Fault Detection and Diagnosis

of Gear Transmission System

via Vibration Analysis
Pratesh Jayaswal* and Amit Aherwar**

Vibration analysis for condition assessment and fault diagnosis has a long history of
application to power and mechanical equipment. In the present work, vibration
signals of a gearbox are acquired and analyzed with the help of vibration analysis
techniques. In order to understand the basic principles of vibration condition
monitoring, a test rig with common machine faults (i.e., rolling element bearing
damage, gear failure and shaft misalignment), designed and constructed at Madhav
Institute of Technology and Science, Gwalior, India, was used. The methods used
for extracting and identifying the type of faults are described. It is shown that this
experimental setup provides a good illustration of the practical applications of
basic theory included in vibration analysis and condition monitoring.

Keywords: Gear defects, Condition monitoring,Vibration analysis, Machine fault simulator,

Fault diagnosis

Introduction
Gears are universally used in today’s high speed rotating machinery. They have been
applied in devices ranging from household appliances such as vacuum cleaners and
lawn mowers to sophisticated systems such as aircraft and automobile drive trains.
Presently, fault/wear detection is an important problem associated with high speed
rotating machinery. If a crack or fault is detected in its early stages, corrective action
can be taken promptly. Also, preventive maintenance can be made in advance to
replace the damaged part of the machinery. Thus, the detection of excessive wear in
machine components could be a valuable diagnostic and prognostic tool for today’s
machinery. Currently, there are three types of approaches for the detection of faults
in geared systems: acoustic signal analysis, debris monitoring and vibration analysis.
The vibration-based diagnosis has been the most popular monitoring technique because
of the ease of measurement. There are many vibration-based monitoring techniques
* Reader, Department of Mechanical Engineering, Madhav Institute of Technology & Science, Gwalior, Madhya
Pradesh, India; and is the corresponding author. E-mail: [email protected]
** Lecturer, Department of Mechanical Engineering, Madhav Institute of Technology & Science, Gwalior, Madhya
Pradesh, India. E-mail: [email protected]
© 2011 IUP. All Rights Reserved.
26 The IUP Journal of Mechanical Engineering, Vol. IV, No. 3, 2011

Electroniccopy
Electronic copy available
available at:
at:https://ssrn.com/abstract=2097703
http://ssrn.com/abstract=2097703
currently available for the detection of gear faults. According to the analysis domain,
they can be classified into time domain analysis and frequency domain analysis.
This article describes how the authors designed and developed an affordable
experimental rig capable of simulating common machine faults, namely, gear damage,
shaft misalignment and rolling element bearing damage for students to personally
perform vibration condition monitoring. The data in time and frequency domains
were recorded using a CSI accelerometer Sl. No. 359 and dynamic signal analyzer
2,400 as a processor. Data analysis consisted of comparing the graphs obtained under
each test condition with those expected for the specific machine faults simulated.
Fundamental frequency spikes determined from the graphs were compared with the
theoretical vibration fault signatures.
A wide range of literature on the topic of gear transmission exists, as demonstrated
by the review of Ozguven (1988). Fundamental publications on gears in mesh can be
found in Walker (1938), Harris (1958), Niemann (1970) and Munro (1994). David
et al. (2009) evaluated fault detection effectiveness as applied to gear tooth pitting
fatigue damage. Three common condition indicators (RMS, FM4 and NA4) were
deduced from the time-averaged vibration data and used with the Oil Debris
Monitoring (ODM) to evaluate their performance for gear fault detection. Saravanan
and Ramachandran (2009) dealt with the application of fast single-shot multiclass
proximal support vector machine for fault diagnosis of a gearbox consisting of 24
classes. The condition of an inaccessible gear in an operating machine can be
monitored using the vibration signal of the machine measured at some convenient
location and further processed to unravel the significance of these signals. Keong
et al. (2007) performed an experimental investigation on spur gears in which natural
pitting was allowed to occur. Throughout the test period, Acoustic Emission (AE),
vibration and spectrometric oil samples were monitored continuously in order to
correlate and compare these techniques to natural life degradation of the gears. Combet
and Gelman (2007) proposed a sensor-less algorithm for angular resampling by using
the mesh vibration of a gearbox and to propose a general methodology that could be
used as an automated scheme for condition monitoring of gear systems without the
need of a speed sensor. Liu et al. (2006) presented two new methods (empirical
mode decomposition and Hilbert spectrum) for adaptive analysis of nonlinear and
non-stationary signals. Vecer et al. (2005) have frequently used Condition Indicators
(CIs). The ability of a selected condition indicator to describe the degree of gearing
wear was tested using vibration signal acquired during durability testing of manual
transmission with helical gears. Peng and Chu (2004) presented a summary of the
application of the wavelet in machine fault diagnostics, including the following main

