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

net/publication/322117035

Gear Condition Monitoring Technics

Article · December 2017

CITATION READS

1 2,093

1 author:

Adeepa Palihawadana
La Trobe University
7 PUBLICATIONS   3 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Building a formula three electric racing car View project

All content following this page was uploaded by Adeepa Palihawadana on 29 December 2017.

The user has requested enhancement of the downloaded file.

Gear Condition Monitoring Technics

Contents
1.0 Types of Gears…………………………………………………………… 02
2.0 Gear Failures…………………………………………………………....... 06
3.0 Gear Condition Monitoring………………………………………………. 09
3.1 Signal Analyzing methods………………………………………… 10
3.2 Detection of gear failures………………………………………….. 15
3.2.1 Gear Wear………………………………………………... 16
3.2.2 Gear Misalignment Identification………………………… 18
3.2.3 Gear Crack………………………………………………... 20
3.2.4 Gear Pitting……………………………………………….. 22

Reference…………………………………………………………………….. 24

~1~
Gear Condition Monitoring Technics

1.0 Types of Gears

Combustion engines and electric motors are the main sources of kinetic power generators in a
machine. To get a motion power has to be transmitted to wheels or other moving parts. For the
transmission, different technics has been developed. Among them belt driven systems, chain
driven systems and gear systems are the most common transmission technics. In belt driven
systems slipping is a common phenomenon. In the chain driven systems when the chain drive is
unduly stretched the chain velocity fluctuates. But in gear slipping will not be happen and gears
transmit the exact velocity ratio.

Figure 01

Belt Drive System

Figure 02

Chain Drive System

~2~
Gear Condition Monitoring Technics

Mechanical gears have some special advantages than the other power transmission technics. They
can be listed as follows,

1. Gears transmit exact velocity.

2. Large power can be transmitted.
3. Can used for small center distance shafts.
4. Efficiency is higher than the other methods of transmission.
5. Gears can be arranged in small place also to transmit power.

Since there are so many advantages of gears there are some disadvantages also. Gear
manufacturing process cost is much higher than the other power transmission technics other than
that gear transmitting is an effective method of transmitting power.

Gears can be categories in to several types. Some of them are listed below,

1. Spur gear
2. Helical gear
3. Herringbone gear
4. Bevel gear
5. Rack and pinion
6. Internal spur gear

Spur gears are the most commonly used gears in machines. Because of its low cost and easy
manufacturing process than the other gear types. These types of gears are used for parallel shafts.
Spur gears are mostly used on motor vehicle gearboxes, machine tool drives, electric motor gear
boxes etc. Childs, Peter R.N.. (2014)

Figure 03

Spur Gear

~3~
Gear Condition Monitoring Technics

In helical gears gear teeth are arranged with an angle. Typically, the helix angles of the helical
gear lie between 150 – 300. Like spur gears helical gears are also used with parallel shafts. Noise
levels are much lower than the spur gear because teeth in mech makes point contacts always rather
than line contacts in spur gears. These gears need special tools to manufacture. Therefore, the
manufacturing costs are higher than the spur gear. These types of gears are normally used in heavy
duty applications such as turbine drivers, locomotive gear boxes etc.. Childs, Peter R.N.. (2014).

Figure 04

Helical Gear

Bevel gear is a special gear than can be able to change the direction of the motion. Bevel gears
have a conical shape and when connecting bevel gears ono to one ratio applied. In motor trans
mission differential, for valve control etc bevel gears are used. Childs, Peter R.N.. (2014)

Figure 05

Bevel Gear

~4~
Gear Condition Monitoring Technics

Worm gears are another type of gears that can be used to reduce large speeds between two
perpendicular and non-intersecting shafts. Because of high speed reduction considerable frictional
heat is generated so continuous lubrication is necessary for the worm gear applications. Worm
gears are used in many applications such as steering gear, winch blocks, rotary tables etc.. Childs,
Peter R.N.. (2014).

