Seminar III Report Final

Vibration-based fault diagnosis of spur bevel gear box using fuzzy technique
1. INTRODUCTION
A faulty gear system could result in serious damage if defects occur to one of the
gears during operation condition. Early detection of the defects, therefore, is
crucial to prevent the system from malfunction that could cause damage or entire
system halt. Diagnosing a gear system by examining the vibration signals is the
most commonly used method for detecting gear failures. The conventional
methods for processing measured data contain the frequency domain technique,
time domain technique, and time frequency domain technique. These methods
have been widely employed to detect gear failures. The use of vibration analysis
for gear fault diagnosis and monitoring has been widely investigated and its
application in industry is well established. This is particularly reflected in the
aviation industry where the helicopter engine, drive trains and rotor systems are
fitted with vibration sensors for component health monitoring. The raw vibration
signal in any mode from a single point on a machine is not a good indicator of the
health or condition of a machine. Vibration is a victorial parameter with three
dimensions and requires to be measured at several carefully selected points. [1]
Fig. 1. Flowchart for bevel gear box health diagnosis.
TSSM’s, Bhivarabai Sawant College of Engineering & Research, M.E. (Mechanical) (Design Engineering) 1
Vibration analysis can be carried out using Fourier transform techniques like
Fourier series expansion (FSE), Fourier integral transform (FIT) and discrete
Fourier transform (DFT) .After the development of large-scale integration (LSI)
and the associated microprocessor technology, fast Fourier transform (FFT)
analyzers became cost effective for general applications. The raw signatures
acquired through a vibration sensor needed further processing and classification of
the data for any meaningful surveillance of the condition of the system being
monitored. Artificial neural network (ANN), support vector machine (SVM) and
in the problem at hand, the nature of the fault itself is fuzzy in nature. Fuzzy
classifier models the physical problem under study more closely the flow chart of
the fault diagnostic. The signals obtained are processed further for machine. [1]
2. LITERATURE SURVEY
Amit Aherwar, Md. Saifullah Khalid has discussed on Vibration Analysis

Techniques for Gearbox Diagnostic. Gears are important element in a variety of
industrial applications such as machine tool and gearboxes. An unexpected failure
of the gear may cause significant economic losses. For that reason, fault diagnosis
in gears has been the subject of intensive research. Vibration signal analysis has
been widely used in the fault detection of rotation machinery. The vibration signal
of a gearbox carries the signature of the fault in the gears, and early fault detection
of the gearbox is possible by analyzing the vibration signal using different signal
processing techniques. In this paper, a review is made of some current vibration
analysis techniques used for condition monitoring in gear fault. After the review
of literature on gear fault analysis, the following points are concluded.
(i) Gearbox vibration signals are usually periodic and noisy. Time-frequency
domain average technique successfully removes the noise from the signal and
captures the dynamics of one period of the signals.
(ii) Time domain techniques for vibration signal analysis as waveform generation,
Indices (RMS value, Peak Level value, and crest factor) and overall vibration
level do not provide any diagnostic information but may have limited application
in fault detection in simple safety critical accessory components. The statistical
moment as kurtosis is capable to identify the fault condition but skewness trend
has not shown any effective fault categorization ability in this present gear fault
condition.
(iii) Spectral analysis may be useful in the detection and diagnosis of shaft faults.
(iv) In frequency domain analysis, it is concluded that FFT is not a suitable
technique for fault diagnosis if multiple defects are presents on gearbox. The
envelope analysis and Power Spectrum Density techniques have shown a better
representation for fault identification. The Hilbert Transform and PSD techniques
are suitable for multiple point defect diagnostics for condition monitoring.
(v) Synchronous signal averaging has the potential of greatly simplifying the
diagnosis of shaft and gear faults (i.e., the safety critical failures) by providing
significant attenuation of nonsynchronous vibrations and signals on which ideal
filtering can be used. Further development needs to done on the implementation of