Fault Detection and Diagnosis of Gear Transmission System via Vibration Analysis 27

Electroniccopy
Electronic copy available
available at:
at:https://ssrn.com/abstract=2097703
http://ssrn.com/abstract=2097703
aspects: time-frequency analysis of signals, fault feature extraction, singularity
detection for signals, denoising and extraction of the weak signals, compression of
vibration signals and system identification. Polyshchuk et al. (2002) presented the
development of a novel method in gear damage detection using a new gear fault
detection parameter based on the energy change in the joint time-frequency analysis
of the vibration analysis of the vibration signal. Wilson et al. (2001) experimentally
investigated the sensitivity and robustness of the currently well-accepted techniques:
phase and amplitude demodulation, beta Kurtosis and wavelet transform.The vibration
signal was measured on the gearbox housing and processed online under three filtering
conditions: general signal average, overall residual and dominant meshing frequency
residual. Choy et al. (2004) have presented a comprehensive procedure in predicting
faults in gear transmission system under normal operating conditions. Results from
the Wigner-Ville procedure are compared to results from a variety of signal analysis
techniques which include time domain analysis methods and frequency analysis
methods. Choy et al. (1996) presented a comprehensive procedure to simulate and
analyze the vibration in a gear transmission system with surface pitting, wear, and
partial tooth fracture of the gear teeth. Choy et al. (2003), demonstrate the use of
vibration signature analysis procedures for health monitoring and diagnostics of a
gear transmission system. Dalpiaz et al. (1998) deals with gear condition monitoring
based on vibration analysis techniques. The detection and diagnostic capability of
some of the most effective techniques are discussed and compared on the basis of
experimental results concerning a gear pair affected by a fatigue crack. Decker (2002)
proposed two new detection techniques. These techniques were applied to a series
of three seeded fault crack propagation tests. Ebersbach et al. (2005) have investigated
the effectiveness of combining both vibration analysis and wear debris analysis in an
integrated machine condition monitoring maintenance program. Irem and Edward
(2003) presented helicopter transmission vibration. The data was collected using
triaxial accelerometers, explored in three different directions, analyzed for content,
and then combined using Principal Components Analysis (PCA) to analyze changes
in directionality. The benefits of using triaxial data for vibration monitoring and
diagnostics are explored by analyzing the changes in the direction of the principal
axis of vibration formed using all three axes of vibration. Zakrajsek et al. (1995)
reviewed recent activities in transmission diagnostics research at NASA Lewis Research
Center that were performed in-house, and through NASA and US Army sponsored
grants. Lebold et al. (2000) have focused on the most traditional features used for
machinery diagnostics and prognostics and presented some of the signal processing
parameters that impact their sensitivity.