Figure 06

Worm Gear

Rack and pinion mechanism is also a special gear type. This type of gears are used to convert the
rotary motion to straight line motion. Rack and pinion mechanism is widely used in machinery
tools and automotive applications such as vehicle steering rack. Childs, Peter R.N.. (2014).

Figure 07

Rack and Pinion Gear Mechanism

~5~
Gear Condition Monitoring Technics

2.0 Gear Failures

Gear damages happens in different forms in gears (Rajagopalan et al., 2006). There are several
types of gears available. Therefore, the same damage in each gear type identification is not similar.
Gear failures can be categorized into several groups and most common failures are gear tooth
failures and gear misalignment failures.

In modern transmission systems because of weight reduction and cost reduction gear systems are
designed to be light weight. So, Failure of gears is a common phenomenon in modern transmission
systems (Guo, Shi and Yi, 2016). Gear wear and tear happens mostly on automobiles and other
machines. Specially in automobiles wear and tear happen even the vehicle maintained properly.
Specially in transmission control gears are subjected to high shocks when gear changes and
therefor wear, heat and noise happens (Rajagopalan et al., 2006).

Several types of gear failures can be happened in a gear system. Bending failure, pitting, scoring,
abrasive wear, case cracking, corrosive wear and misalignment are some of common failures
happen in gears (Rajagopalan et al., 2006) (Guo, Shi and Yi, 2016).

Bending failure happens if the acting load is greater than beam strength of the gear tooth. Due
than unbalance force bending occurs in gear tooth’s. To avoid such bending failures of gear teeth
face width of the gear has to be adjusted to withstand the acting load or the gear material has to
be changed.

Figure 08

Bending Failure

~6~
Gear Condition Monitoring Technics

Pitting failure happens on the surface of the gear teeth. It is a fatigue failure and happens when
the surface contact stress is higher that the endurance limit of the material. Dynamic load between
the gear tooth the gear should be less than the rupture of the tooth surface.

Figure 09

Pitting Failure

Scoring happens when systems runs on lack of lubricants. When there is excessive surface
pressure on gear tooth excessive heat is generated. Therefore, it allows to metal to metal contact
which leads to local welding. It results tearing of local welded surface. This type of failures can
be minimized by providing proper flow of lubricants and by adjusting parameters such as speed,
pressure (Rajagopalan et al., 2006).

Figure 10

Scoring Failure

~7~
Gear Condition Monitoring Technics

Wear is the most common gear failure occurs in gears. Wear can be divided as local wear,
distributive wear and gear backlash (Rajagopalan et al., 2006). If one or two gear teeth are
damaged due to wear it is called local wear. Distributive wear occurs when there is a sliding
motion between gear teeth (Guo, Shi and Yi, 2016). Gear backlash failure is due to thinning of
the gear teeth and it leads to improper gear operations (Rajagopalan et al., 2006). If the lubricant
of gears contaminated with the foreign particles such as dirt, dust gear wear increases. So, by
providing filters for the lubricate and by adding high viscus lubricants can avoid wear.

Figure 10

Gear wear

Gear corrosion is another special type of gear failure. Corrosion of gears happens mainly due to
contact of gears with corrosive elements such as water and other additives existing in the lubricant
oil. To minimize or reduce the corrosive wear anti-corrosive additives has to be added to the
lubricants or minimize the contact of corrosive elements can be done.

Figure 11

Corrosive wear

~8~
Gear Condition Monitoring Technics

3.0 Gear Condition Monitoring

As mention in the above gears have different failures. Early identification of such failures is
essential because if a machine gear fails it might damage other important parts also. So, to save
cost, increase the machine life time and to maintain procedures of the machine, premature failure
detection of gears needed.

Faults identification and prediction is done by analyzing vibration signals taken from the machine
gear box casing (Dalpiaz, Rivola and Rubini, 2017). Acetometers are used to convert the vibration
signals into electrical signals to analyze the data (Lebold, et al., 2000). To analyse the vibration
data signal various signal processing method were developed. Those methods were developed on
time domain, frequency domain and time-frequency domain (Guo, Shi and Yi, 2016).