synchronous averaging techniques and the analysis of results.
(vi) Expert system based on ANN and fuzzy logic can be developed for robust
fault categorization with the use of extracted features from vibration signal. [2]
Shaban Ghavami Jolandan, Hossein Mobli, Hojat Ahmadi, Mahmoud Omid,
Seyed Saeid Mohtasebi has discussed on Fuzzy-Rule-Based Faults Classification
of Gearbox Tractor. They presented a fault Classification method based on a fuzzy
inference system. The vibration signal from a piezoelectric transducer is captured
for the following conditions of MF 285 gearbox: ‘Healthy Gearbox’ (H), ‘Gear
with tooth face worn’ (W) and ‘Gear with tooth face broken’ (B), at three working
speed (700, 1500 and 1800 rpm). The features of signal were extracted using
descriptive statistic parameters. The output of the J48 algorithm is a decision tree
that was employed to produce the crisp if-then rule and membership function sets.
Results showed that the total classification accuracy for 700, 1500 and 1800 rpm
conditions were about 79.63%, 100% and 96.3%.Fuzzy inference system (FIS)
have been presented to perform fault diagnosis of a gearbox. The implementation
of J48-FIS based classifier requires two consecutive steps. Firstly, method
Correlation-based Feature Selection (CFS) and the J48 algorithm are utilized to
select the relevant features in the data set obtained from feature extraction part.
The output of the J48 algorithm is a decision tree that is employed to produce the
crisp if-then rule and membership function sets. Secondly, the structure of the FIS
classifier is defined based on the obtained rules, which were fuzzified in order to
avoid classification surface discontinuity. The classification results and statistical
measures are then used for evaluating the J48-FIS model. The total classification
accuracy for 700, 1500 and 1800 rpm conditions were 79.63%, 100% and 96.3%
respectively. Therefore, fault diagnosis is more reliable in higher speeds of
gearbox using this technique. The results indicate that the proposed J48-FIS model
can be used in diagnosing gearbox faults and developing an online condition
monitoring test. [3]
S. Devendiran and Dr. K. Manivannan has worked on Vibration Based
Condition Monitoring and Fault Diagnosis Technologies for Bearing and Gear
Components. All machines with moving parts give rise to sound and vibration and
each machine has a specific vibration signature related to the construction and the
state of the machine. If the state of the machine changes the vibration signature
will also change and a change in the vibration signature can be used to detect
incipient defects before they become critical. This is the basics of vibration based
condition monitoring (CM) methods. The CM technique is based on detecting the
presence of a fault, diagnosing the root cause of the fault, assessing its level of
severity and making arrangements for its correction. A broad review of the state-
of-art of condition monitoring and fault diagnosis techniques has been carried out
for improving the accuracy and ability of condition monitoring and prognosis
systems for bearing and gear components. Recent advances in automated
condition monitoring technologies have significantly improved the diagnosis and
prognosis of bearing and gear component health condition and it reduced the
operation and maintenance costs of the rotating machinery. However, accurate
fault detection and reliable prognosis in rotating machinery remain challenging
due to the congenital complexities in their mechanical systems and operation
conditions. In order to accomplish robust condition monitoring for bearing and
gear component, the following research areas need to be further strengthened. [4]
TANG Jiafu WANG Dingwei, Richard Y K FUNG, Kai-Leung Yung has
worked for Understanding of Fuzzy Optimization: Theories and Methods. A brief
summary on and comprehensive understanding of fuzzy optimization is presented.
This summary is made on aspects of fuzzy modeling and fuzzy optimization,
classification and formulation for the fuzzy optimization problems, models and
methods. The importance of interpretation of the problem and formulation of the
optimal solution in fuzzy sense are emphasized in the summary of the fuzzy
optimization. An extensive study on fuzzy optimization is conducted in this paper,
which leads to the following concluding remarks that the basic procedure of fuzzy
optimization problems is to transform a fuzzy model into a crisp one, and the most
important thing is how to make this transformation have an appropriate and
reasonable interpretation. During the transformation, the first thing to do is to
understand the problem and interpret the optimal solution, and then try to an
appropriate interpretation, and propose some concepts and theory to support the
interpretation, and finally transform the fuzzy model into a crisp one. The
interpretation and formulation are the key constituent parts of the approaches, and
they also bridge the gap between the fuzzy optimization and the application in
solving practical problems. [5]
3. OBJECTIVES
The objectives of this seminar are:
 To understand fuzzy logic technique.

 To study experimental set up of fault diagnosis of faulty gear box.
 To study vibration signals from faulty gear box.
 To study generation of rules for various gearbox conditions and
discussions.
 To study decision tree and use of J48 Algorithm.
4. EXPERIMENTAL STUDIES
The fault simulator with sensor is shown in Fig.2 and the pinion and gear shown
in Fig. 3. A variable speed DC motor (0.5HP) with speed up to 3000 rpm is
the basic drive. A short shaft of 30 mm diameter is attached to the shaft of the
motor through a flexible coupling; this is to minimize effects of misalignment
and transmission of vibration from the motor. The shaft is supported at it sends
through two roller bearings. From this shaft the drive is transmitted to the bevel
gear box by means of a belt drive.
Fig.2. Fault simulator setup.
Fig. 3. Inner view of the bevel gear box.

The gear box is of dimension 150 mm x 170 mm x 120 mm and the full
lubrication level is 110 mm and half lubrication level is 60 mm. SAE 40 oil was
used as a lubricant. An electromagnetic spring-loaded disc brake was used to load
the gear wheel. A torque level of 8 N-m was applied at the full-load condition.
The various defects are created in the pinion wheels and the mating gear wheel is
not disturbed. With the sensor mounted on top of the gear box vibrations signals
are obtained for various conditions. The selected area on the top of the gearbox for
mounting the sensor is made flat and smooth to ensure effective coupling between
the sensor and the gearbox. The sensor used is a piezoelectric accelerometer
(Dytran model) which is mounted on the flat surface using direct adhesive
mounting technique.
The accelerometer is connected to the signal conditioning unit (DACTRAN FFT
analyzer), where the signal goes through the charge amplifier and an analogue-to-
digital converter (ADC). The vibration signal in digital form is fed to the
computer through a USB port. The software RT Pro series that accompanies the
signal-conditioning unit is used for recording the signals directly in the
computer’s secondary memory. The signal is then read from the memory and
processed to extract different features. [1]
Table1: Details of faults under investigation
Gears Fault description Dimension(mm)

