturing errors and operating conditions. Due to sudden failure of bearing the machine
may cause serious accident, economical loss and time. Therefore, it is necessary to predict the
faults in bearing well in advance preventive maintenance of machinery. The vibration based
condition monitoring is most effective method for identifying fault in the bearing during its
operation. In this project a test rig has been designed and developed for vibration
measurement of shaft-rolling bearing system. The development of test rig includes design and
manufacturing of shaft, selection of support and test bearing along with suitable pedestal,
selection of belt, pulley and motor. The vibrations data has been captured by accelerometer
for test bearing in presence of normal and defective bearing outer raceway. The measured
vibrations has been analysed in frequency domain to predict the fault characteristics. The
vibrations generated by bearing depends on bearing geometry, load and shaft speed. It has
been also study the effect of defect size, load and shaft speed on vibration response of healthy
and defective bearing.
 INDEX

Sr No. Content
1 Introduction
1.1 Introduction
1.2 What is bearing?
1.3 Types of bearing
1.4 Objectives
2 Literature Review
3 Design of test rig for vibration monitoring
3.1 Selection of motor
3.2 Selection of belt and pulley
3.3 Design of shaft
3.4 Selection of bearing
4 Working
5 Expected Conclusion
Referances
 Nomenclature

C - Centre distance
d - Ball diameter
D - Pitch diameter
Dmp - Motor pulley diameter
Dsp - Shaft pulley diameter
fs - Shaft frequency
ksz - Size factor
ksur - Surface finish factor
krel - Reliability factor
kf - Fatigue stress concentration factor
L - Belt length
n - Number of balls in rolling element bearing
S - Slip of belt
 - Contact angle
Ø - Diameter
σe’ - Endurance limit stress
σe - Equivalent Endurance limit stress
1. INTRODUCTION

1.1 Introduction

The rolling element bearing is widely used in industrial machinery. The main function
of bearing is to support the load and permit relative motion. Due to longer period of use for
rolling bearing it may generate local defects viz, crack, pits, spalls, etc., due to fatigue of
mating parts of bearing. Even defect free bearing may produce vibrations due to time varying
contact forces between mating parts or due to varying compliance effect.
The nature of vibration response changes due to absence and presence of defect on contacting
elements of bearing. Due to failure of bearing in the machine, it will cause sudden break
down or in some cases possible catastrophic failure of machine. The failure of bearing in the
machine will cause shut down the production process and accident to the operating personnel.
The various condition monitoring techniques are used for identifying health of machinery.
The main techniques for condition monitoring of machinery are vibration measurement,
temperature measurement, wear debris and sound measurement which are used in the
industry. The vibration measurement is a most effective technique and widely used in
industry for fault identification in a rolling element bearing. The main elements of rolling
bearing are inner race, outer race, cage (Retainer) and rolling element (ball / roller). The
mainly two types of defect (Localized and Distributed defect) are observed on elements of
bearing. The bearing faults like spalls, pits, dirt, dent, and crack on the rolling surfaces,
brinelling and contaminations in lubricant are considered as localized defect while surface
roughness, manufacturing errors, off size rolling element, raceway or rolling element
waviness and misaligned races are classified as a distributed fault in a bearing. In vibration
analysis the vibration signals are captured from machine parts and analysed in time and
frequency domain. The presence of fault in the element of bearing indicates peak in the
vibration spectrum. However amplitude of vibration peak is depends on size, shape, location
and type of defect. The characteristics defect frequencies for inner, outer and cage are termed
as ball pass frequency for inner / outer (BPFO / BPFI), ball spin frequency (BSF) and
fundamental train frequency (FTF) respectively. These characteristics defect frequencies are
calculated using equations (1 -4). The defect frequency of bearing elements depends on the
shaft rotational speed and bearing geometry.
 Many of the researchers have studied theoretically and experimentally the vibration
response of healthy and defective bearing in presence of varying defect size, shape and its
type. It is required to improve the existing proposed dynamic model for the better prediction
of fault in bearing elements through vibration monitoring technique. In theoretical model
researchers have considered additional deflection or excitation force of rolling element due to
interaction of defective races with rolling element. A test rig has been designed and
developed to perform the experiment and measure the vibrations generated by defective and
healthy bearing for validation of theoretical results and bearing condition monitoring. The
developed test rig is also useful for testing the bearing vibrations when it is not possible to
check on site. The vibration measurement of healthy and defective deep groove ball bearing
have been captured through a vibration measuring device. The theoretical characteristics
defect frequencies for elements of bearing have been calculated using equations (1 -4) which
are verified with experimental defect frequencies obtained using performed experiment. The
developed test rig must have sufficient stiffness for support structure of system to avoid
unnecessary vibrations due to some other sources. It is only possible to see clearly the
dominent frequencies and vibration amplitude for test bearing if the test rig supports are rigid
in nature.