Analysis of vibration signals is done by comparing previous signals and currents signals. Before
a signal to be analyzed signal conditioning has to be done (Lebold, et al., 2000). The vibration
data signal might contain noises, DC offsets and other unwanted signal. Before the signals is to
be processed those unwanted signals have to be removed from the vibration data set. There are
several signal conditioning techniques that used in signal processing. Signal correction, low pass
filtering, amplifying, time synchronize averaging and mean value removal are some of those
signal conditioning methods (Lebold, et al., 2000).

In gear vibration tooth meshing frequency, harmonics of the meshing frequency and side bands
are the most important parameters in gear vibration analysis. By analyzing the changes of
amplitudes and number of side bands faults conditions can be predicted and identified (Dalpiaz,
Rivola and Rubini, 2017). But analysis of such components of the vibration signal is difficult in
simple spectral analysis in early stage of gear faults. So different methods of signal analyzing have
been developed by researches (Dalpiaz, Rivola and Rubini, 2017) to overcome those scenarios.

A single crack on the surface of the gear wheel create a unusual signal in the vibration signal
which is mounted on the gear box casing (Dalpiaz, Rivola and Rubini, 2017). Larger the failure
much longer and considerable bursts generate in the vibration signal (Guimarães & Viana Duarte,
2000). Vibration signals are not measured when the machine is accelerating. Because of the
acceleration shaft speeds wary and therefore the frequency components so the vibration signals
will be changed.

~9~
Gear Condition Monitoring Technics

3.1 Signal Analyzing Methods

Signal processing methods of gear fault detection can be divided into five main groups (Lebold,
et al., 2000). They are,

1. Raw signal (RAW)

- Calculated from the conditioned or mean removal signal from the sensor
- RMS & Delta RMS, Crest factor, Enveloping and Demodulation are some of
Raw signal processing technics.

2. Time synchronous averaged signal (TSA)

- Used to extract repeated signals from the noisy raw signals.
- Interpolation factor, the number of revolutions and number of averages changes
the TSA.
- FM0 is an example for TSA

3. Residual signal (RES)

- Use time synchronized average signal to calculate RES.
- Residual signal contains time synchronous average signal with main gear tooth
meshing frequency with removed harmonics of the meshing frequency.
- Not all the harmonics are removed.
- NA4 is a residual signal method.

4. Difference signal (DIF)

- In the difference signal preprocessing method, an average signal has been taken
by removing the original gear frequency signal by the fault signal.
- All the primary meshing, shaft frequencies and their harmonics are removed in
this DIF method.
- M6A and M8A are these types of signal processing methods

~ 10 ~
Gear Condition Monitoring Technics

5. Band Pass Mesh signal (BPM)

- In this method of preprocessing signal is filtered by using a band pass filter
around primary gear mesh frequency and including side band frequencies.
- Then Hilbert transform is applied to the signal.
- After than getting the magnitude of the signal the analysis can be done.
- NB4 is an example for the BPM preprocessing.

Below table shows some details about various preprocessing technics that using in gear fault
detection condition monitoring.

Signal
Analyzing Description Gear Faults can be Identified
Method
RMS and ✓ Root mean square and delta root ✓ Cannot be identified the
Delta RMS mean square are time analysis exact failure of the
method of a vibration signal. component. But can
✓ Delta RMS is the difference of detect the noise deferent
the present and the previous of the original gear and
RMS values. the faulting gear.