28 The IUP Journal of Mechanical Engineering, Vol. IV, No. 3, 2011

Electronic copy available at: https://ssrn.com/abstract=2097703

2. Vibration Analysis Techniques
2.1 Time Domain Analysis
The time domain methods try to analyze the amplitude and phase information of the
vibration time signal to detect the fault of gear-rotor-bearing system.The time domain
is a perceptive that feels natural and provides physical insight into the vibration. It is
particularly useful in analyzing impulsive signals from bearing and gear defects with
non-steady and short transient impulses.
2.1.1 Time Waveform Analysis
Prior to the commercial availability of spectral analyzers, almost all vibration analyses
were performed in the time domain. By studying the time domain waveform using
equipments, such as oscilloscopes, oscillographs or vibrographs, it was often possible
to detect changes in the vibration signature caused by faults. Time-waveform analysis
involves the visual inspection of the time-history of the vibration signal. The general
nature of the vibration signal can be clearly seen and distinctions made between
sinusoidal, random, repetitive, and transient events. Non-steady-state conditions are
most easily captured and analyzed using time waveforms. High-speed sampling can
reveal such defects as broken gear teeth and cracked bearing races, but can also
result in extremely large amounts of data being collected - much of which is likely to
be redundant and of little use. The digitized data of 8,202 vibration samples of 1.934
seconds time length is used to develop the time domain signatures via MATLAB
algorithm.
2.1.2 Indices
Indices have also been used in vibration analysis. The peak value, Root Mean Square
(RMS) level and their ratio crest factor are often used to quantify the time signal. The
peak level is not a statistical quantity and hence may not be reliable in detecting
damage in continuously operating systems. The RMS value, however, is more
satisfactory for steady-state applications. The crest factor, defined as the ratio of the
peak value to RMS level, has been proposed as a trending parameter as it includes
both parameters. Crest factors are reliable only in the presence of significant
impulsiveness.
Peak: The peak level of the signal is defined simply as half the difference between the
maximum and minimum vibration levels:
Peak  Max (A) ...(1)

Fault Detection and Diagnosis of Gear Transmission System via Vibration Analysis 29

Electronic copy available at: https://ssrn.com/abstract=2097703

Root Mean Square: The RMS value of the signal is the normalized second statistical
moment of the signal (standard deviation):

 A n 
N 2

RMS  n 1 ...(2)
N
where A(n) is the amplitude of the nth digitized point in the time domain, and N is the
number of points in the time domain.
The RMS of the signal is commonly used to describe the steady-state or continuous
amplitude of a time varying signal.
Crest Factor: The crest factor is defined as the ratio of the peak value to the RMS of
the signal:

Peak Level
Crest Factor  ...(3)
RMS Level

2.1.3 Statistical Methods

Statistical analysis can also be carried out on time domain data.
Kurtosis: Kurtosis is the normalized fourth statistical moment of the signal. For
continuous time signals, this is defined as:
N

  yn    
4

K n 1 ...(4)
N    2 2

where y(n) is the data; n = 1, 2, 3, ..., N; N is the total number of data samples,  is the
mean; and  is the standard deviation.
The Kurtosis level of a signal is used in a similar fashion to the crest factor, that is,
to provide a measure of the impulsive nature of the signal. Raising the signal to the
fourth power effectively amplifies isolated peaks in the signal.
Skewness: Skewness is a measure of symmetry, or more precisely, the lack of
symmetry. A distribution of dataset is symmetric if it looks the same to the left and
right of the center point. Equation (5) is used to calculate the values of skewness.
N

  y n    
3

S n 1 ...(5)
N   3

30 The IUP Journal of Mechanical Engineering, Vol. IV, No. 3, 2011

Electronic copy available at: https://ssrn.com/abstract=2097703

2.2 Frequency Domain Analysis
The frequency domain methods include Fast Fourier Transform (FFT), Hilbert
transform method and power cepstrum analysis. They use the difference of power
spectral density of the signal due to the fault of gear and/or bearing to identify the
damage of elements. Any real world signal can be broken down into a combination
of unique sine waves. Every sine wave separated from the signal appears as a vertical
line in the frequency domain. Its height represents its amplitude and its position
represents the frequency. The frequency domain representation of the signal is called
the signal. The frequency domain completely defines the vibration. Frequency domain
analysis not only detects the faults in rotating machinery but also indicates the cause
of the defect.
Theoretically, time domain can be converted into frequency domain using the
Fourier transforms and vice versa. The Fourier transform is a generalization of the
complex Fourier series in the limit as L. Replace the discrete An with the
continuous F(k)dk while letting n/Lk. Then change the sum to an integral, and the
equation becomes:

f  x    F  k e 2 i k xdk ...(6)


f  k    F  x e 2 i k xdx ...(7)