G1 Good –
G2 Gear tooth breakage 8
(GTB)
G3 Gear with crack at 0.8 x0.5
root (GTC)
G4 Gear with face wear 0.5
Table2: Gear wheel and pinion details
Parameter Gear wheel Pinion wheel
No. of teeth 35 25
Module (mm) 2.5 2.5
Normal pressure angle (°) 20 20
Shaft angle (°) 90 90
Top clearance (mm) 0.5 0.5
Addendum (mm) 2.5 2.5
Whole depth (mm) 5.5 5.5
Chordal tooth thickness
(mm) 3.93 0.150 3.92 0.110
Chordal tooth height 2.53 2.55
(mm)
Material EN8 EN8
5. EXPERIMENTAL PROCEDURE
In the present study, four pinion wheels whose details are as mentioned in Table1
were used. One was a new wheel and was assumed to be free from defects. In the
other three pinion wheels, defects were created using electron discharge machine
(EDM) in order to keep the size of the defect under control. The details of the
various defects are depicted in Table 2 and its views are shown in Fig.4 a, b & c.
The size of the defects is in line with work reported. The vibration signal from
the piezoelectric pick up mounted on the test bearing was taken, after allowing
initial running of the bearing for some time. The sampling frequency was
12,000Hz and sample length was 8192 for all speeds and all conditions. The
sample length was chosen arbitrarily, however, the following points were
considered. Statistical measures are more meaningful, when the number of
samples is more. On the other hand, as the number of samples increases the
computation time increases. To strike a balance, sample length of around 10000
was chosen. In some feature extraction techniques, which will be used with the
same data, the number of samples is to be 2n. The nearest 2n–10,000 is 8192 and
hence, it was taken as sample length. Many trials were taken at the set speed and
vibration signal was stored in the data file. The raw vibration signals acquired for
various experimental conditions form the gear box using FFT are shown in Fig.5.
[1]
Fig.4 (a). View of good pinion wheel.
Fig.4 (b). Views of pinion wheel with face wear (GFW).
Fig.4 (c). View of pinion wheel with tooth breakage (GTB).
Fig.5. (a) Vibration signals for good pinion wheel under different
lubrication and loading conditions. (b) Vibration signal for pinion wheel
with teeth breakage under different lubrication and loading conditions.
(c) Vibration signal for pinion wheel with crack at root under different
lubrication and loading conditions. (d) Vibration signals for pinion
wheel with teeth face wear under different lubrication and loading
conditions.
6. FAULT DETECTION AND DIAGNOSIS FROM VIBRATION ANALYSIS

Diagnostics is understood as identification of a machine's condition/faults on the
basis of symptoms. Diagnosis requires a skill in identifying machine's condition
from symptoms. The term diagnosis is understood here similarly as in medicine. It
is generally thought that vibration is a symptom of a gearbox condition. Vibration
generated by gearboxes is complicated in its structure but gives a lot of
information. We may say that vibration is a signal of a gearbox condition. To
understand information carried by vibration one have to be conscious/ aware of a
relation between factors having influence to vibration and a vibration signal. In
order to detect (and diagnosis) an impending failure, a good understanding of the
evidence relating to the failure mode and methods of collecting and quantifying
the evidence is needed. Although many faults may be easily detectable by
physical examination of a component, using techniques such as microscopy, X-
ray, dye penetrates, magnetic rubber, etc., these methods usually cannot be
performed without removal of, and in some cases physical damage to, the
component. Whilst physical examination techniques still play a critical role during
manufacture, assembly and overhaul, they are impractical in an operational large
transmission system and other (non-intrusive) fault detection methods need to be
employed for routine monitoring purposes. Most modern techniques for gear
diagnostics are based on the analysis of vibration signals picked up from the
gearbox casing. The common target is to detect the presence and the type of fault
at an early stage of development and to monitor its evolution, in order to estimate
the machine’s residual life and choose an adequate plan of maintenance. It is well
known that the most important components in gear vibration spectra are the gear
meshing frequency (GMF) and its harmonics, together with sidebands due to
modulation phenomena. The increment in the number and amplitude of such
sidebands may indicate a fault condition. Moreover, the spacing of the sidebands
is related to their source. source identification and fault detection from vibration
signals associated with items which involve rotational motion such as gears, rotors
and shafts, rolling element bearings, journal bearings, flexible couplings, and
electrical machines depend upon several factors: (i) the rotational speed of the
items, (ii) the background noise and/or vibration level, (iii) the location of the
monitoring transducer, (iv) the load sharing characteristics of the item, and (v) the
dynamic interaction between the item and other items in contact with it. The main
causes of mechanical vibration are unbalance, misalignment, looseness and
distortion, defective bearings, gearing and coupling in accuracies, critical speeds,
various form of resonance, bad drive belts, reciprocating forces, aerodynamic or
hydrodynamic forces, oil whirl, friction whirl, rotor/stator misalignments, bent
rotor shafts, defective rotor bars, and so on. Some of the most common faults that
can be detected using vibration analysis are summarized in Table 1. [2]
Table 3: Some typical faults and defects that can be detected with vibration
analysis
Item Fault
Gear Tooth meshing faults, misalignment,
cracked and / or worn teeth, Eccentric
gear
Rotors and shaft Unbalance, Bent shaft, misalignment,
Loose components, Rubs, Critical speed,
Cracked shaft, Blade loss, Blade
resonance.
Rolling element bearing Pitting of race and ball / roller, Spalling,
other rolling elements defect.
Flexible coupling Misalignment, Unbalance
Unbalanced magnetic pulls, Broken/
Electrical machines damaged rotor bars, Air gap geometry
variations, Structural and foundation
faults, Structural resonance, Piping
resonance.
7. FEATURE EXTRACTION
Statistical analysis of vibration signals yields different primary and secondary
parameters. Research work reported uses these in combinations to elicit
information regarding bearing faults. Such procedures use allied logic often
based on physical Considerations. A fairly wide set of these parameters is
selected as a basis for our study, as detailed below.
a. Mean b. Standard error c. Median d. Standard deviation e. Sample variance
f. Kurtosis g. Skewness h. Range i. Minimum j. Maximum k. Sum.
All the above mentioned statistical features were extracted for the
vibration signals obtained for various conditions and fed as an input to J 48
algorithm for selecting the best features which classify the different fault
conditions.
7.1. Descriptive statistics

