machine tools of Development non-contact structural health monitoring systemfor Technology Journal of Applied Researchand

(1)

Journal of Applied Research and Technology

www.jart.ccadet.unam.mx JournalofAppliedResearchandTechnology14(2016)245–258

Original

Development of non-contact structural health monitoring system for machine tools

Deepam Goyal

^∗

, B.S. Pabla

DepartmentofMechanicalEngineering,NationalInstituteofTechnicalTeachers’Training&Research(NITTTR),Chandigarh160019,India Received22September2015;accepted4June2016

Availableonline3August2016

Abstract

Inthiseraofflexiblemanufacturingsystems,areal-timestructuralhealthmonitoring(SHM)isparamountformachiningprocesseswhichare ofgreatrelevancetoday,whenthereisaconstantcallforbetterproductivitywithhighqualityatlowprice.Duringmachining,vibrationsare alwaysbroughtforthbecauseofmechanicaldisturbancesfromvarioussourcessuchasanengine,asound,andnoise,amongothers.ASHM systemprovidessignificanteconomicbenefitswhenappliedtomachinetoolsandmachiningprocesses.Thisstudydemonstratesanoncontact SHMsystemformachinetoolsbasedonthevibrationsignalcollectedthroughalow-cost,microcontrollerbaseddataacquisitionsystem.The examinationtestsofthisdevelopedsystemhavebeencarriedoutonavibrationrig.Thereadingshavealsobeencalibratedwiththeaccelerometer tovalidatetheproposedsystem.Thedevelopedsystemresultsinquickmeasurement,enablesreliablemonitoring,andiscosteffectivewithno needtoalterthestructureofthemachinetool.Itisexpectedthatthesystemcanforewarntheoperatorfortimelybasedmaintenanceactionsin additiontoreducingthecostsofmachinedowntimeandacquiringequipmentswithreductionincomplexityofmachinetools.

Keywords:Noncontactstructuralhealthmonitoring;Vibrationsignatures;Microcontroller;Dataacquisitionsystem

1. Introduction

Theincreaseincomplexityofmachinetools,togetherwith theincreaseinthecostof acquiringequipmentsandmachine downtime,hasattractedtheattentionofresearcherstowardthe developmentof non contactstructural healthmonitoring sys- tems.Thenewdevelopmentsintheglobalmarkethaveincreased thecompetitionbytransformingmodernmanufacturingtechnol- ogyfrommassproductiontoleanmanufacturingforenhancing SHMthroughsensorintegrationandonlinefaultdiagnosis.The purposeofSHMistoavoidwastefulactivitiestooptimizepro- fitabilityofproducts andservices,toimprove theinformation obtainedabouttheconditionofthemachinetoolsbeingused, thestatusoftheworkpiecebeingprocessed,andinthemanu- facturingprocessitself.Themonitoringofindustrialoperations andequipmentisanessential componentofacriticalscheme that drives manufacturing industries towardbeing leaner and

∗Correspondingauthor.

E-mailaddress:[email protected](D.Goyal).

PeerReviewundertheresponsibilityofUniversidadNacionalAutónomade México.

morecompetitive(Frankowiak,Grosvenor,&Prickett,2005).

Machiningisthemostwidelyusedprocessinthemanufactur- ingindustrythatremovesunwantedmaterialfromaworkpiece intheformofchipstoachievethedesiredproductshape,size, accuracy, and surface quality. US industries spend US $100 billion annually tomachine metals because the hugebulk of manufactured partsrequire machining atsome phaseintheir production (De Garmo, Black, &Kohser, 1997). Monitoring machinehealthplaysareallysignificantpartincreatinghigh- quality,low-costcomponents.Intermsoftheleanmanufacturing concept,regularlyscheduledmachinemaintenanceandtooling replacementarevitalinkeepingthemachinerunningatoptimum conditionstoavoidhumanandeconomiclosses.Astoolingis quiteexpensive,ideally,cuttingtoolsshouldbemaximallyuti- lizedtoreducemanufacturingcosts,sorecentresearchfocuses on the integration of online tool monitoring system method that can be performed automatically. Rotating machines are consideredas thebackbone ofnumerous industries,fromgas turbinesforproducingelectricitytoturbo-machineryforgener- atingpowerintheaerospaceindustry,andthenitisimportantto runthesemachinessafelyovertimetopreventanycatastrophic failures.

http://dx.doi.org/10.1016/j.jart.2016.06.003

(2)

