Ener gy Resear ch and Devel opment Di vi si on

FI NAL PROJECT REPORT

ASSESSMENTOFPIEZOELECTRIC
MATERIALSFORROADWAYENERGY
HARVESTING

CostofEnergyandDemonstration
Roadmap

Preparedfor: CaliforniaEnergyCommission
Preparedby: DNVKEMAEnergy&Sustainability
J ANUARY 2014
CEC 500 2013 007

PREPARED BY:

Primary Author(s):
Davion Hill, Ph.D., DNV
Arun Agarwal, Ph.D., DNV
Nellie Tong, KEMA Inc., Project Manager

DNV KEMA Energy & Sustainability
155 Grand Avenue, Suite 500
Oakland, California 94612
www.dnvkema.com

Contract Number: 500-11-029


Prepared for:

California Energy Commission

J esselyn Rosales
Contract Manager

David Chambers
Project Manager

Fernando Pina
Office Manager
Energy Efficiency Research Office

Laurie ten Hope
Deputy Director
ENERGY RESEARCH AND DEVELOPMENT DIVISION

Robert P. Oglesby
Executive Director

DISCLAIMER

This report was prepared as the result of work sponsored by the California Energy Commission. It
does not necessarily represent the views of the Energy Commission, its employees or the State of
California. The Energy Commission, the State of California, its employees, contractors and
subcontractors make no warranty, express or implied, and assume no legal liability for the
information in this report; nor does any party represent that the uses of this information will not
infringe upon privately owned rights. This report has not been approved or disapproved by the
California Energy Commission nor has the California Energy Commission passed upon the
accuracy or adequacy of the information in this report.

ACKNOWLEDGEMENTS
DNVKEMAwouldliketoacknowledgethefollowingpeopleandorganizationsforaidingin
theresearchforthisassessment:
MikeGatto,MemberoftheCAStateAssembly,43
rd
District,forhisleadershipandsupport
MikeGravely,CaliforniaEnergyCommission
DavidChambers,CaliforniaEnergyCommission
KaterinaRobinson,CaliforniaAssemblymanGattosOffice
RandyCopperman,ChannelTechnologiesGroup
EnderKuntsal,ChannelTechnologiesGroup
EricWeaver,FederalHighwayAdministrationOfficeofInfrastructureResearchand
Development
JonLazarus,OregonDepartmentofTransportation
ElizabethRedmond,POWERleapInc.
CharlesMilgrom,Innowattech
CraigNear,Genziko

i
PREFACE
TheCaliforniaEnergyCommissionEnergyResearchandDevelopmentDivisionsupports
publicinterestenergyresearchanddevelopmentthatwillhelpimprovethequalityoflifein
Californiabybringingenvironmentallysafe,affordable,andreliableenergyservicesand
productstothemarketplace.
TheEnergyResearchandDevelopmentDivisionconductspublicinterestresearch,
development,anddemonstration(RD&D)projectstobenefitCalifornia.
TheEnergyResearchandDevelopmentDivisionstrivestoconductthemostpromisingpublic
interestenergyresearchbypartneringwithRD&Dentities,includingindividuals,businesses,
utilities,andpublicorprivateresearchinstitutions.
EnergyResearchandDevelopmentDivisionfundingeffortsarefocusedonthefollowing
RD&Dprogramareas:
BuildingsEndUseEnergyEfficiency
EnergyInnovationsSmallGrants
EnergyRelatedEnvironmentalResearch
EnergySystemsIntegration
EnvironmentallyPreferredAdvancedGeneration
Industrial/Agricultural/WaterEndUseEnergyEfficiency
RenewableEnergyTechnologies
Transportation

AssessmentofPiezoelectricMaterialsforRoadwayEnergyHarvestingisthefinalreportforthePiezo
BasedEnergyHarvestingTechnologyproject(ContractNumber50011029,Work
AuthorizationNumber1)conductedbyDNVKEMAEnergyandSustainability.The
informationfromthisprojectcontributestoEnergyResearchandDevelopmentDivisions
EnergySystemsIntegrationProgram.

FormoreinformationabouttheEnergyResearchandDevelopmentDivision,pleasevisitthe
EnergyCommissionswebsiteatwww.energy.ca.gov/research/orcontacttheEnergy
Commissionat9163271551.