The statistical features are explained below.
7.1.1 Standard deviation

This is a measure of the effective energy or power content of the
vibration signal and clearly indicates deterioration in the bearing
condition. The following formula was used for computation of standard
deviation.
Standard deviation=√(𝑛∑𝑥 2 − (∑𝑥)2 )/(𝑛(𝑛 − 1))

7.1.2Skewness
Skewness characterizes the degree of a symmetry of a distribution around its
mean. The below shown expression was used to calculate the skewness, where
‘n’ is the sample size and‘s’ is the sample standard deviation
7.1.3Kurtosis
Kurtosis indicates the flatness or the spikiness of the signal. Its value is very
low for good bevel gearbox and high for faulty gearbox due to the spiky
nature of the signal.
where‘s’ is the sample standard deviation

7.1.4.Standard error
Standard error is a measure of the amount of error in the prediction of ‘y’ for
an individual ‘x’ in th regression, where x and y are the sample meansand ‘n’ is
the sample size.
Standard error often predicted as
7.1.5. Sample variance

It is variance of the signal point sand the following formula was used for
computation of standard variance
:
7.1.6. Range
It refers to the difference in maximum and minimum signal point values for a
given signal.
7.1.7. Minimum value
It refers to the minimum signal point value in a given signal. As the gear parts
(crack, breakage, face wear) get degraded, the vibration levels seem to go high.
Therefore, it can be used to detect faulty gears.
7.1.8. Maximum value

It refers to the maximum signal point value in a given signal.
7.1.9. Sum
It is the sum of all signal point values in a given signal.
8. USINGJ48ALGORITHMINTHEPRESENTWORK
A standard tree induced with c5.0 (or possibly ID3 or c4.5) consists of a number
of branches, one root, a number of nodes and a number of leaves. One branch is a
chain of nodes from root to a leaf; and each node involves one attribute. The
occurrence of an attribute in a tree provides the information about the importance
of the associated attribute. A Decision Tree is a tree based knowledge
representation methodology used to represent classification rules. J48 algorithm
(A WEKA implementation of c4.5 Algorithm) is a widely used one to construct
Decision Trees .The Decision Tree algorithm has been applied to the problem
under discussion. Input to the algorithm is set of statistical features of vibration
signatures. It is clear that the top node is the best node for classification. The other
features in the nodes of Decision Tree appear in descending order of importance.
It is to be stressed here that only features that contribute to the classification
appear in the Decision Tree and others do not. Features, which have less
discriminating capability, can be consciously discarded by deciding on the
threshold. This concept is made use for selecting good features. The algorithm
identifies the good features for the purpose of classification from the given
training data set, and thus reduces the domain knowledge required to select good
features for pattern classification problem. The decision trees shown in Figs. are
for various lubrication and loading conditions of different faults compared with
good conditions of the pinion gear wheel.
Based on the output of J 48 algorithm, the decision tree various statistical
parameters are selected for the various conditions of the gearbox. The values
appearing between various nodes in the decision tree are used for generating the
fuzzy rules to classify the various conditions of the gearbox under study. [1]
8.1Applicationofdecisiontreeforfeatureselection
The algorithm has been applied to the problem under discussion for feature
selection. Input to the algorithm is the set of statistical features extracted from
raw vibration signatures; the output is the Decision Tree. It is clear there from
that the top node is the best node for classification. The other features appear in
the nodes in Decision Tree in descending order of importance. It is to be stressed
here that only features that contribute to the classification appear in the Decision
Tree and others do not. The level of contribution is not same and all statistical
features are not equally important. The level of contribution by individual feature
is given by a statistical measure within the parent thesis in the Decision Tree.
The first number in the parent thesis indicates the number of data points that can
be classified using that feature set. These second number indicates the number of
samples against this action. If the first number is very small compared to the total
number of samples, then the corresponding features can be considered as outlier
sand hence ignored. Features that have less discriminating capability can be
consciously discarded by deciding on the threshold. This concept is made use of
in selecting good features.
9. METHODOLOGY ADOPTED FOR FUZZY CLASSIFICATION
Fig.6. Methodology of classification using Fuzzy.