1.1. Vibrationasadiagnostictool

Diagnosisisthe processof determiningmachinecondition from its signs or symptoms to detect what the fault is and whichcomponentof the systemisaffected. Variousmonitor- ing approacheslike directandindirecttechniques canbeput ontoanyapplicationbyadaptationsandalterations,toaccom- plish anacceptable level of consequences.Direct techniques, such as visionand opticalmethods, measurethe actual geo- metric changes inthe cuttingtool. These techniques are not widely used in real time operation due to permanent con- tactbetweenthetoolandworkpiece,andalsothepresenceof coolantandchips.Indirecttechniques,suchasforce,vibration, temperatureandacousticemissions,areeasytoimplementto correlatethesensorsignalswiththemachinehealth.Researchers haveusedmeasurementsofspindlemotorcurrents,feed,forces (Nouri,Fussell,Jerard,Gao,&Linder,2012;Saglam&Unuvar, 2003), acoustic emission (Houshmand, Kannatey-Asibu, &

Herrin, 1995; Niu, Wong, Hong, &Liu, 1998; Srinivasa Pai

&RamakrishnaRao,2002),vibrations(Berger,Minis,Harley, Rokni, & Papadopoulos, 1998; Dilma & Lister, 2000; El- Wardany,Gao,&Elbestawi,1996),toestimatethetool wear.

In contrast to the aforementioned variables, vibration is one of the features of modern mechanical machinerythat isnow continuously monitoredinmanysignificant applications.The commonlymonitoredvibrationsignalsareacceleration,velocity,anddisplacementforthepurposeofdeterminingmachinery integrity. Thevibrationsignaturesof the machineapprisethe operatortomakeacrucialdecisionfor conditionbasedmain- tenance.Theamplitudeofthesesignaturesgivesanindication oftheacutenessofthefault,whilstthefrequencycanindicate the originof the defect(Peng& Kessissoglou,2003).Vibra- tion monitoring is generallycarried outby analyzing signals collectedontherunningmachine.Thesemeasurements,which representinfactsomepartsofthemachine,aredifficulttointer- pretduetothenatureofthevibrationsignalandtheinterferences.

Itemergesaneedtodevelopamonitoringanddiagnosissystem ofhighleveldatatoprovidepreciseextractionofdataaboutthe healthofthemachine.

The various vibration monitoring methods and signal processingtechniquesforstructuralhealthmonitoringhasbeen reviewed (Goyal & Pabla, 2015a, 2015b). Vibration analysis ismostprevalenttechniqueusedformonitoring,detectingand analyzing the machine condition in an industry due to non- destructiveinnaturethatallowssustainablemonitoringwithout any interferinginthe operation.The variouscondition moni- toringindicatorsforfaultdiagnosisofdynamicmachineshave been discussed (Goyal, Pabla,& Dhami, 2016).Areview of theutilizationofsoftcomputingtechniquesincondition-based maintenance has been introduced (Goyal, Pabla, Dhami, &

Lachhwani,2016).

1.2. Structuralhealthmonitoring

ASHMsystemisacontinuousandanautonomoustoolused forfault detection,breakdownanticipationandproblemdiag- nosticsinthemachine.AneffectiveSHMsystemcanimprove

productivity and ensure work piecequality contributing toa majorinfluenceonmachiningeconomics.Theeffectivenessof theSHMsystemreliesonthesensitivityofsensorused,position ofsensormounting,numberofsensorstobeusedandtypesof sensorstobepracticed.Currenthealthmonitoringapproaches aremuchabletodeterminethe(Rytter,1993):

• presenceofdamageinstructure(level1);

• geometriclocationofdamage(level2);

• typeorextent/severityofdamage(level3);

• anticipationoftheremaininglifetime(level4);

• selfhealingstructures(level5).

Park,Muntges,andInman(2001)suggests‘Level5’inthe contextof smart structureswithemergenceof shapememory alloys i.e. Cu-Al-Ni and Ni-Ti (NiTi) alloys. The develop- menttowardahigherfunctionallayeristakenasanimportant strideforwardintheevolutionofautonomousmonitoringofthe integrityofmachinetools.

The accelerometer is the most widely used contact type instrumentsthatusethepiezoelectriceffecttomeasuremechan- icalinputsuch asforce, torque,strain,pressure,acceleration, acousticemissionandvibrationbyconvertingthemintoelec- trical charge. Both accelerometers and velocity transducers measure the vibration of the non rotating part of a machine, buttheycannotprovideanydirectinformationabouttherela- tivemotionoftherotorandstator(Wei,Zhou,Huang,&Tan, 2013).

The present workdevelopsanon contact structural health monitoringsystemusingalowcostmicrocontrollerbaseddata acquisitionsystem(DAS).Thesystemusedthelight depend- entresistorsthatworkontheprincipleofphotoconductivityto measurethevibrationsignalofthemachine.Theproducedvibra- tionsignalisthenanalyzedusingNIbasedLabVIEWsoftware todeterminethehealthofthemachine.Anexperimentiscar- riedoutusingthevibrationrigforthemachinetocollectsignal datafortheanalysis.Byapplyingamicrocontroller,theproject proposestodevelopacost-effectivedataacquisitionsystemfor monitoringthemachinecondition.Theobjectivesofthiswork includethefollowing:

a) Todevelopaneconomicalnoncontactvibrationbasedinstru- menttomonitorthecurrentstatusofthemachineforreliable structuralhealthmonitoringusingmicrocontrollerinterfaced withLabVIEW.

b) Toacquire,visualizeandanalyzethevibrationsignaturesof amachinetool.

c) Tocorrelatethevibrationsignatureswithmachinehealth.

d) Toreduceregularperiodicinspections,betterqualitycontrol andscheduledmaintenance.

e) RemoteMonitoring.