1.2 What is bearing ?

A bearing is a machine element that constrains relative motion to only the desired

motion, and reduces friction between moving parts. The design of the bearing may, for
example, provide for free linear movement of the moving part or for free rotation around a
fixed axis; or, it may prevent a motion by controlling the vectors of normal forces that bear
on the moving parts. Most bearings facilitate the desired motion by minimizing friction.
Bearings are classified broadly according to the type of operation, the motions allowed, or to
the directions of the loads (forces) applied to the parts.

Rotary bearings hold rotating components such as shafts or axles within mechanical

systems, and transfer axial and radial loads from the source of the load to the structure
supporting it. The simplest form of bearing, the plain bearing, consists of a shaft rotating in a
hole. Lubrication is often used to reduce friction. In the ball bearing and roller bearing, to
prevent sliding friction, rolling elements such as rollers or balls with a circular cross-section
are located between the races or journals of the bearing assembly. A wide variety of bearing
designs exists to allow the demands of the application to be correctly met for maximum
efficiency, reliability, durability and performance.

The term "bearing" is derived from the verb "to bear";[1] a bearing being a machine
element that allows one part to bear (i.e., to support) another. The simplest bearings
are bearing surfaces, cut or formed into a part, with varying degrees of control over the form,
size, roughness and location of the surface. Other bearings are separate devices installed into
a machine or machine part. The most sophisticated bearings for the most demanding
applications are very precise devices; their manufacture requires some of the highest
standards of current technology.

1.3 Types of bearing

1. Plain Bearings

Plain bearings are the simplest type of bearing and are composed of just the bearing surface
with no rolling elements. They have a high load-carrying capacity, are generally the least
expensive and, depending on the materials, have much longer lives than other types.
 2. Rolling Element Bearings

Rolling element bearings place balls or rollers between two rings – or “races” – that
allows motion with little rolling resistance and sliding. These bearings include ball bearings
and roller bearings.

Ball bearings are the most common type of rolling element bearing. These bearings
can handle both radial and thrust loads but are usually used where the load is relatively small.
Because of its structure, there is not a lot of contact with the balls on the inner and outer
races.  If the bearing is overloaded the balls would deform and ruin the bearing. Roller
bearings are able to handle a much heavier, radial load, like conveyor belts, because they
don’t use balls. Instead, they have cylinders allowing more contact between the races,
spreading the load out over a larger area. However this type of bearing is not designed to
handle much thrust loading

3. Jewel bearing

Jewel bearings are plain bearings with a metal spindle that turns in a jewel-lined pivot
hole. They carry loads by rolling the axle slightly off-center and are usually used in
mechanical watches or clocks. This is due to their low and predictable friction that improves
watch accuracy.
 4. Fluid Bearings

Fluid bearings support their load using a thin layer of gas or liquid and can be
classified into two types: fluid-dynamic bearings and hydrostatic bearings. Fluid-dynamic
bearings use rotation to form the liquid into a lubricating wedge against the inner surface. In
hydrostatic bearings, the fluids – usually oil, water, or air – rely on an external pump.
Fluid bearings are used in high load, high speed or high precision applications that ordinary
ball bearings either couldn’t handle or would suffer from increased vibration and noise.

5. Magnetic Bearing

Magnetic bearings support moving parts without physical contact, instead relying on

magnetic fields to carry the loads. They require continuous power input to keep the load
stable, thus requiring a back-up bearing in the case of power or control system failure.
Magnetic bearings have very low and predictable friction and the ability to run without
lubrication or in a vaccum. They are increasingly used in industrial machines like turbines,
motors, and generators.

6. Flexure Bearing

A typical flexure bearing is one part joining two others, like a hinge, in which motion is
supported by a load element that bends. These bearings require repeated bending, so material
selection is key. Some materials fail after repeated bending, even at low loads, but with the
right materials and bearing design the flexure bearing can have an indefinite life. Another
notable characteristic of this bearing is its resistance to fatigue. Many other bearings that rely
on balls or rollers can fatigue as the rolling elements flatten against each other.