Fuzzy Logic provides a precise approach for dealing with uncertainty. Fuzzy
inference is a method that interprets the values in the input vector and, based on
some set of rules, assigns values to the output vector. The point of fuzzy logic is
to map an input space to an output space, and the primary mechanism for doing
this is a list of ‘if then’ statements called rules. Rules are the inputs for building
a fuzzy inference engine. The
methodologyadoptedforfuzzyclassificationisshowninFig.6.Allrulesareevaluatedi
nparallel,andtheorderoftherulesisunimportant.Therealworlddatadonothavesharpl
ydefinedboundarieswhereinformationisoftenincompleteorsometimes unreliable.
Inquestfor precision,scientistshavegenerallyattemptedto
manipulatetherealworldintoartificialmathematicalmodelsthatmakenoprovisionfo
rgradation.BecauseFuzzyLogicprovidesthetoolstoclassifyinformationintobroad,c
oarsecategorizationsorgroupings,ithasinfinitepossibilitiesforapplicationwhichha
veproventobemuchcheaper,simplerandmoreeffectivethanothersystemsinhandling
complexinformation.
Fortheproblemathand,theconditionofthegearbox,goodorfaultyisbasicallyfuzzy
innature.Allthefaultsdonotoccurinthegearboxinstantly.Itcomesgradually.Inthatca
se,thereisnothresholdvalue(crispdata)basedonwhichthedecisiononthecondition
ofthegearboxcanbetaken.(Whethergearboxisnowgoodorfaulty).Theproblemsoft
hiskindcanbemodeledusingfuzzylogicmoreclosely. [ 1]
10. MEMBERSHIP FUNCTION

Amembershipfunction(MF)isacurvethatdefineshoweachpointintheinputspacei
smappedtoamembershipvalue(ordegreeofmembership)between0and1.Observing
thevaluesofthefeature,basedonwhichthebranchesoftheDecisionTreearecreatedfor
differentconditionsofthegearbox,themembershipfunctionsforthecorrespondingfe
aturesaredefined.Therearefourpossibleoutcomesfromafuzzyclassifier,namely:go
odbevelgear,bevelgearwithtoothbreakage(GTB),bevelgearwithcrackatrootofthet
ooth(GTC),andbevelgearwithfacewearoftheteeth(TFW)forvariousloadingandlub
ricationconditions.Hence,fourmembershipfunctionsaredefinedwithequalrangefor
theoutput.[1]
11. RULE GENERATION FROM DECISION TREE

Artificialneuralnetworkandsupportvectormachineareusedtogenerateruleforclassifi
cationproblems.Inthisstudy,DecisionTreeisusedforthatpurpose.DecisionTreeshow
stherelationbetweenfeaturesandtheconditionofthegearbox.Tracingabranchfromthe
rootnodeleadstoaconditionofthegearbox(ReferFigs.7,13,18,22,27and31)anddecod
ingtheinformationavailableinabranchintheformof‘if then’statement gives the rules
for classification using fuzzy for
variousconditionsofthegearbox.Hencetheusefulnessofthedecisiontreeinformingth
erulesforfuzzyclassificationisestablished.[1]
Fig.7.DecisiontreefromJ48Algorithmfordry-lubricationno-loadcondition
12. GENERATION OF RULES FOR VARIOUS GEARBOX CONDITIONS