𝑁
1
𝑅𝑀𝑆 = √ × ∑ 𝑥𝑛2
𝑁
𝑛=1

Kurtosis ✓ By kurtosis normal distribution ✓ Gear wears and breaks

of vibration signal data is can be identified by
compared with fourth moment of using this method.
the distribution.
✓ Flatness of the distribution is
calculated
∑𝑁
𝑛−1(𝑦(𝑛) − 𝜇)
4

𝑁 × (𝜎 2 )2

~ 11 ~
Gear Condition Monitoring Technics

Crest Factor ✓ Tooth breakage on a

𝑃𝑒𝑎𝑘 𝐿𝑒𝑣𝑒𝑙 gear
𝐶𝑟𝑒𝑠𝑡 𝐹𝑎𝑐𝑡𝑜𝑟 =
𝑅𝑀𝑆 ✓ Can be identified early
✓ Crest factor is the ratio between
stage gear damages
the input signal peak levels and
the RMS level of the signal.
✓ Crest factor will increase when
the when there are peaks in the
time series data signal.
✓ Impulsive vibration effects
pattern of a signal can be
identified
✓ Crest factor for normal operation
of gear is around 2-6. More than
6 relates to a machine failure.
Enveloping ✓ In enveloping acetometer data is ✓ low levels of defected
fed to a band-pass filter in the frequency pulses of an
beginning of the processing to accelerometer signal can
filter the unwanted noises. be identified by using
✓ Then the signal is fed into a full this method.
wave of half wave rectifier then ✓ That means small gear
the signal is smoothened. failures can be detected
✓ This technic is a signal by this method.
conditioning method
Demodulation ✓ Demodulation is a gear fault ✓ For teeth wear
identification process.
✓ By demodulation can identifies
the gear faults by analyzing
amplitude differences of the gear
mesh frequencies and its
harmonics of the fault gear
signal data and the original gear
signal data.

~ 12 ~
Gear Condition Monitoring Technics

✓ First the raw data of the

accelerometer signal is high
passed at 85%ofthe GMF (gear
mesh frequency)
✓ Low pass filtering is done after
that for percentage of 115% of
the GMF.
✓ Then the signals are analyzed by
using power spectral diagram
FM0 ✓ FM0 can be found by dividing ✓ Major tooth faults
peak to peak average by the ✓ Heavy wear
summation of the amplitude at ✓ Miner tooth damages of
the gear meshing frequency. the gear cannot be found
✓ For heavy wear, the FM0 value by using FM0 analysis
is a constant.
✓ In major tooth failures peak to
peak level increases.
𝑃𝑃𝐴
𝐹𝑀0 =
∑𝑛𝑛=1 𝐴(𝐹)

NA4 ✓ Onset of gear damage

𝑁 × ∑𝑁
𝑖=1(𝑟𝑖 − 𝑟)
4
𝑁𝐴4 =
1 2
∑𝑁 𝑁 2
𝑚 × 𝑗=1[∑𝑖=1(𝑟𝑖𝑗 − 𝑟𝑗 ) ]
✓ Calculated by dividing the fourth
statistical moment by the current
signal runtime average variance

FM4 ✓ Calculated by fourth normalized ✓ Developed to detect

statistical moment. damages on limited no
∑𝑁
𝑖=1(𝑑𝑖 − 𝑑)
4 of gear teeth
𝐹𝑀4 = 2
[∑𝑁 2
𝑖=1(𝑑𝑖𝑗 − 𝑑) ]

~ 13 ~
Gear Condition Monitoring Technics

M6A and M8A ✓ More sensitive to peaks ✓ Detection of the surface

✓ Applied to the difference signal damage
✓ Same as the FM4
✓
𝑁 2 ∑𝑁
𝑖=1(𝑑𝑖 − 𝑑)
6
𝑀6𝐴 = 3
[∑𝑁 2
𝑖=1(𝑑𝑖𝑗 − 𝑑) ]

𝑁 3 ∑𝑁
𝑖=1(𝑑𝑖 − 𝑑)
8
𝐹𝑀4 = 4
[∑𝑁 2
𝑖=1(𝑑𝑖𝑗 − 𝑑) ]