2. Structureoftheproposeddataacquisitionsystem Thefunctionalinformationflowoftheproposeddataacquisi- tionsystemisshowninFigure1.Duringmachiningoperation, real-time vibration data intwo directions has been collected

(3)

Laser beam

Vibrating object

LDR transducer Reflected

light

Vibration data

Vibration

parameters Microcontroller

Signal visualization

Fig.1.Informationflowfornoncontactvibrationmeasurement.

throughalightdependantresistor.Theacquireddataisthencal- ibratedwiththecontacttypesignaltransducer(accelerometer).

3. Structuralhealthmonitoringsystemthroughthe microcontroller-baseddataacquisitionsystem

Machineconditionmonitoringwasimplementedthroughthe analysisof vibrationdata that wascaptured through thedata acquisition system. The following four sections describe the developmentofthehardwaresystem,softwaresystem,integra- tionandtestingofthedataacquisitionsystem,andcalibration alongwiththevibrationdataanalyses.

3.1. Hardwaresystemdevelopment

The non contact data acquisition system employed off- the-shelf equipment in a novel integrated approach. During machining operation, real-time vibration data was collected throughalightdependantresistor.Therawsignalwasprocessed toextractthecharacteristicfeaturesfromthemachiningsignals.

Thedynamicstreamofthecharacteristicfeaturescanthenbe visualizedinrealtime.Thevariousequipmentthatisusedfor developingasystemisdescribedbelow:

3.1.1. Laseremitter

Alaserdiodeproducesalaserbeamandanarrayofphoto- diode sensors receives the reflected beam, whichis reflected from the reflectorplaced on the vibratingobject for acquisi- tionofvibrationdata.Theanglebetweentheincidentrayand reflectedrayhasbeenfoundtobe7.5^◦asthesensorarrangement waspositionedinfrontofmirrortopreventanylossofreflected lightoverthearrayoflightdependantresistors.Itisincorporated withpowersupplystandandlaserbeampositioningmechanism toreachtheflexiblechangesinitscoherence.

3.1.2. Vibrationrig

Anelectricmotorisfittedinasteelframe.Anironframehas beenemployedwithtwoassemblies.Firstly,therotorassembly hasa pulley anda rotating shaft withan outer diameter, the pulleybeingmountedforrotationontherotatingshaft(Fig.2).

Secondly,thestatorassemblyhasabearingsleevewithabottom opening,therotatingshaftbeinginsertedintothebearingsleeve forrotation. A20mm diametershaft issupportedbyanother bearingplacedattheotherend.Abeltisalsoaddedtorotate

AC motor

Supporting frame

V belt

Accelerometer

V-groove pulley Shaft Bearing 1

Bearing 2

Fig.2.Topviewofvibrationrig.

Connecting pins Connecting wire Pin 4 Pin 1

Pin 2

PCB

Resistor Pin 0

Pin 5

Pin 3

Fig.3.Schematicdiagramofthenon-contactvibrationmeasurementcircuit.

theshaftbyapplyinganelectricpowertothemotor(rotatesat 1100rpm).

3.1.3. Machinevibrationmeasurementsensor

Sixlightdependantresistors(LDR)sensorsembeddedina printed circuitedboardare habituatedtoreceivethe vibration signalbystudyingthechangeintheintensityofthelaserbeam.

Theprintedcircuitboardincludesasensingelementandaresis- torinterfaceabletocarryinformationfromthesensingelement andsendananalogsignaltothemicrocontroller.Figure3shows two-dimensionalcross-shapedfivesensors,i.e.,PIN0,PIN1, PIN2,PIN4,andPIN5areplacedtogethertoreceivethevari- ationinintensityoflaserlightduetovibrations,whilstthesixth sensor, i.e., PIN 3 is placed at some distance to receive the

(4)

Light falls on reflector

Light

received LDR sensor

Microcontroller Interfacing

Vibration parameters

Calibration with accelerometer

Data analysis

Results &

discussions Software developement

system

Connected to Laser

source

Vibrating rig (containing

reflector)

Fig.4.Experimentalsetupfornoncontactvibrationmeasurementsystem.

surrounding light. The PIN 3 is positioned separately as to removethenoise(intheformoflight)interferencesthatcome fromthesurroundingsduringoperation.PIN1actsasarefer- encepin,changeintheintensityoflightbetweenPIN0andPIN 2areusedtomeasuretheX-axisvibrations,whilebetweenPIN 4andPIN5areusedtomeasureY-axisvibrations.