AND DISCUSSIONS
Theprecedingsectiondescribeshowtheclassificationhasbeencarriedoutusingfuzzyte
chnique.
9.1.Dry-lubricationandno-loadcondition
FromFig. 7wecanseethat
standarderror,kurtosisandvarianceplayadecisiveroleinclassifyingthevariousgearb
oxfaultsunder drylubricationandno-loadcondition.Thisoutputof thedecisiontreeis
usedtodesignthemembershipfunctionforfuzzyclassifierasshowninFig.8–10 as
follows.Amembershipfunction(MF)isacurvethatdefineshoweachpointintheinputs
paceismappedtoamembership
value(ordegreeofmembership)between0and1.Inthepresentstudy,trapezoidalmem
bershipfunctionisused.Theselectionofthismembershipfunctionistosomeextentarb
itrary.However,thefollowingpointswereconsideredwhileselectingmembershipfu
nction.TheDecisionTreefortheselectedthreefeaturesisshowninFig.7.Observingth
evaluesofthefeature,basedonwhichthebranchesoftheDecisionTreeiscreated,them
embershipfunctionsforallthreefeaturesaredefinedforstandarderror,kurtosisandvar
iance,respectively
Fig.8.Membershipfunctionfor‘ ‘standarderror”.
Fig.9.Membershipfunctionfor‘‘kurtosis”.
Fig.10.Membershipfunctionfor‘‘variance”.
9.1.1. Rulesdesignedforthedry-lubricationandno-loadcondition
1. If(stderrisnotstderr)then(Output1isGTC).
2. If(stderrisstderr)and(kurtosisisKur)then(Output1isGOOD).
3. If(stderrisstderr)and(kurtosisisnotKur)and(varianceisVar)then(Output1isGTB)
.
4. If(stderrisstderr)and(kurtosisisnotKur)and(varianceisnotVar)then(Output1isT
FW).
ThemembershipvalueoftheconditionbeingGTCiswhen thestandarderror valueis
less than or equal
to0.000175(fromFig.7)whichisthethresholdvalue.Hence,uptothisthresholdvaluethe
membershipfunctiongeneratesthevalue‘0’andafterwardsitincreaseslinearly(assump
tion).Thetrapezoidalmembershipfunctionsuitsthisphenomenonandhenceitwasselect
edtomapeachpointintheinputspacetoamembershipvalue.To
review,thethresholdvaluesaregivenbydecisiontreeandtheslopeisdefinedbytheuserth
roughheuristics.Thethresholdvalue(0.000175)isdefinedbasedontherepresentativetra
ining dataset. Ifstandarderror value is less thanor equal to 0.000175, a
membership function which is defined on a 0–1 scale gives a value of 0 which
means that it is not a standard error. If threshold value is greater than 0.000175,
the membership function generates a value of 1. Similarly membership functions
for other features are designed accordingly and shown in Figs. 9 and 10.
Therearefourpossibleoutcomesfromafuzzyclassifier,namely:Good,GTC,GTBand
TFWHence,fourmember-
shipfunctionsaredefinedwithequalrangeandshowninFig.11
Fig. 11. Membership functions for condition. (Output).
13. FUZZYINFERENCEENGINE
Afterdefiningmembershipfunctionsandgeneratingthe‘if-
then’rules,thenextstepistobuildthefuzzyinferenceengine.Thefuzzytoolboxav
ailableinMATLAB7wasusedforbuildingfuzzyinferenceengine.Eachrulewas
takenatatimeandusingmembershipfunctionsandfuzzyoperatorstherulesweree
ntered.Theruleswereobtainedfromatrainingdataset(150trialsineachcondition).
Fortestingthebuiltmodelaportionofthedata(100trialsineachcondition)calledt
estingdatawaskeptaside.Usingthetestingdata,thefuzzyinferenceenginewasev
aluatedanditsperformancewaspresentedasconfusionmatrixinTable3.Thediag
onalelementsintheconfusionmatrixtable(3)showthenumberofcorrectlyclassifie
dinstances.Inthefirstrow,thefirstelementshowsthenumberofdatapointsbelong
ingto‘good’classandclassifiedbyfuzzylogicas‘good’.Thesecondelementsho
wsthenumberofdatapointsbelongingto‘GTC’classandclassifiedbyfuzzylogic
as‘GTC’.Thethirdelementshowsthenumberofdatapointsbelongingto‘GTB’c
lassandclassifiedbyfuzzylogicas‘GTB’.Thefourthelementshowsthenumbero
fdatapointsbelongingto‘TFW’classandclassifiedbyfuzzylogicas‘TFW’.Tabl
e3illustratesthepowerfulnessofthefuzzyrulesdesignedwiththeaidofthedecisiont
reesbytheauthors.[1]
Table 4. Number of correctly classified instances

condition GOOD GTC GTB TFW
GOOD 100 0 0 0
GTC 0 100 0 0
GTB 0 0 100 0
TFW 0 0 0 100
Fig.12. Ruleviewerforoneofthetestdata.
Fig.12illustratestheapplicationoftherulesdesigned.Hereeachrowcorrespondstoeachr
uleasdiscussedabove.Thefirstthreeblocksinrowsrepresentsthemembershipfunctiono
fstandarderror,kurtosis,and
variance,respectively.Thefourthblockcorrespondstothemembershipfunctionsforout
putasshowninFig.11.Withthehelpofsampleinputsforstandarderror,kurtosisandvaria
ncetherulesaretestedasfollows,forasampleinputofstandarderroras0.0005,kurtosisas
10andvarianceas
0.005whichsatisfiesthesecondrulecompletelyandthecorrespondingoutputconditioni
sGOOD,whichisshownintheoutputblockofthesecondrowintheruleviewershowninFi
g.12.
9.2.Dry-lubrication and full-load condition
Fig.13. Decision tree from J48 Algorithm for dry-lubrication full-load condition
10.1.1. Rules designed for the dry-lubrication and full-load condition. (See Figs.
14–16.)
1. If (stderr is stderr) then (output1 is GTC).
2. If (stderr is not stderr) and (variance is var3) then (output1 is GTB).
3. If (kurtosis is kur) and (variance is var2) then (output1 is GOOD).
4. If (kurtosis is not kur) and (variance is var1) then (output1 is TFW).
Fig.14. Membershipfunctionfor‘‘stderr”
Here there are three membership functions to represent three threshold values of
variance in the decision tree.
Fig. 17 is the rule viewer for the following test data. If stderr = 0.0005, kurtosis =
10, Variance = 0.005, then the output is 6.25, i.e., the condition is GTB.
Fig.15. Membershipfunctionfor‘‘kurtosis”.
Fig.16. Membershipfunctionsfor‘‘variance”.
10.1.2. Confusionmatrix
Table 5. Confusionmatrix
Condition GOOD GTC GTB TFW
GTC 0 100 0 0
GTB 0 0 100 0
GTC 0 100 0 0
TFW 0 0 0 100
Fig.17. Ruleviewerforoneofthetestdata.
10.2. Half-lubricationandno-loadcondition
Fig.18.shows the DecisiontreefromJ48Algorithmforhalf-lubno-loadcondition
Fig.18. DecisiontreefromJ48Algorithmforhalf-lubno-
loadcondition.
10.2.1.Rulesdesignedforhalf-lubricationandno-loadcondition
SeeFigs.19–21.
1. If(stderrisstderr1)then(output1isGTC).
2. If(stderrisnotstderr2)then(output1isGOOD).
3. If(stderrisstderr2)and(kurtosisiskur)then(output1isGTB).
4. If(stderrisstderr2)and(kurtosisisnot kur)then(out-put1isTFW).
Fig.19. Membershipfunctionsfor‘‘stderr”
Fig.20. Membershipfunctionfor‘‘kurtosis”.
Fig.21.RuleviewerforoneofthetestdataIfstderr=0.0005,kurtosis=10.Theoutputis
6.25,hencetheconditionisGTB.
10.2.2. Confusionmatrix
Table 6. Confusionmatrix