1.4 Objectives

The work has been carried out in the following stages:

 To construct a single rotor test rig supported in conventional ball bearings. That could be
used to simulate and determine vibration analysis of bearing with different faults.

 Specifically vibration analysis of healthy ball bearing, bearing with inner race crack,
bearing with outer race crack, ball bearing with ball having spall and bearing without
lubrication are the targeted faults.

 Under investigation both supports acceleration has been carried out for the vibration
signatures due to different faults mention with the help of accelerometer and vibration
measuring device.

2. LITERATURE REVIEW
 In 2002, N.G. Nikolaou et.al presented rolling element bearing fault using wavelet
packets (WPT). A time-frequency decomposition of vibration signals is analyzed and those
components of signal having important information are selected for processing. The method
is evaluated using simulated and experimental signals. They have used a PCMCIA DAQ
Card-1200 data acquisition card. He concluded that compared to other methods, using
filters or continuous wavelet transform, it has the advantage of flexibility and efficient
computational implementation.

In 2007, V.K. Rai, A.R. Mohanty [2007] have discussed bearing fault diagnosis
using FFT of intrinsic mode functions in Hilbert–Huang transform. They have used the
conventional data acquisition card for measurement of vibration data and calculated the CDFs
analytically. They have taken readings for Healthy bearings, IR fault and OR faults. The
HHT technique provides multi-resolution in various frequency scales and takes the signals
frequency content their variation into consideration. In the presentwork, FFT of IMFs from
HHT process has been incorporated to utilize efficiency of HT in frequency domain. The
comparative analysis is presented in this paper indicates the effectiveness of using frequency
domain approach in HHT and its efficiency as one of the best-suited techniques for bearing
fault diagnosis (BFD). They have done comparative analysis and indicated the effectiveness
of HT in frequency domain. They have shown presence of amplitude modulation as well as
capturing the defect frequencies accurately.

In 2009, Tuncay Karacay et.al have discussed experimental diagnostics of ball

bearings using statistical and spectral methods. They have conducted test on test rig with new
bearings and data acquisition card is used to measure the vibrations. The bearings run
throughout their lifespan under constant speed and loading conditions. The formation and
development of localized defect in ball bearings are investigated running a bearing through
failure. Vibration signatures produced are recorded and statistical measures are calculated
during the test. In this study, the formation and development of localized defect in ball
bearings are investigated running a new bearing through to failure. During the test, vibrations
are acquired and traditional metric values such as peak-to-peak value, RMS, Crest factor and
kurtosis are calculated. They have found that first an inner-race defect is formed in the ball
bearing together with a slight ball defect. While at the end of test an inner-race and ball
localized defects developed at the end of the test, an outer race defect followed them.
 In 2009, P.K. Kankar, et.al have discussed fault diagnosis of a rotor bearing
system using response surface method. They have considered the distributed defects such as
internal radial clearance and surface waviness of the bearing components. They have studied
ball waviness, inner race waviness and outer race waviness individually and also taken
readings of combined waviness. They investigate an accurate performance prediction, which
is essential to the design of high performance rotor bearing system. They used mathematical
formulation the contacts between the rolling elements and the races are considered as
nonlinear springs, whose stiff nesses are obtained by using Hertzian elastic contact
deformation theory. From the obtained responses following conclusions are drawn:
1)Nonlinear dynamic responses are found to be associated with large internal radial clearance
and distributed defects.

2)It is shown that the system exhibit dynamic behaviors that are extremely sensitive to small
variations of the system parameters, such as internal radial clearance and ball waviness.
3)The system shows periodic nature, when ball waviness is at its maximum level.

In 2010, P.K. Kankar et.al have focused on fault diagnosis of ball bearings using
machine learning methods. They have used supervised machine learning methods like
artificial neural network (ANN) and support vector machine (SVM) to extract the features. A
test rig of high speed rotor supported on rolling bearings is used. The vibration response are
obtained and analyzed for the various defects of ball bearings. The specific defects are
considered as crack in outer race, inner race with rough surface and corrosion pitting in balls.
A comparative experimental study of the effectiveness of ANN and SVM is carried out. This
study presents a procedure for detection of bearing fault by classifying them using two
machine learning methods, namely, ANNs and SVMs. Features are extracted from
timedomain vibration signals using statistical techniques. it is observed that bearing shows
severe vibrations occur under bearing with rough inner race surface and ball with corrosion
pitting. It is also observed that the classification accuracy for SVM is better than of ANN.