3.1.4. Microcontroller-baseddataacquisitiondevice

Amicrocontroller(Arduino)dataacquisitiondevicehasbeen used as a central data acquisition hub. This device helps to interface vibration signals with the computer program using LabVIEWinstalledonaPCviaaUSBport.Usingastandard highspeedUSBport,thisdeviceenableslowcost,PC based analoganddigital input/outputdataacquisition. Amicrocon- trollerisassociated withthe sensor,andbrings forthadigital outputafterburningtheprogramforthephotoresistor.

3.2. Softwaresystemdevelopment

Twocomputerprogramshavebeenwrittenseparatelyindif- ferentsoftwaredevelopmentalenvironmentswithoutusingany anti-alisaingfilter.Thefirstispreparedformicrocontrollerand theotherisbasedonVI(VirtualInstrumentation)forLDRsen- sorinLabVIEWtocaptureanddisplaythevibrationsignalsin themannerofrealtime.Bothofthemaresuccessfulininterfac- ingwiththehardwaresystemforreal-timedataacquisition.An amplificationfactorof1000isusedtoconvertthemVintovolts foranaccelerometer.

3.3. Integratingandtestingofthedataacquisitionsystem TheschematicoftheexperimentalsetupisshowninFigure4.

Duringtheoperation,thesensorarrangementwaspositionedin frontofamirrortopreventanylossofreflectedlightoverthe arrayoflightdependantresistors.

Theinstallationofthesensorandtheimplementationofthe data acquisition system for vibration measurements are dis- playedinFigure5.

Laser source

Microcontroller LDR Circuit

Vibration rig

Accelerometer LabVIEW program

NI cDAQ

Fig.5.Schematicdiagramofcomponentintegration.

Threesetsof experimentswereconductedonthedesigned vibration rig for different conditions of bearings. One set included two healthybearingsonbothends ofthe shaft.The othersetincludedtwofaultybearings,andthethirdoneincluded healthy andfaulty bearingsonboth sidesrespectively. Three experiments were performed for each setto collectvibration datageneratedintheoperation.Forthefirsttwoexperiments, 30k sampleswere collected,andfor the nextexperiment 25k samplesweretakentovisualizeandanalyzethevibrationdata.

3.4. Calibration

Acontacttypeinstrument,i.e.,ICP@triaxialaccelerome- terembeddedinavibrationrighasbeenusedtocalibrate the resultsacquiredbythedevelopedsystemusingNIcDAQ-9178.

Two dimensional data have been extracted from the tri-axial accelerometerbyremovingtheZ-axisvibrationdata.

4. Resultsanddiscussion

Basedontheanalysisandthetheoreticalframeworkprovided inthe earliersections,theexperimenthasbeen performedon thevibrationrig,andresultsobtainedfromtheanalysisphase in thedevelopment of non contactSHM systemfor machine tools are discussed inthis section.A numberof experiments havebeenconductedtochecktheaccuracyinthemeasurement ofthedevelopedsystembycalibratingitwithanaccelerometer.

Thepointsusedfortheanalysispartaremarkedtodistinguishthe variousexperimentsasshowninTable1.Usingthisdata,various graphsbetweenvibrationamplitudeandnumberofsampleshave beenplottedandanalyzedfordifferentconditionofbearingsin chronologicalorder:

(5)

Table1

Analysisviewfordifferentconditionofbearings.

Points Numberofsamples Statusofthemachine Time(ins)

BeforeA 5k Non-operating 1.6

AtoB 30k Operating 10

BtoC 5k Nonoperating 1.6

CtoD 30k Operating 10

DtoE 5k Nonoperating 1.6

EtoF 25k Operating 8.5

Axisnotation

Y X

Samplingfrequency 3k

2.0

1.5

0.5

0.0

–0.5 1.0

–1.5

0 200 400

Number of samples (x 10²)

X axis LDR (volts)

600 800 1000

–1.0

–2.0

X axis LDR

D F

C E

A B

Fig.6.X-axisvibrationsignatureforLDRvs.numberofsamples[0–100k].

0.5

0.0

–0.5 1.0

0 200 400

Y axis LDR (volts)

600 800 1000

–1.0

Y axis LDR

F

C D E

A B

Fig.7.Y-axisvibrationsignatureforLDRvs.numberofsamples[0–100k].

4.1. Experimentonhealthybearings

The first experiment was conducted on the healthy bearings tomeasure the vibrations using the photosensor andan accelerometer.

Figures6and7showthecombinedviewofallexperiments conductedtoanalyzethe variationof thevibration amplitude withanumberofsamplesintheX-axisandY-axisforthepho- tosensor.