GOOD 100 0 0 0
GTC 0 100 0 0
GTB 0 0 100 0
TFW 0 0 0 100
10.3. Half-lubrication and full-load condition

Fig.22.shows the DecisiontreefromJ48Algorithmforhalf-lubfull-loadcondition.
Fig.22. DecisiontreefromJ48Algorithmforhalf-lubfull-loadcondition.
10.3.1. Rulesdesignedforhalf-lubricationandfull-loadcondition
Figs.23–26
3. If(stderrisstderr2)and(kurtosisisnotkur)then(out-put1isTFW).
4. If(stderrisstderr2)and(kurtosisiskur)and(minimumismin)then(output1isGTB)[1
]
Fig.23. Membershipfunctionsfor‘‘stderr”.
Fig.24. Membershipfunctionsfor‘‘kurtosis”
Fig.25. Membershipfunctionsfor‘‘minimum”
Fig.26.Ruleviewerforoneofthetestdata.IfStderr=0.0005,kurtosis=1
0,minimum=0,output1=6.25,i.e.,theconditionisGTB.
10.3.2. Confusion matrix

10.3.3. Table 7:Confusion matrix

GOOD 100 0 0 0
GTC 0 100 0 0
GTB 0 0 100 0
TFW 0 0 0 100
10.4. Full-lubricationandno-loadcondition
SeeFig.27.
Fig.27. DecisiontreefromJ48Algorithmforfull-lubno-loadcondition.
10.4.1. Rulesdesignedforfull-lubricationandno-loadcondition
Figs.28–30.
3. If(stderrisstderr2)and(varianceisvar)then(output1is GTB).
4. If(stderrisstderr2)and(varianceis notvar)then(output1is TFW).
5. If(stderrisstderr2)and(kurtosisiskur)and(minimumisnotmin)then(output1
isTFW).
Fig.28. Membershipfunctionfor‘‘stderr.”
Fig.29. Membershipfunctionsfor‘‘variance”.
Fig.30.Ruleviewerforoneofthetestdata.Ifstderr=0.0005,varia
nce=0.005,thenoutput1=6.25.

Table 8. Confusion matrix
GOOD 96 0 0 4
GTC 0 100 0 0
GTB 10 0 90 0
TFW 0 0 0 100
10.5. Full-lubrication and full-load condition

See Fig. 31.
Fig.31. DecisiontreefromJ48Algorithmforfull-lubfull-loadcondition.
10.5.1. Rules designed for full-lubrication and full-load condition

Figs. 32 and 33.
1. If (stderr is stderr1) then (output1 is GTC).
2. If (stderr is not stderr3) then (output1 is GOOD).
3. If (stderr is stderr2) then (output1 is GTB).
4. If (stderr is stderr3) then (output1 is TFW).
Fig. 32. Membership function for ‘‘stderr”.
Fig. 33. Rule viewer for one of the test data. If stderr = 0.0002 then output1 =
1.25, ie, the condition is Good.
Table 9. Confusion matrix

GOOD 100 0 0 0
GTC 0 100 0 0
GTB 0 0 100 0
TFW 0 0 0 100
14. SUMMERY AND CONCLUSION:
FaultdiagnosisofGearboxisoneofthecoreresearchareasinthefieldofconditionmo
nitoringofrotatingmachines.Theworkconducted,proposingthemethodofthegearbo
xfaultidentificationbasedonfuzzylogictechnique,showsthegreatpotentialityandthe
strongabilitytoclassifyandidentifymachineryfaults.Thisworkhasinvestigatedtheu
seofbasicfuzzylogicprincipleasafaultdiagnostictechniqueforspurbevelgearbox.T
heworkconductedhasdemonstratedthe potentialoffuzzylogicto
classifythelikelyfaultconditionswhicharerepresentinrotatingmachinery.Fromthes
tudycarriedoutandpresentedinthispaper,thediagnosistechniquebasedonthefuzzylo
gicprincipleisfoundtobepracticalfortheconditionrecognitionofthegearbox.Alsothi
sworkbringsoutthepotentialofdecisiontreestogeneratetherulesautomaticallyfromt
hefeaturesetwhichprovestobeagreatassetingeneratingfuzzyrules.This
workhasoutlinedtheprocedureoffuzzydiagnosistechniquebyusingthecharacteristi
cvariableswhichrepresentaparticularrunningconditionofthegearboxtodetermineth
efuzzymembershipfunction.
REFERENCES
[1].N. Saravanan , S. Cholairajan, K.. Ramachandran,“Vibration-based fault diagnosis of spur