Here, Equation (7) is called forward (–i) Fourier transform and Equation (6) is
called the inverse Fourier transform.
2.2.1 Fast Fourier Transform
The FFT is simply a class of special algorithms which implement the Discrete Fourier
Transform (DFT) with considerable savings in computational time. It must be pointed
out that the FFT is not a different transform from the DFT, but rather just a means of
computing the DFT with a considerable reduction in the number of calculations
required.The FFT is a discrete Fourier transform algorithm which reduces the number
of computations needed for N points from 2N 2 to 2N log N, where log is the base-2
logarithm.
The vibration characteristics of any rotating machine are to some extent unique
due to the various transfer characteristics of the machine. In the FFT plot various
peaks with large and small amplitudes are presented corresponding to characteristic
frequencies to show the origin of defects. The spectra of FFT may produce peaks at
identified fault frequencies. These peaks may or may not represent the indicated

Fault Detection and Diagnosis of Gear Transmission System via Vibration Analysis 31

Electronic copy available at: https://ssrn.com/abstract=2097703

fault. One must look for harmonics to determine if the identified frequencies were
generated from the indicated fault.
• If peaks appear at the fundamental fault frequency and also at frequency
two times of fundamental frequency, it shows strong indication of reality of
fault.
• If no peak appears at the fundamental fault frequency but peaks are present
at two, three, and maybe four times the fundamental fault frequency, then
this also represents a strong indication that the indicated fault is valid.
FFT for Determination of the Severity of the Fault:
• One way to determine the fault’s severity is to compare its amplitude with
the past readings taken under consistent conditions.
• Another way is to compare the amplitude to the other readings obtained by
similar machines running under the same conditions. A higher than normal
reading indicates a problem.
3. Experimental Setup
In the present study, an experimental work was carried out on a vibration simulator,
as shown in Figure 1. The test rig consists of a three shaft with gears on the first and
second shafts and two overhang pulleys on the second and third shafts, which are
supported on the bearings. The specifications of the test rig are shown in Tables 1-4.

Figure 1: Experimental Setup

32 The IUP Journal of Mechanical Engineering, Vol. IV, No. 3, 2011

Electronic copy available at: https://ssrn.com/abstract=2097703

 Table 1: Electrical Specifications
S. No. Parameters Value
1. Electrical AC Drives
2. Drive 1 HP Single Phase
3. RPM 1,440
4. Voltage 220 Single Phase, 50 Hz

Table 2: Mechanical Specifications

S. No. Parameters Value
1. Shaft Diameter 28 mm
2. Number of Bearings 6, Ball Bearings
3. Bearing Housing 6, Cast Iron
4. Gear Helical Tooth Gear
5. General Structure Steel Frame

Table 3: Physical/Other Specifications

S. No. Parameters Value
1. Operating Weight 120 kg
2. Size L = 1300 mm, W = 86 mm, H = 218 mm
3. PCD of Gear 210 mm
4. PCD of Pinion 156 mm
5. Helix Angle 35
6. Number of Teeth (Gear) 70
7. Number of Teeth (Pinion) 52
8. Module of Gear 3.08/3
9. Bigger Pulley Diameter 128
10. Half Groove Angle ( ) 22
11. Small Pulley Diameter 63.4 mm
12. Bearing Number and Make GYE25KRRB (Make NRB)
13. First Shaft L = 600, D = 28
14. Second Shaft L = 600, D = 28
15. Third Shaft L = 600, D = 28
16. Wheel Bearing Grease CM/L-2051936

Fault Detection and Diagnosis of Gear Transmission System via Vibration Analysis 33