2.0

1.5

1.0

0.5

0.0

–0.5

–1.0

–1.5

–2.0

0 200 400 600 800 1000

X axis accelerometer (volts) A B C D E F

Number of samples (x 10²) X axis accelerometer

Fig.8.X-axisvibrationsignaturefortheaccelerometervs.numberofsamples [0–100k].

1.0

0.5

0.0

–0.5

–1.0

0 200 400 600 800 1000

Y axis accelerometer (volts)

A B C D E F

Number of samples (x 10²) Y axis accelerometer

Fig.9.Y-axisvibrationsignaturefortheaccelerometervs.numberofsamples [0–100k].

Figures8and9showthecombinedviewofallexperiments conducted toanalyzethevariationof the vibrationamplitude with a number of samples in the X-axis and Y-axis for an accelerometer.

Figure 10 shows the comparison of variation in vibration amplitudemeasuredbybothcontactandnoncontacttypeinstru- mentsintheX-axistomonitorthehealthofthemachine.Ithas beenobservedthatthedevelopedsystemgivesalmostthesame valueslikethecontacttypeinstrument.

• For0–35krangeof samples,thevalueofbothtransducers, i.e., atnon operatingtime(0–50),approaches tozero. The maximumvalue forLDR andtheaccelerometerwere 1.13 and 1.32V,whereas the minimum values were −1.25and

−1.29V.

• For35–70krangeofsamples,themaximumandminimum valuesforLDRwereobservedtobe1.294and−1.19V.The maximumandminimumvaluesfor theaccelerometerwere observedtobe1.29and−1.25V.

• For 70–100k range of samples, the maximum values for the photosensorandthe accelerometerare0.85and0.91V,

(6)

2.0

1.5

1.0

0.5

0.0

–0.5

–1.0

–1.5

–2.0

0 200 400 600 800 1000

2.0

1.5

1.0

0.5

0.0

–0.5

–1.0

–1.5

–2.0

X axis LDR (volts)

X axis accelerometer (volts) A B C D E F

X axis LDR X axis accelerometer

Fig. 10.Comparison of X-axis vibration signature for the LDR and the accelerometervs.numberofsamples[0–100k].

1.0

0.5

0.0

–0.5

–1.0

0 200 400 600 800 1000

–1.0 –0.5 0.0 0.5 1.0

Y axis accelerometer (volts) A B C D E F

Number of samples (x 10²) Y axis LDR

Y axis accelerometer

Y axis LDR (volts)

Fig. 11.Comparison of Y-axis vibration signature for the LDR and the accelerometervs.numberofsamples[0–100k].

whereas the minimum values for both transducers were

−0.981and−0.983V,respectively.

Figure11 showsthe comparisonof variationinthe vibra- tionamplitudemeasuredbybothcontactandnoncontacttype instrumentsintheY-axistomonitorthehealthofthemachine.

• For0–35krangeofsamples,themaximumvalueforLDRand theaccelerometerare0.496and0.49V,whereastheminimum valuesare−0.38and−0.32V.

• For35–70krangeofsamples,themaximumandminimum values for LDR were observedto be0.496 and−0.497V.

The maximum andminimum values for the accelerometer wereobservedtobe0.45and−0.498V.

• For70–100krangeofsamples,themaximumvaluesforthe photosensor and the accelerometer were 0.33 and 0.34V, whereas the minimum values for both transducers were

−0.330and−0.32V.

3.0 2.5 2.0 1.5 1.0 0.5 0.0 –0.5 –1.0 –1.5 –2.0 –2.5 –3.0

0 200 400 600 800 1000

X axis LDR (volts)

A B C D E F

X axis LDR

Fig.12.X-axisvibrationsignatureforLDRvs.numberofsamples[0–100k].

3.0 2.5 2.0 1.5 1.0 0.5 0.0 –0.5 –1.0 –1.5 –2.0 –2.5 –3.0

0 200 400 600 800 1000

Y axis LDR (volts)

A B C D E F

Y axis LDR

Fig.13.Y-axisvibrationsignatureforLDRvs.numberofsamples[0–100k].

4.2. Experimentonfaultybearings

Figures12and13showthecombinedviewofallexperiments conducted toanalyzethe variationofthe vibrationamplitude withanumberofsamplesintheX-axisandY-axisforthepho- tosensor.Itwasfoundthatthemagnitudeofvibrationamplitude increasesincomparisontothehealthybearingforbothaxis.

Figures14and15displaythecombinedviewof allexper- iments conducted to analyze the variation of the vibration amplitudewithanumberof samplesintheX-axisandY-axis fortheaccelerometer.Here,itwasagainobservedthatthemag- nitudeofvibrationintheY-axisislessascomparedtoX-axis.