NB4 ✓ Time synchronous averaging is ✓ For all types of failures

done to the raw data signal.
✓ Then the signal is band-pass
filtered within in the rage of its
primary gear meshing frequency
and its harmonics.
✓ After filtering Hilbert transform
is done and the magnitude will
be taken to analysis.
✓ Then the NB4 is determined by
the following equation.
𝑁 × ∑𝑁
𝑖=1(𝐸𝑖 − 𝐸)
4
𝑁𝐵4 =
1 2
× ∑𝑚 [∑ 𝑁
𝑖=1(𝐸𝑖𝑗 − 𝐸𝑗 ) 2]
𝑚 𝑗=1

Table 1

Comparison of signal analyzing methods

(Lebold, et al., 2000)

~ 14 ~
Gear Condition Monitoring Technics

3.2 Detection of gear Failures

As mentioned in the above gears may fault because of different conditions and there are different
gear failures also. Most of the gear failures are happened because of fatigue conditions of the gear
(Onsy, et al., 2013). Early detection of such failures is very important because it will save the
machine, time and money.

Analysis of gear failure is done by using various methods as mentioned in the signal analyzing
methods section previously. Mainly in vibration analysis of a signal cepstrum analysis technique
is used. It is the simplest method of detecting gear faults. But identification of all failures by
cepstrum analysis is somewhat difficult than other methods like Demodulation Analysis,
cyclostationary analysis and wavelet transform (Dalpiaz, Rivola and Rubini, 2017).

Cepstrum analysis is a development of the spectrum analysis. Cepstrum analysis have a good
resolution than the spectrum analysis (Dalpiaz, Rivola and Rubini, 2017). Spectrum analysis is
done by using fast fourier transform and the x axis of the amplitude graph of the spectrum analysis
labeled as frequency. But is cepstral analysis the cepstrum analysis graph represent the logarithmic
value of the amplitude and the frequency of the amplitude spectrum is replaced by quefrency in
the cepstral analysis graph. Quefrency represents modulation period and the reciprocal of the
signals modulating frequency (Dalpiaz, Rivola and Rubini, 2017).

Figure 12

Spectrum analysis of a gear signal

(Guo, Shi and Yi, 2016)

~ 15 ~
Gear Condition Monitoring Technics

Figure 13

Cepstrum analysis of a gear signal

(Dalpiaz, Rivola and Rubini, 2017).

3.2.1 Gear Wear

In a frequency spectrum of a gear which has wear typically gives the rotating shaft frequency of
the gear, gear meshing frequency and harmonics of those frequency except the noise. With respect
to the development of the gear wear in the gear wheel the second and third gear mesh harmonics
increase more significantly than the noises of the shaft rotating frequency and first gear mesh
frequencies (Guo, Shi and Yi, 2016).

The following figure 14 shows a healthy gear frequency spectrum which gear shaft and pinion
shaft rotation frequencies are 20 Hz and 47 Hz respectively (Guo, Shi and Yi, 2016).

Figure 14

Healthy gear fequency spectrum

(Guo, Shi and Yi, 2016)

~ 16 ~
Gear Condition Monitoring Technics

GMF is the gear mesh frequency and other multiples of GMF are side bands of the gear mesh
frequencies. The following table show variations between several wear failure signals that taken
in several intervals of the gear fault. There are six signals different signals. As we can see in the
table 02 the second and the third GMF harmonics of the gear increase when the gear wear increase.
Anyhow, the gear mesh frequency is not changed (Guo, Shi and Yi, 2016).

If gear meshing second and third harmonics are increasing than the GMF states that the gear has
gear wearing problem. When the second and third harmonics amplitude increase the gear wear
also increases (Guo, Shi and Yi, 2016).

Table 02

Wear signal information

(Guo, Shi and Yi, 2016)

Therefore, a gear that has a gear wear fault can be identified by analyzing the frequency spectrum
of the gear if second and third GMF harmonics of the gear increase slightly than the gear mesh
frequency.