In 2011, Matej Tadina, et.al have developed the numerical bearing model to
investigate the vibrations of the ball bearing during run-up. The numerical bearing model was
developed with the assumptions that the inner race has only 2 DOF and that the outer race is
deformable in the radial direction, and is modelled with finite elements. They have
considered centrifugal load effect and the radial clearance effect. The contact force for the
balls is described by a nonlinear Hertzian contact deformation. They have considered outer
vrace, inner race and ball defects into account. They used the continuous wavelet transform
(CWT) and envelope analysis to identify simulated bearing faults. The continuous wavelet
transform (CWT) found better for vibration analysis of a bearing with ball fault.

In 2011, P.K. Kankar, Satish C. Sharma, S.P. Harsha have discussed bearing fault
diagnosis using continuous wavelet transform (CWT). They have used three machine
learning techniques for fault classifications out of which two are supervised machine learning
techniques i.e. support vector machine (SVM), artificial neural network (ANN) and other one
is an unsupervised machine learning technique i.e. selforganizing maps (SOM). The result
showed that the SVM identifies the fault categories of rolling element bearings more
accurately compared to ANN and SOM. The fault diagnosis method consists of three steps,
firstly the six different base wavelets are considered in which three are from real valued and
other three from complex valued. Out of these six wavelets, the base wavelet is selected
based on wavelet selection criterion to extract statistical features from wavelet coefficients of
raw vibration signals. Two wavelet selection criteria Maximum Energy to Shannon Entropy
ratio and Maximum Relative Wavelet Energy are used and compared to select an appropriate
wavelet for feature extraction. Finally, the bearing faults are classified using these statistical
features as input to machine learning techniques. This study presents, a methodology for
detection of bearing faults by classifying them using three machine learning methods, like
SVM, ANN, and SOM. This methodology incorporates most appropriate features, which are
extracted from wavelet coefficients of raw vibration signals.

In 2012, Mohit Lal, Rajiv Tiwari, have developed an identification algorithm to

estimate parameters of multiple faults in a turbine-generator system model based on the
forced response information. A simple discrete model of the system has been developed with
the assumption of the rigid rotor, flexible-bearings and the flexible-coupling. The advantage
of the present method in respect to existing methods is that it does not assume equivalent
forces and moments of faults, which often suffers from erroneous estimation of fault
parameters due to severe ill-conditioning of regression equations. The effect of change in
assumed unbalance parameters and disc masses on estimates of residual unbalance is
discussed and it is observed that the estimates are showing good agreement with change in
these parameters.

In 2012, Tiago Cousseau et.al, have studied influence of grease rheology on thrust
ball bearing friction torque. They were measured friction torque amd operating temperature
for seven different types of greases.they concluded that grease rheology is stronglty
dependent on the thickner type also the bearing rolling torque and bearing sliding torque
mainly dependes upon viscosity and specific film thickness of the oil. In 2013, Bubathi
muruganatham et.al has used singular spectrum analysis for extraction of features from
rolling bearing. An artificial neural network (ANN) is used for fault diagnosis. They have
considered four bearing conditions for study; healthy bearings, bearing with a point defect on
inner race, bearing with a point defect on outer race and bearing with a point defect on ball.
3.DESIGN OF TEST RIG FOR VIBRATION MONITORING

This section deals with design and development of test rig parts for the vibration
measurement of shaft-bearing system. The main parts of test rig are electric motor, belt, shaft,
support bearing, test bearing, test bearing housing, support bearing housing, table, shaft
pulley and motor pulley. In design of test rig includes selection and design of test rig parts as
per the following procedure. Following photograph shows the experimental set up.
3.1 Selection of Electric Motor:

In the present study a single phase AC motor has been selected with following
specifications:

Power Rating: 500w

Speed: 2800 RPM,

Current: 0.6 Amps

A variation in speed of AC motor is controlled by AC Speed regulator.

3.2 Selection of Belt and pulleys :

A V belt of Fanner make B1360B52 has been selected due to its major advantage of less
chances of slippage during operation. As per IS- 2494- 1974 for power rating and selection of
V–belt, A–cross section of V- belt has been selected. Based on the power rating of motor and
maximum rotation speed of motor shaft. A belt has length of A55.