Figure16depictsthecomparisonofvariationinthevibration amplitudemeasuredbybothcontactandnoncontacttypeinstru- mentsintheX-axistodeterminetheconditionofthemachine.

• For0–35k range of samples, the maximumand minimum valuesforLDRwereobservedtobe2.15 and−1.6V.The maximumandminimumvaluesfortheaccelerometerwere observedtobe2.12and−1.58V.

(7)

3.0 2.5 2.0 1.5 1.0 0.5 0.0 –0.5 –1.0 –1.5 –2.0 –2.5 –3.0

0 200 400 600 800 1000

X axis accelerometer (volts)

A B C D E F

X axis accelerometer

3.0 2.5 2.0 1.5 1.0 0.5 0.0 –0.5 –1.0 –1.5 –2.0 –2.5 –3.0

0 200 400 600 800 1000

A B C D E F

0 200 400 600 800 1000

2.0 2.5 3.0

1.5 1.0 0.5 0.0 –0.5 –1.0 –1.5 –2.0 –2.5 –3.0 2.0

2.5 3.0

1.5 1.0 0.5 0.0 –0.5 –1.0 –1.5 –2.0 –2.5 –3.0

X axis LDR (volts)

A B C D E F

X axis LDR X axis accelerometer

0 200 400 600 800 1000

2.0 2.5 3.0

1.5 1.0 0.5 0.0

–1.5 –1.0 –0.5

–3.0 –2.5 –2.0 2.0

2.5 3.0

1.5 1.0 0.5 0.0

–1.5 –1.0 –0.5

–3.0 –2.5 –2.0

Y axis LDR (volts)

A B C D E F

Y axis LDR Y axis accelerometer

Fig. 17.Comparison of Y-axis vibration signature for the LDR and the accelerometervs.numberofsamples[0–100k].

• For 35–70krange of samples, the maximum valuefor the LDRandtheaccelerometerwere 2.41and2.37V,whereas theminimumvalueswere−1.99and−1.98V.

• For70–100krangeofsamples,themaximumvaluesforthe photosensor and the accelerometer were 2.38 and 2.37V, whereastheminimumvaluesforbothtransducerswere−1.84 and−1.82V.

Figure 17showsthe comparisonof variationinthe vibra- tionamplitudemeasuredbybothcontactandnoncontacttype instrumentsintheY-axisformonitoringthehealthofmachine tools.

• For0–35krangeofsamples,themaximumvaluesforthepho- tosensorandtheaccelerometerare2.018and2.09V,whereas the minimumvalues for both transducers were −1.87and

−1.85V.

• For 35–70krange of samples, the maximum valuefor the LDRandtheaccelerometerwere2.301and2.310V,whereas theminimumvalueswere−1.79and−1.80V.

• For70–100krangeofsamples,themaximumandminimum valuesforthe LDRwereobservedtobe2.03and−1.91V.

The maximum andminimum values for the accelerometer wereobservedtobe2.08and−1.92V.

4.3. Experimentonhealthyandfaultybearings

Thecombinedviewofallexperimentsconductedtoanalyze the variation of vibration amplitude with number of samples in X-axis and Y-axis for the photosensor is shown in Figures18and19.Itisfoundthat thiscasehasfollowedthe characteristicsofbothpreviouscases.

Figures20and21displaythecombinedviewofallexperi- mentsconductedtoanalyzethevariationofvibrationamplitude with a number of samples in the X-axis and Y-axis for the accelerometer.Themagnitudeofvibrationsmeasuredbyboth transducersintheY-axishaslessvalueascomparedtoX-axis data.

(8)

2.5 2.0 1.5 1.0 0.5 0.0 –0.5 –1.0 –1.5 –2.0 –2.5

0 200 400 600 800 1000

X axis LDR (volts)

A B C D E F

X axis LDR

Fig.18.X-axisvibrationsignaturefortheLDRvs.numberofsamples[0–100k].

2.5 2.0 1.5 1.0 0.5 0.0 –0.5 –1.0 –1.5 –2.0 –2.5

0 200 400 600 800 1000

Y axis LDR (volts)

A B C D E F

Y axis LDR

Fig.19.Y-axisvibrationsignaturefortheLDRvs.numberofsamples[0–100k].

2.5 2.0 1.5 1.0 0.5 0.0 –0.5 –1.0 –1.5 –2.0 –2.5

0 200 400 600 800 1000

A B C D E F

X axis accelerometer

Figure22 showsthe comparisonof variationinthe vibra- tionamplitudemeasuredbybothcontactandnoncontacttype instrumentsintheX-axis tocorrelatethe vibrationsignatures withmachinehealth.