~ 17 ~
Gear Condition Monitoring Technics

3.2.2 Gear Misalignment Identification

Frequency spectrum of a healthy gear typically gives shaft frequency, gear mesh frequency and
its harmonics (Guo, Shi and Yi, 2016). Gear misalignment is a considerable problem for gear
because it could be the beginning for some serious faults in gears. So, by minimizing the
misalignment of gears we can protect the gear from serious injurious. The following figure 15 is
about a frequency spectrum of a healthy gear.

Figure 15

Healthy gear frequency spectrum

(Guo, Shi and Yi, 2016)

Table 03

Misalignment signal information

(Guo, Shi and Yi, 2016)

~ 18 ~
Gear Condition Monitoring Technics

When the misalignment occurs in a gear special changes happens in the frequency spectrum. Side
band responses around the shaft frequency GMF and other GMF harmonics are a common
phenomenon for misalignment gears. According to the table 03 there are seven different vibration
signals are analyzed using spectrum analysis. In every signal, the 2nd gear meshing frequency and
its side band are changing other than the other frequency signals such as first GMF and third GMF
and their side bands.

Figure 16

Gear misalignment frequency spectrum

(Guo, Shi and Yi, 2016)

The figure 16 demonstrates how the misalignment faulting gear frequency spectrum differs from
the original gear frequency spectrum. As I mentioned above in the figure 16 we can clearly identify
the fundamental frequencies of the gear and its side bands. As in the table 03 we can see that the
second harmonic GMF frequency has a higher amplitude than the other component of the
spectrum.

Different kind of misalignment failures are there in gears. Anyway, in the frequency spectrum, all
the other frequencies are same as the original gear signal frequency spectrum other than the second
harmonic gear mesh frequency. When the gear misalignment increases it’s second gear harmonic
frequencies and its side bands are also assumed to be increased (Guo, Shi and Yi, 2016).

~ 19 ~
Gear Condition Monitoring Technics

3.2.3 Gear Teeth Crack

Gear crack and broken tooth problem appears mostly because of gear fatigue. If there is a single
crack on a gear tooth, when contacting that tooth there will be a little noise on the frequency
spectrum. If the fault intensity is larger the fault duration will be also larger than the single crack
problem (Guimarães & Viana Duarte, 2000) (Dalpiaz, Rivola and Rubini, 2017). So there will be
a family of sidebands on the frequency spectrum with equal spacing (Dalpiaz, Rivola and Rubini,
2017).

Figure 17

Gear crack, gear with broken tooth,

Gear with two broken teeth

(JIn, et al., 2016)

In the figure 17 shows three types of gear crack failures first one shows a gear with a one tooth
crack. Second one is a gear with one tooth broken. The third one is two teeth broken gear.
Following figure 18 shows Frequency spectrums of healthy gear, one tooth crack gear, one tooth
broken gear and two teeth broken gear. Broken tooth problem happens because of gear tooth
cracks (JIn, et al., 2016).

(fm1, fm2 are gear meshing frequencies / fr1, fr2, fr3 are shaft frequencies of input, pinion and output)

Figure 18.a gives the healthy gear frequencies of the gear. 60Hz, 180 Hz and 300Hz are some
ghost frequencies appeared in the gear box.

In the gear crack case fm1 with sidebands and fr3 rotation frequency appear in the frequency
spectrum than the healthy gear frequency spectrum (figure 18.b).

In the gear one tooth and two teeth broken cases fm1 and fm2 with sidebands and fr3 appears in the
frequency spectrums of two vibration signals (figure 18.c & 18.d).

~ 20 ~
Gear Condition Monitoring Technics

Figure 18

Frequency Spectrums of a). Healthy gear

b). Gear crack, c).gear with broken tooth, d).Gear with two broken teeth

(JIn, et al., 2016)

~ 21 ~
Gear Condition Monitoring Technics

3.2.4 Gear Pitting

Detection of pitting gears is not an easy task for human eye. So further transformations of wave
analysis need for those examinations. FFT is a common gear fault identification method that
generates amount of amplitude for frequency axis. By that we can identify frequency variations
in the faulty and good signals by comparing the difference of those FFT’s.