Following photographs shows the used pulley and bearing in this set up :
3.3 Design of Shaft:

A shaft has been designed with considering subjected to fluctuating load with 25mm
uniform diameter. A shaft has been designed with a maximum of 1000 N load to be
sustained. A mild steel material has been selected for the shaft design. A Mild steel usually
contains 40 points of carbon at most. One carbon point is 0.01 percent of carbon in the steel.
This means that it has at most 0.4 percent carbon. Most steels have other alloying elements
other than carbon to give them certain desirable mechanical properties. 1018 steel, a common
type of mild steel, contains approximately 0.6 percent to 0.9 percent manganese, up to 0.04
percent phosphorus, and up to 0.05 percent sulphur. Varying these chemicals affects
properties such as corrosion resistance and strength. Maximum length of shaft of 300 mm has
been considered. The variable and mean bending moment has been find out considering
central shaft weight, weight of pulley at left side and load applied at right side. Following
figure shows the line diagram of shaft :
 3.4 Selection of bearing :

a) Non defective Bearing :

Non defective bearing is used as a reference

specification:

Dynamic load rating (Cr) 14,022 N

Static load rating (Cor) 7,843 N

Shaft Dia. (Fw) 25.00 mm

Housing width (b) 38.00mm

Shaft height (h) 36.5mm

Mounting hole center to center (e) 105.00mm

Housing length (a) 140.00mm

Mounting slot length (S2) 19.00mm

Mounting slot width (S1) 13.00mm

Housing base thickness (g) 13.00mm

Housing height (w) 71.00mm

Bearing inner race width (Bi) 34.00mm

Bearing inner race width short side (n) 14.30mm

Bearing inner race width extended side (m) 19.70mm

Mounting bolt size 10.00mm

Bearing number UCP 205

Weight 816.00gm

Material Cast iron housing , chrome steel bearing

B) Defective Bearing :
 The artificial defect on outer raceway has been generated using electro discharge machining
(EDM) process.

2 defective bearings are used as follows :

1) Bearing with 01 mm defect on outer casing.

2) Bearing with 02 mm defect on outer casing.

Specification:

Dyanamic load rating (Cr) 14,022 N

Static load rating (Cor) 7,843 N

Shaft Dia. (Fw) 25.00 mm

Housing width (b) 38.00mm

Shaft height (h) 36.5mm

Mounting hole center to center (e) 105.00mm

Housing length (a) 140.00mm

Mounting slot length (S2) 19.00mm

Mounting slot width (S1) 13.00mm

Housing base thickness (g) 13.00mm

Housing height (w) 71.00mm

Bearing inner race width (Bi) 34.00mm

Bearing inner race width short side (n) 14.30mm

Bearing inner race width extended side (m) 19.70mm

Mounting bolt size 10.00mm

Bearing number UCP 205

Weight 816.00gm

Material Cast iron housing , chrome steel bearing

Figure shows the defect generated on outer casing :
Bearing housing :

A suitable split type test bearing housing has been designed and developed for the ease of
dismantling and change of bearing during performing experimental work. A cast iron pulley
is used in which boring operation of diameter equal to the outer diameter of bearing is done.

The photographic view of test bearing housing is shown below.

Spring Balance :
A spring balance is used in order to apply load on the bearing. This is done by suitable
arrangement where, as the nut is tighten it causes stretching of spring and due to which load
is applied on the bearing. The load applied on bearing can be seen on the linear scale which is
marked on spring balance device.
4.WORKING

The experiments have been conducted on designed and developed test rig for
vibration monitoring of deep groove ball bearing. The vibrations responses have been
measured for the healthy (defect free) bearing for varying load of 2 kg, 5 kg and 10 kg with
shaft rotational speed of 1000 rpm and 1500 rpm. The experiments have been also conducted
for study of vibration response for outer raceway localized defective bearing with fault size of
0.5 mm, 1 mm and 1.5 mm diameter having uniform depth of 1 mm for shaft speed upto 2800
rpm. The artificial defect on outer raceway has been generated using electro discharge
machining (EDM) process. The shaft rotational speed has been varied using speed regulator
for electric motor. The experimental study for defective test bearing has been
performed at a steady radial load of 100 N which has been applied at the bottom of test
bearing housing by means of hook and load attachment. The weight of test bearing housing
and load attachment is neglected. The vibrations generated for healthy and defective outer
raceway of bearing is captured by vibration measuring device.
5.EXPECTED CONCLUSION