bevel gear boxusing fuzzy technique”, Expert Systems with Applications 36 (2009) 3119–
3135,Available online at www.sciencedirect.com
[2]. Amit Aherwar, Md. Saifullah Khalid, “VIBRATION ANALYSIS TECHNIQUES FOR
GEARBOXDIAGNOSTIC: A REVIEW”,International Journal of Advanced Engineering
Technology E-ISSN 0976-3945.
[3].Shaban Ghavami Jolandan, Hossein Mobli, Hojat Ahmadi, Mahmoud Omid, Seyed Saeid
Mohtasebi “Fuzzy-Rule-Based Faults Classification of Gearbox Tractor”,WSEAS
TRANSACTIONS on APPLIED and THEORETICAL MECHANICS .
[4].S. Devendiran Dr. K. Manivannan ,“Vibration Based Condition Monitoring and Fault
Diagnosis Technologies For Bearing and Gear Components-A Review”, International Journal
of Applied Engineering Research ISSN 0973-4562 Volume 11, Number 6 (2016) pp 3966-
3975.
[5].TANG Jiafu, WANG Dingwei Richard, Y K FUNG Kai-Leung
Yung,“UNDERSTANDING OF FUZZY OPTIMIZATION:THEORIES AND
METHODS” Vol. 17 No. 1 Journal of Systems Science and Complexity Jan., 2004
[6].Sajid Hussain , “Gearbox Health Monitoring by Evolutionary Optimization”,Journal of
Industrial and Intelligent Information Vol. 3, No. 3, September 2015.
[7]. H Shah and H Hirani,“Online condition monitoring of spur gears”,ONLINE CM
Submitted 27.11.12 Accepted 04.04.14
[8]. Persi Pamela., Gayathri. P and N. Jaisankar ,“A Fuzzy Optimization Technique for
thePrediction of Coronary Heart Disease Using Decision Tree”,Persi Pamela. et.al /
International Journal of Engineering and Technology (IJET).

Seminar III Report Final

Uploaded by

Copyright:

Available Formats

Seminar III Report Final

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Seminar III Report Final

Uploaded by

Copyright:

Available Formats

Vibration-based fault diagnosis of spur bevel gear box using fuzzy technique

Fig. 1. Flowchart for bevel gear box health diagnosis.

Amit Aherwar, Md. Saifullah Khalid has discussed on Vibration Analysis

filtering can be used. Further development needs to done on the implementation of

The objectives of this seminar are:

 To understand fuzzy logic technique.

Fig.2. Fault simulator setup.

Fig. 3. Inner view of the bevel gear box.

Gears Fault description Dimension(mm)

Table2: Gear wheel and pinion details

Parameter Gear wheel Pinion wheel

Fig.4 (a). View of good pinion wheel.

Fig.4 (b). Views of pinion wheel with face wear (GFW).

Fig.4 (c). View of pinion wheel with tooth breakage (GTB).

6. FAULT DETECTION AND DIAGNOSIS FROM VIBRATION ANALYSIS

7.1. Descriptive statistics

7.1.1 Standard deviation

Standard deviation=√(𝑛∑𝑥 2 − (∑𝑥)2 )/(𝑛(𝑛 − 1))

where‘s’ is the sample standard deviation

7.1.5. Sample variance

7.1.8. Maximum value

9. METHODOLOGY ADOPTED FOR FUZZY CLASSIFICATION

Fig.6. Methodology of classiﬁcation using Fuzzy.

10. MEMBERSHIP FUNCTION

11. RULE GENERATION FROM DECISION TREE

12. GENERATION OF RULES FOR VARIOUS GEARBOX CONDITIONS

Fig. 11. Membership functions for condition. (Output).

Table 4. Number of correctly classified instances

9.2.Dry-lubrication and full-load condition

1. If (stderr is stderr) then (output1 is GTC).

2. If (stderr is not stderr) and (variance is var3) then (output1 is GTB).

3. If (kurtosis is kur) and (variance is var2) then (output1 is GOOD).

4. If (kurtosis is not kur) and (variance is var1) then (output1 is TFW).

variance in the decision tree.

Condition GOOD GTC GTB TFW

10.3. Half-lubrication and full-load condition

10.3.2. Confusion matrix

Condition GOOD GTC GTB TFW

4. If(stderrisstderr2)and(varianceis notvar)then(output1is TFW).

10.4.2. Confusion matrix

Condition GOOD GTC GTB TFW

10.5. Full-lubrication and full-load condition

10.5.1. Rules designed for full-lubrication and full-load condition

Fig. 32. Membership function for ‘‘stderr”.

10.5.2. Confusion matrix

Table 9. Confusion matrix

14. SUMMERY AND CONCLUSION:

[1].N. Saravanan , S. Cholairajan, K.. Ramachandran,“Vibration-based fault diagnosis of spur

You might also like