2.5 2.0 1.5 1.0 0.5 0.0 –0.5 –1.0 –1.5 –2.0 –2.5

0 200 400 600 800 1000

A B C D E F

0 200 400 600 800 1000

2.0 2.5

1.5 1.0 0.5 0.0 –0.5 –1.0 –1.5 –2.0 –2.5 2.0

2.5

1.5 1.0 0.5 0.0 –0.5 –1.0 –1.5 –2.0 –2.5

X axis LDR (volts)

A B C D E F

X axis LDR X axis Accelerometer

• For0–35k range of samples, the maximumand minimum valuesforLDRwereobservedtobe1.86and−1.17V.The maximumandminimumvaluesfortheaccelerometerwere observedtobe1.83and−1.36V.

• For35–70krangeofsamples,themaximumvaluefortheLDR and the accelerometer were 1.70 and 1.74V, respectively, whereastheminimumvalueswere−1.91and−1.64V.

• For70–100krangeofsamples,themaximumvaluesforthe photosensor and the accelerometer were 1.86 and 1.99V, whereastheminimumvaluesforbothtransducerswere−1.86 and−1.77V.

Figure 23 shows the comparison of variation in vibration amplitudemeasuredbybothcontactandnoncontacttypeinstru- mentsintheY-axistovalidatethedevelopedsystemforhealth monitoring.

• For0–35k range of samples, the maximumand minimum valuesfor theLDRwereobservedtobe1.88and−1.59V.

(9)

0 200 400 600 800 1000 2.0 2.5

1.5 1.0 0.5 0.0 –0.5 –1.0 –1.5 –2.0 –2.5 2.0

2.5

1.5 1.0 0.5 0.0 –0.5 –1.0 –1.5 –2.0 –2.5

Y axis LDR (volts)

Y axis accelerometer (volts) A B C D E F

Y axis LDR Y axis accelerometer

Fig. 23.Comparison of Y-axis Vibration signature for the LDR and the accelerometervs.numberofsamples[0–100k].

2500 3000 3500 4000

2000

1500

1000

500

0

0 500 1000 1500 2000 2500 3000 3500 4000

Power spectrum (g2/Hz)

Frequency (Hz)

ACC LDR

Fig.24.X-axispowerspectrumgraphfortheaccelerometerandtheLDR[A–B].

The maximumand minimum valuesfor the accelerometer wereobservedtobe1.84and−1.56V.

• For35–70krangeofsamples,the maximumvaluesfor the photosensor and the accelerometer were 1.52 and 1.77V, whereastheminimumvaluesforbothtransducerswere−1.55 and−1.01V.

• For70–100k rangeofsamples,themaximumvalueforthe LDRandtheaccelerometerwere1.837and1.74V,whereas theminimumvalueswere−1.46and−1.18V.

5. Signalanalysis

Mostof thepastresearchonthetopichasbeenperformed onthemeasurementandanalysisof the machiningsignalsin thetimedomainonly.Recentstudieshavefoundthefrequency domainanalysistobemoreusefulinconditionmonitoring.For thiswork,theFFTandpowerspectrumwereappliedtoextract thefeaturesfromthevibrationdataasshowninFigures24–41.

FFTwasappliedforconvertingthesignalsfromtimedomain tofrequencydomain,whereasthepowerspectrumwasusedfor spectral measurements. The parameter, powerspectrum peak

2500 3000 3500 4000

2000

1500

1000

500

0

0 500 1000 1500 2000 2500 3000 3500 4000

Frequency (Hz)

ACC LDR

Fig.25.X-axispowerspectrumgraphfortheaccelerometerandtheLDR[C–D].

2500 3000 3500 4000

2000

1500

1000

500

0

0 500 1000 1500 2000 2500 3000 3500 4000

Frequency (Hz)

ACC LDR

Fig.26.X-axispowerspectrumgraphfortheaccelerometerandtheLDR[E–F].

2500

2000

1500

1000

500

0

0 500 1000 1500 2000 2500 3000 3500 4000

Frequency (Hz)

ACC LDR

Fig.27.Y-axispowerspectrumgraphfortheaccelerometerandtheLDR[A–B].

wasextractedfromreal-timesignaltochecktheaccuracyinthe measurementofbothtransducers.

Thepowerspectrumpeakvaluesatdifferentfrequenciesin boththeX-axisandtheY-axiswereobserved.Itwasobserved thatthepeakamplitudeofpowerspectruminallthreecases,i.e.,

(10)

2500

2000

1500

1000

500

0

0 500 1000 1500 2000 2500 3000 3500 4000

Frequency (Hz) Power spectrum (g2/Hz)

ACC LDR

Fig.28.Y-axispowerspectrumgraphfortheaccelerometerandtheLDR[C–D].

Table2

PowerspectrumpeakamplitudefortheLDRandtheaccelerometerinbothaxis.