The comparisons indicates the good matching of theoretical results and experimental
results obtained for normal and defective test bearing by performing experiments on
developed test rig. The developed test rig can be used for vibration measurement with
varying shaft rotational speed and loading conditions in presence and absence of bearing
defects. In experimental results the peaks only at the shaft rotational frequency and its
multiple harmonics are clearly visible for defect free bearing. Whereas for outer raceway
defect the main peaks at shaft rotational frequency, bearing characteristics defect frequency
(BPFO) and its harmonics are visible in the spectra. Good agreements between
theoretical and experimental defect frequencies have been observed. The experimental results
show that amplitude of vibration is increased with increase in defect size. It is also
noticed from the experimental results that there is rise in magnitude of vibrations due to
increase in radial load on test bearing and speed of shaft rotation. The test rig design and
developed in this project is multifaceted in nature.
REFERENCES

[1] N.G. Nikalaou, I.A. Antoniadis, rolling element fault diagnosis using wavelet packets, NDT&E
International 35 (2002) page 197-205

[2] V. Purushothama, S. Narayanana, Suryanarayana A.N. Prasadb, Multi-fault diagnosis of rolling

bearing elements using wavelet analysis and hidden Markov model based fault recognition, NDT&E
International 38 (2005) page 654–664

[3] V.K. Rai, A.R. Mohanty Bearing fault diagnosis using FFT of intrinsic mode functions in Hilbert–
Huang transform, Mechanical Systems and Signal Processing 21 (2007) pages 2607–2615

[4] Tuncay Karacay, Nizami Akturk, Experimental diagnostics of ball bearings using statistical and
spectral methods, Tribology International 42 (2009) pages 836– 843

[5] P.K. Kankar, S.P. Harsha, Pradeep Kumar, Satish C. Sharma Fault diagnosis of a rotor bearing
system using response surface method European Journal of Mechanics A/Solids 28 (2009) pages 841–
857

[6] P.K. Kankar, Satish C. Sharma, S.P. Harsha, Fault diagnosis of ball bearings using machine
learning methods, Expert Systems with Applications 38 (2011) pages 1876–18

[7] P.K. Kankar, Satish C. Sharma, S.P. Harsha, Fault diagnosis of ball bearings using continuous
wavelet transform, Applied Soft Computing 11 (2011) pages 2300–2312

[8] Matej Tadina, MihaBoltezar, Improved model of a ball bearing for the simulation of vibration
signals due to faults during run-up, Journal of Sound and Vibration 330 (2011) pages 4287-4301

[9] P.K. Kankar, Satish C. Sharma, S.P. Harsha, Rolling element bearing fault diagnosis using
wavelet transform, Neurocomputing 74 (2011) pages 1638– 1645

[10] Mohit Lal, Rajiv Tiwari, Multi-fault identification in simple rotor-bearing-coupling systems
based on forced response measurements, Mechanism and Machine Theory 51 (2012) pages 87–109
[11] Tiago Cousseau, Beartriz M. Graca, Armando V. Campos, Jorge H.O. Seabra, Influence of
grease rheology on thust ball bearings friction torque, tribology internatonal, 46 (2012) pages 106–
113

[12] Feiyun Cong, JinChen, Guangming Dong, Michael Pecht, Vibration model of rolling element
bearings in a rotor-bearing system for fault diagnosis, Journal of Sound and Vibration 332 (2013)
pages 2081–2097

[13] Min-Chun Pan, Wen-Chang Tsao [2013] Using appropriate IMFs for envelope analysis in
multiple fault diagnosis of ball bearings, International Journal of Mechanical Sciences 69 (2013)
pages 114–124

[14] Hamdi Taplak, Selcuk Erkaya, Ibrahim Uzmay [2013] Experimental analysis on fault detection
for a direct coupled rotor-bearing system, Measurement 46 (2013) pages 336–344

[15] Rajesh Kumar, Manpreet Singh, Outer race defect width measurement in taper roller bearing
using discrete wavelet transform of vibration signal, Measurement 46 (2013) pages 537–545

[16] Bubathi Muruganathamn, M.A.Sanjith, B. Krishnakumar, S .A. V. SatyaMurty [2013] Roller

element bearing fault diagnosis using singular spectrum analysis, Mechanical Systems and Signal
Processing 35 (2013) pages 150–166

[17] R.K. Upadhyay, L.A. Kumaraswamidhas, Md.Sikandar Azam, and Rolling element bearing
failure analysis: A case study, Case Studies in Engineering Failure Analysis 1 (2013) pages 15–17

[18] Design of machine elements, V.B. Bhandari, McGrawHill, Third Edition, 2010

[19] Vibration and Noise for Engineers, Kewal Pujara, Dhanpat Rai & Co., Third Editiona

[20] SKF Rolling bearing catalogue, SKF India Limited, February 2013