Electronic copy available at: https://ssrn.com/abstract=2097703

 Table 4: Bearing Specifications
S. No. Parameters Value
1. Number of Rollers 10
2. Outer Diameter 52 mm
3. Inner Diameter 25 mm
4. Pitch Diameter 39 mm
5. Roller Diameter 8 mm

4. Data Acquisition and Analysis Figure 2: (a)Tooth with Crack

(b) Enlarged Image ofTooth
Signals are acquired at intervals of one with Crack (C) Accelerometer
hour, 10 signals per day using CSI Location
accelerometer Sl. No. 359 as pickup and
dynamic signal analyzer 2,400 as processor, (a)
data presenter and keeper with initializing
filter set at 1,656.6 Hz and sampling
frequency at 4,240.89 Hz (Figure 2).
Signals are taken for the gear, with each
signal containing 8,202 sample points for
time domain signal; all the signals are taken
at the shaft speed 1,420 rpm and readings
are taken at point 1 of simulator. FFTs of
signals are analyzed to identify the faults.
The acquired signals from the analyzer (b)
are fed to a Pentium computer through a
PCI-E series data cable. A driver program,
written in C++, is applied to
accommodate the data into MATLAB
environment so as to utilize its signal
processing toolboxes.
With the use of MATLAB software,
RMS, peak value, crest factor, Kurtosis and
skewness of gear condition are calculated.
(c)
The curves are generated to find the
suitability of these techniques, as discussed
in Section 5. A total of 40 signals of the
gear are used to calculate the peak values,
RMS values, crest factor, Kurtosis and
skewness. A total 40 values of the gear are
used to draw five curves. The objective of
curve generation is to check the trend of
these features in gear condition.
34 The IUP Journal of Mechanical Engineering, Vol. IV, No. 3, 2011

Electronic copy available at: https://ssrn.com/abstract=2097703

 The characteristic frequencies and their harmonic frequencies of the gear up to
the third multiple, of which the fault presence is to be identified are shown in the
Table 5.
Table 5: Characteristic Frequencies andTheir Harmonic
Frequencies up toThird Multiple
Characteristic Multiples (Hz)
Frequencies 1st 2nd 3rd
Gear Frequency (Frg) 24 48 72
Pinion Frequency (Frp) 32 64 96
Tooth Mesh Frequency (Fm) 1,657 3,314 4,971

Therefore,
Gear Frequency = Gear speed/60 Hz
(Frg) = 1420/60 = 23.66 Hz
Pinion Frequency = Pinion speed/60 Hz
(Frp) = 1911/60 = 31.85 Hz
Tooth Mesh Frequency = (Frg)  Ng
(Fm) = 31.85  70 = 1656.6 Hz
5. Results and Discussion
The main focus of this paper is to find out the presence of fault and to predict the
nature of the fault in the gear used in this work. The data for the gearbox were
analyzed by using time domain analysis and FFT. Decomposition is done up to 40
nodes, and then for each node, RMS value is calculated and analyzed for finding the
nodes which give significant differences in different faulty or healthy conditions.
5.1 Case 1: Time Domain Analysis
Three major analysis techniques are applied—waveform analysis, indices calculations
and statistical moment’s calculation. The findings are as follows:
5.1.1 Waveform Analysis
The waveform is generated as shown in Figure 3a. The acceleration level of vibration
signals in gear condition is generated in the form of waveform. It is found that the
acceleration level of vibration of gear is much higher than the level of healthy gear
condition.
Fault Detection and Diagnosis of Gear Transmission System via Vibration Analysis 35

Electronic copy available at: https://ssrn.com/abstract=2097703

 Figure 3a: Waveform of Vibration for Duration of 1.934 s of Gearbox

50
Fault
40
30 Alert
Acceleration (g)

20
10

0
–10

–20
–30
0 1,000 2,000 3,000 4,000 5,000 6,000 7,000 8,000
Time (s)