Figure 19 shows a healthy gear frequency spectrum and frequency spectrum of a same type of
gear that have pitting failure figure 20. The following gear are meshed with a input shat speed of
3000 rpm and 50 Nm of load (Morsy & Achtenova, 2014).

In the healthy gear, constant gear teeth meshing frequency is generated with very low vibration
amplitudes both in pinion gear and gearwheel (Morsy & Achtenova, 2014) (Figure 19).

Gear meshing frequencies increase when a gear fault develops. So when comparing two frequency
spectrums we can clearly see that there is an amplitude incase in the fault signal spectrum. The
value that increased the most is the 2500 Hz frequency.

Figure 19

Healthy gear frequency spectrum

(Morsy & Achtenova, 2014)

~ 22 ~
Gear Condition Monitoring Technics

Increase of a certain frequency amplitude might be a gear misalignment problem. To identify such
problem advanced analysis needed. Unless in the fear misalignment case in hear we can see a huge
amplification in the 2500 Hz frequency. That is because gear meshing is not smoother in this case
than the healthy gear. When gear pitting happens gear, meshing is not smoother for the reason that
is gear teeth surface has small damages. So this case can be identified as a gear pitting case but
for further study advanced technological methods are needed (Morsy & Achtenova, 2014).

Figure 20

Pitting Failure gear frequency spectrum

(Morsy & Achtenova, 2014)

~ 23 ~
Gear Condition Monitoring Technics

References
Childs, Peter R.N.. (2014). Mechanical Design Engineering Handbook. (pp. 317 - 457). Elsevier.
Online version available at:
http://app.knovel.com/hotlink/toc/id:kpMDEH0002/mechanical-design-engineering/mechanical-
design-engineering

DALPIAZ, G., RIVOLA, A. and RUBINI, R. (2017). [online] Available at:

http://diem1.ing.unibo.it/mechmach/rivola/pub13.pdf [Accessed 26 Jul. 2017].

Guimarães, T. A. & Viana Duarte, M. A., 2000. Detection of Faults in Gearboxes Through
Vibration Signals Time-Frequency Analysis. SAE Technical Paper Series, 2000-01-
3176(Congresso SAE Barsil).

Guo, D., Shi, Q. and Yi, P. (2016). Psychoacoustic Analysis of Gear Noise with Gear Faults.
SAE International Journal of Passenger Cars - Mechanical Systems, 9(2).

JIn, X. et al., 2016. A Comparative Study on Vibration- and Current-Based Approaches for
Drivetrain Gearbox Fault Diagnosis. IEEE.

Lebold, M. et al., May 1-4, 2000,. REVIEW OF VIBRATION ANALYSIS METHODS FOR
GEARBOX DIAGNOSTICS AND PROGNOSTICS. 54th Meeting of the Society for Machinery
Failure Prevention Technology, Virginia Beach, pp. p. 623-634.

Morsy, M. E. & Achtenova, G., 2014. Robust Diagnostic Concept for Vehicle Gearbox with
Artificial Pitting Defect in Gear Using Vibration Measurements. SAE TEchnical Paper, Issue
2014-32-0047.

morsy, M. E. & Achtenova, G., 2017. Value of Optimal Wavelet Function in Gear Fault Diagnosis.
SAE Technical Paper, Issue 2017-01-1771.

~ 24 ~
 Gear Condition Monitoring Technics

Onsy, A., Bicker, R. & Shaw, B. A., 2013. Predictive Health Monitoring of Gear Surface Fatigue
Failure Using Model-Based Parametric Method Algorithms; An Experimental Validation. SAE
International.

Rajagopalan, S., Habetler, T., Kral, C. and Pirker, F. (2006). Monitoring and Diagnostics for
Electric Drivetrain Components in HEVs. SAE Technical Paper Series.

~ 25 ~

View publication stats