Points Axis Powerspectrumpeak amplitude

(accelerometer)(g²/Hz)

Powerspectrum peakamplitude (LDR)(g²/Hz)

AtoB X 2095.11 2418.69

CtoD X 2190.29 2270.37

EtoF X 2432.55 2380.92

AtoB Y 1442.63 1308.57

CtoD Y 1680.81 1791.15

EtoF Y 2102.19 1965.74

healthybearings,faultybearings,healthyandfaultybearingsfor boththeLDRandtheaccelerometerliebetweenthefrequency rangeof500–1100Hz.

5.1. Healthybearings

On performing the experimentsusing healthybearings on bothendoftheshaft,theobtainedpowerspectrumpeakvalues atdifferent frequenciesinboththe X-axisandthe Y-axis has beentabulatedinTable2.

ThepowerspectrumgraphsinboththeX-axisandtheY-axis fortheLDRandtheaccelerometerisshowninFigures24–29.

5.2. Faultybearings

Theobtained power spectrumpeakvaluesat differentfre- quenciesinboththeX-axisandtheY-axiswhileperformingthe experimentsusingfaultybearingsonbothendsoftheshafthas beentabulatedinTable3.

The powerspectrum graphsin bothX-axis andY-axis for LDRandaccelerometerhasbeenshowninFigures30–35.

5.3. Healthyandfaultybearings

Theobtained power spectrumpeakvaluesat differentfre- quenciesinboththeX-axisandtheY-axiswhileperformingthe

Table3

PowerspectrumpeakamplitudefortheaccelerometerandtheLDRinbothaxis.

AtoB X 3957.01 3891.37

CtoD X 4102.29 3978.31

EtoF X 4628.22 4502.14

AtoB Y 3399.01 3472.92

CtoD Y 3461.13 3503.21

EtoF Y 3718.08 3605.71

2500

2000

1500

1000

500

0

0 500 1000 1500 2000 2500 3000 3500 4000

ACC LDR

Fig.29.Y-axispowerspectrumgraphfortheaccelerometerandtheLDR[E–F].

0 500 1000 1500 2000 2500 3000 3500 4000

ACC LDR 5000

4500 4000 3500 3000 2500 2000 1500 1000 500 0

Table4

PowerspectrumpeakamplitudefortheaccelerometerandtheLDRinbothaxis.

AtoB X 3242.75 3196.09

CtoD X 3399.73 3472.41

EtoF X 3741.20 3623.80

AtoB Y 2629.24 2451.31

CtoD Y 2904.51 3056.29

EtoF Y 3242.71 3196.19

(11)

0 500 1000 1500 2000 2500 3000 3500 4000 Frequency (Hz)

ACC LDR 5000

4500 4000 3500 3000 2500 2000 1500 1000 500 0

0 500 1000 1500 2000 2500 3000 3500 4000

ACC LDR 5000

4500 4000 3500 3000 2500 2000 1500 1000 500 0

4000

3500

3000

2500

2000

1500

1000

500

0

0 500 1000 1500 2000 2500 3000 3500 4000

ACC LDR

4000

3500

3000

2500

2000

1500

1000

500

0

0 500 1000 1500 2000 2500 3000 3500 4000

ACC LDR

4000

3500

3000

2500

2000

1500

1000

500

0

0 500 1000 1500 2000 2500 3000 3500 4000

ACC LDR

4000

3500

3000

2500

2000

1500

1000

500

0

0 500 1000 1500 2000 2500 3000 3500 4000

ACC LDR

(12)

4000

3500

3000

2500

2000

1500

1000

500

0

0 500 1000 1500 2000 2500 3000 3500 4000

ACC LDR

4000

3500

3000

2500

2000

1500

1000

500

0

0 500 1000 1500 2000 2500 3000 3500 4000

ACC LDR

4000

3500

3000

2500

2000

1500

1000

500

0

0 500 1000 1500 2000 2500 3000 3500 4000

ACC LDR

4000

3500

3000

2500

2000

1500

1000

500

0

0 500 1000 1500 2000 2500 3000 3500 4000

ACC LDR

4000

3500

3000

2500

2000

1500

1000

500

0

0 500 1000 1500 2000 2500 3000 3500 4000

ACC LDR

experimentsusinghealthyandfaultybearingsonbothendsof theshafthasbeentabulatedinTable4.

ThepowerspectrumgraphsinboththeX-axisandtheY-axis fortheLDRanditisdisplayedinFigures30–35.

6. Conclusions

With evolution in recent developments in sensors, signal processingandcomputingmethods,structuralhealthmonitoring hasbecomeanimportantfieldofresearch.Thepresentworkhas beenaimedtodevelopanoncontactstructuralhealthmonitor- ingsystemformachinetoolstomonitorthehealthofmachine for decisionmaking about maintenancestrategies. Following conclusionshavebeenobtainedfromtheexperimentationand analysisofresults:-

1) The developedsystemcorrelatesthe vibrationsignatures withmachinehealth.

2) Realtimeconditionmonitoringcanbecarriedoutbyusing thedevelopedsystem.