5.1.2 Indices Computation

Peak Level: From Figure 3b, it is observed that some peak values of acceleration
signal are above the alert line. Hence, these values indicate a local defect on gear
tooth.
RMS Level: From Figure 3c, it is observed that some RMS values of acceleration
signal are above the alert line. Hence, these values indicate a local defect on gear
tooth.
Crest Factor: Typical values of crest factor for a gear in a healthy (good) condition
range from 3.4 to 4. From Figure 3d, it is observed that some crest factor values of
acceleration signal are above the alert line. Hence, these values indicate a local defect
on gear tooth.
5.1.3 Statistical Moments
Kurtosis: Typical values of Kurtosis for gear in a healthy (good) condition range
from 3 to 4. From Figure 3e, it is observed that some Kurtosis values of acceleration
signal are above the alert line. Hence, these values indicate a local defect on gear
tooth.

36 The IUP Journal of Mechanical Engineering, Vol. IV, No. 3, 2011

Electronic copy available at: https://ssrn.com/abstract=2097703

 Figure 3b: Curve of PeakValues for Gearbox

60

55

50
Fault
Peak Value

45
Alert
40

35

30

25
0 5 10 15 20 25 30 35 40

Number of Samples

Figure 3c: Curve of RMSValues for Gearbox

13.0

12.5

12.0
Fault
RMS Value

11.5
Alert
11.0

10.5

10.0

9.5
0 5 10 15 20 25 30 35 40

Number of Samples

Fault Detection and Diagnosis of Gear Transmission System via Vibration Analysis 37

Electronic copy available at: https://ssrn.com/abstract=2097703

 Figure 3d: Curve of Crest FactorValues for the Gearbox

5.5

5.0
Fault
4.5
Crest Factor

Alert
4.0

3.5

3.0

2.5
5 10 15 20 25 30 35 40
Number of Samples

Figure 3e: Curve of KurtosisValues for the Gearbox

12

10

8
Kurtosis

6 Fault

Alert
4

0
0 5 10 15 20 25 30 35 40
Number of Samples

Skewness: Typical values of skewness for a gear in a healthy (good) condition range
from 2.5 to 3. From Figure 3f, it is observed that some skewness values of acceleration
signal are above the alert line. Hence, these values indicate a local defect on gear
tooth.
38 The IUP Journal of Mechanical Engineering, Vol. IV, No. 3, 2011

Electronic copy available at: https://ssrn.com/abstract=2097703

 Figure 3f : Curve of SkewnessValues for the Gearbox

4 Fault
Skewness

Alert
3

0
0 5 10 15 20 25 30 35 40
Number of Samples

5.2 Case 2: Frequency Domain Analysis

The spectra of FFT may produce peaks at identified fault frequencies. These peaks
may or may not represent the indicated fault. One must look for harmonics to
determine if the identified frequencies were generated from the indicated fault.
• If peak appears at the fundamental fault frequency and another peak appears
at two times the fundamental frequency, it is a very strong indication that
the fault is real.
• If no peak appears at the fundamental fault frequency but peaks are present
at two, three, and maybe four times the fundamental fault frequency, then
this also represents a strong indication that the indicated fault is valid.
5.2.1 FFT Spectrum
Figure 3g shows the spectrum of defected gear vibration signal. In this figure, the
peaks are found at Fm and its second multiple frequency, but there are some other
peaks due to modulation effect of the signal.
Figure 3h shows the spectrum of defected gear vibration signal. In this figure, the
peaks are found at Frg and Frp and its second and third multiple frequency, but there
are some other peaks due to modulation effect of signal.

Fault Detection and Diagnosis of Gear Transmission System via Vibration Analysis 39

Electronic copy available at: https://ssrn.com/abstract=2097703

 Figure 3g: FFT Spectrum of Vibration Signal for Gearbox

3,500 Fm = Tooth Mesh Frequency

Amplitude (Reactive Unit)

3,000

2,500

2,000

1,500

1,000

500
0
500 1,000 1,500 2,000 2,500 3,000 3,500

Frequency (Hz)

Figure 3h: Zoom View of FFT Spectrum ofVibration Signal for Gearbox
Amplitude (Reactive Unit)

2,000

1,500

1,000

500

0
50 100 150 200 250 300

Frequency (Hz)

These peaks show a strong indication that the gear is in faulty condition and the
nature of fault is chipped tooth pitch point to the top at 2% thickness.

40 The IUP Journal of Mechanical Engineering, Vol. IV, No. 3, 2011

Electronic copy available at: https://ssrn.com/abstract=2097703

Conclusion
In the present work, fault identification techniques have been studied, implemented
and tested for gearbox fault diagnostics. On the basis of analysis and results, the
following are the main conclusions:
• The results of the effect of the proposed time domain feature extraction
demonstrate that this technique extracts the fault-specific features.
– The results further show that the waveform generation in case of multiple
faults at gear contact surfaces is only useful to find the healthy or faulty
condition but not capable of identifying the categories of fault.
– The trends of peak level, RMS level, and crest factors shows that peak
value and RMS level can identify the fault on gear tooth surface. The
crest factor is not able to distinguish the gear defects categories.
– The statistical moment, Kurtosis is capable of identifying the fault
condition, but the skewness trend has not shown any effective fault
categorization ability in this present gear fault condition
• In frequency domain, FFT were able to show the impulses at fault
characteristic frequencies and its multiple frequencies, but other peaks are
also present due to signal modulation effect. By this technique, identification
of fault categories is difficult.
From the analysis results of gear data, it can be seen that the characteristic
frequencies are quite clear in the FFT spectrum of gear vibration signals. FFT spectrum
of vibration signals captured from gearbox is capable to distinguish the fault category.
In the present work, waveform analysis and FFT have been used for gear condition
monitoring. In continuation of this work, the following can be considered for future
research and investigation.
• The short-time Fourier transform, envelop analysis and wavelet transform
of gear vibration signals;
• Features extraction through envelop analysis;
• Development of expert system with these extracted features;
• Other condition monitoring approaches such as temperature monitoring,
noise monitoring and lubricant monitoring; and
• Generalized defect detection.
Acknowledgment: The authors would like to thank Madhav Institute ofTechnology and Science,
Gwalior, India, for providing experimentation facility and support throughout the work.

Fault Detection and Diagnosis of Gear Transmission System via Vibration Analysis 41

Electronic copy available at: https://ssrn.com/abstract=2097703

References
1. Choy F K, Huang S, Zakrajsek J J et al. (1994), “Vibration Signature Analysis of
a Faulted Gear Transmission System”, pp. 1-14, 30th Joint Propulsion Conference
Cosponsored by the AAIA, ASME, NASA.
2. Choy F K, Polyshchuk V, Zakrajsek J J et al. (1996), “Analysis of the Effects of
Surface Pitting and Wear on the Vibration of Gear Transmission System”, Tribology
International, Vol. 29, No. 1, pp. 77-83.
3. Choy F K, Mugler D H and Zhou J (2003), “Damage Identification of Gear
Transmission Using Vibration Signatures”, Journal of Mechanical Design, Vol. 125,
June, pp. 394-403.
4. Combet F and Gelman L (2007), “An Automated Methodology for Performing
Time Synchronous Averaging of a Gearbox Signal Without Speed Sensor”,
Mechanical Systems and Signal Processing, Vol. 21, pp. 2590-2606.
5. Cooley J W and Tukey J W (1965), “An Algorithm for the Machine Calculation
of Complex Fourier Series”, Mathematics of Computing, Vol. 19, pp. 297-301.
6. Dalpiaz G, Rivola A and Rubini R (1998), “Gear Fault Monitoring Comparison
of Vibration Analysis Techniques”, pp. 623-632, Italian National Research Council.
7. David G L, Dempsey P J, Heath G F and Shanthakumaran P (2009), “Gear Fault
Detection Effectiveness as Applied to Tooth Surface Pitting Fatigue Damage”,
Glenn Research Center, NASA.
8. Decker Harry J (2002), “Gear Crack Detection Using Tooth Analysis”, National
Aeronautics and Space Administration, John H Glenn Research Center, Lewis
Field, Cleveland, Ohio.
9. Ebersbach S, Penga Z and Kessissoglou N J (2005), “The Investigation of the
Condition and Faults of a Spur Gearbox Using Vibration and Wear Debris Analysis
Techniques”, International Conference on Wear of Materials, Wear, Vol. 260,
Nos. 1 & 2, pp. 16-24.
10.Irem T Y and Edward M H (2003), “Analysis of Triaxial Vibration Data for Health
Monitoring of Helicopter Gearboxes”, Journal of Vibration and Acoustics, Vol. 125,
pp. 120-128.
11.Keong Tan Chee, Irvinga P and Mba D(2007), “A Comparative Experimental
Study on the Diagnostic and Vibration and Spectrometric Oil Analysis for Spur
Gears”, Mechanical Systems and Signal Processing, Vol. 21, No. 1, pp. 208-233.

42 The IUP Journal of Mechanical Engineering, Vol. IV, No. 3, 2011

Electronic copy available at: https://ssrn.com/abstract=2097703

12.Lebold M, McClintic K, Campbell R et al. (2000), “Review of Vibration Analysis
Methods for Gearbox Diagnostics and Prognostics”, Proceedings of the 54th Meeting
of the Society for Machinery Failure Prevention Technology, pp. 623-634, Virginia Beach,
VA.
13.Liu B, Riemenschneider S and Xu Y (2006), “Gearbox Fault Diagnosis Using
Empirical Mode Decomposition and Hilbert Spectrum”, Mechanical Systems and
Signal Processing, Vol. 20, No. 3, pp. 718-734.
14.Peng Z K and Chu F L (2004), “Application of the Wavelet Transform in Machine
Condition Monitoring and Fault Diagnostics: A Review with Bibliography”,
Mechanical Systems and Signal Processing, Vol. 18, pp. 199-221.
15.Polyshchuk V V, Choy F K and Braun M J (2002), “Gear Fault Detection with
Time Frequency Based Parameter NP4”, International Journal of Rotating Machinery,
Vol. 8, No. 1, pp. 57-70.
16.Saravanan N and Ramachandran K I (2009), “A Case Study on Classification of
Features by Fast Single-Shot Multi Class PSVM Using Morlet Wavelet for
Fault Diagnosis of Spur Bevel Gear Box”, Expert Systems with Applications, Vol. 36,
pp. 10854-10862.
17.Stewart R M (1977), “Some Useful Data Analysis Techniques for Gearbox
Diagnostics”, University of Southampton Report MHM/R/10/77, July.
18.Vecer P, Kreidl M and Šmíd R (2005), “Condition Indicator for Gearbox
Condition Monitoring System”, Acta Polytechnica , Vol. 45, No. 6, pp. 35-43.
19.Wilson Q W, Fathy I and Golnaraghi F M (2001), “Assessment of Gear Damage
Monitoring Techniques Using Vibration Measurements”, Mechanical Systems and
Signal Processing, Vol. 15, No. 5, pp. 905-922.
20.Zakrajsek J J, Handschuh R F, Lewicki D G and Decker H J (1995), “Detecting
Gear Tooth Fracture in a High Contact Ratio Face Gear Mesh”, Proceedings of
the 49th Meeting of the Mechanical Failure Prevention Group, Virginia Beach,
VA, April 18-20.

Reference # 60J-2011-08-02-01

Fault Detection and Diagnosis of Gear Transmission System via Vibration Analysis 43

Electronic copy available at: https://ssrn.com/abstract=2097703

Copyright of IUP Journal of Mechanical Engineering is the property of IUP Publications and its content may
not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written
permission. However, users may print, download, or email articles for individual use.

Electronic copy available at: https://ssrn.com/abstract=2097703