ii
ABSTRACT
TheCaliforniaEnergyCommissionseekstobetterunderstandthecurrentstatusof
piezoelectricbasedenergyharvestingtechnologyinroadwayandrailwayapplications.This
projectassessedthevalueofpiezoelectricbasedenergyharvestingtechnologytodetermineif
theearlyresultsfromprototypedemonstrationswarrantamoredetaileddemonstrationin
California.Piezoelectriccrystalsgiveanelectricaldischargewhenmechanicallystressed.If
piezoelectricbasedtechnologyhasthepotentialtomatchtheperformance,reliability,andcosts
ofexistingoremergingrenewableenergysources,thenitcanpotentiallydiversifyCalifornias
resourceportfolio,andultimatelyincreasegridreliabilityandreducecoststoratepayers.This
reportassessesthecommercialstatusandfuturepotentialofpiezoelectricbasedenergy
harvestingtechnologyinroadwayandrailwayapplications.Thereportreviewsdatafrom
recentdemonstrationsandcalculatesthelevelizedcostofenergy.Althoughthecostofenergyis
computedusingvendorsuppliedinformation,simplifiedtrafficmodelsareusedtocrosscheck
vendorclaimsanddeterminewhichfactorscontributetouncertaintyaboutthecostofenergy.
Usingvendorsupplieddata,theanalysisshowsthattheestimatedrangeoflevelizedcostof
energyforpiezoelectricsystemisbetween$0.08$0.20/kWh,dependingontrafficconditions
andvehiclecharacteristics.Simplifiedtrafficmodelsandknowndemonstrationdatarevealthat
somevendorclaimsaremutuallyexclusive.Thislikelyindicatesthepresentationofbestvalues
frommutuallyexclusiveconditions.Untilthepoweroutputpermoduleistransparently
quantified,costofenergyestimateswillcontaininherentuncertainty.Withtheinformation
currentlyavailable,itappearsthatpowerdensitiesof300W/ft
2
ormoreareneededtoapproach
theeconomicviabilityclaimedbyvendors.Theresultsofthisresearchindicateademonstration
andfurtherevaluationofthetechnologyshouldattempttoquantifythepoweroutput,
durability,andlifetimeofthesysteminadditiontoitsperformanceasafunctionoftraffic
volumeiswarranted.

Keywords:Piezoelectric,energyharvesting,wastevibration,highway

Pleaseusethefollowingcitationforthisreport:
Hill,Davion,NellieTong,(DNVKEMA).2013.AssessmentofPiezoelectricMaterialsfor
RoadwayEnergyHarvesting.CaliforniaEnergyCommission.PublicationNumber:
CEC5002013007.
iii
TABLE OF CONTENTS
Acknowledgements...................................................................................................................................i
PREFACE...................................................................................................................................................ii
ABSTRACT..............................................................................................................................................iii
TABLEOFCONTENTS.........................................................................................................................iv
LISTOFFIGURES..................................................................................................................................vi
LISTOFTABLES..................................................................................................................................viii
EXECUTIVESUMMARY........................................................................................................................1
Introduction............................................................................................................................................1
CHAPTER1:BasicsofPiezoelectricMaterials...................................................................................3
1.1 WhatArePiezoelectricMaterials?...........................................................................................3
1.2 BenefitsofEnergyHarvesting..................................................................................................3
1.3 WhatArePiezoelectricMaterialsCommonlyUsedfor?......................................................4
1.4 WhatAreSomeRelevantCostandEnergyMetrics?............................................................5
1.5 TwoWaystoHarvestEnergyWithPiezoelectricMaterials................................................9
1.5.1 EnergyDensityofaCompressionBasedSystem........................................................10
1.5.2 EffectofWheelbaseonCapacityFactoronkW/km....................................................13
1.5.3 IncreasedCapacityFactorThroughLongerPowerDuration...................................17
CHAPTER2:CostAnalysisofaPiezoelectricRoadwayEnergyHarvestingSystem................19
2.1 VendorClaimsandDemoData.............................................................................................19
2.2 RelationshipBetweenTrafficParametersandHarvesterCharacteristics.......................22
2.3 WhatIsthePowerOutputRequiredperHarvester?..........................................................24
2.4 CostofElectricityofaCompressionBasedPiezoelectricRoadwayEnergyHarvesting
SystemVendorClaims.....................................................................................................................27
2.4.1 ComparisonWithTrafficModel....................................................................................30
2.5 CostofElectricityFromVibrationBasedRoadwayEnergyHarvestingSystemBased
onVendorClaims................................................................................................................................31
2.5.1 ComparisonWithTrafficModel....................................................................................32
2.6 DNVKEMAEstimatesBasedonRestrictedAssumptions................................................34
iv
2.7 ComparisonoftheProjectedLCOEtoDistributedRenewableEnergyGeneration
Sources...................................................................................................................................................35
2.8 AddedValue:DataandReducedInspectionCosts............................................................39
2.9 RoadwaysversusRailways....................................................................................................39
CHAPTER3:ConclusionsandRecommendations...........................................................................42
3.1 StageGateEvaluation.............................................................................................................42
3.1.1 PhaseI:LabScaleTests...................................................................................................43
3.1.2 PhaseII:AcceleratedTests..............................................................................................44
3.1.3 PhaseIII:FieldTests........................................................................................................46
3.1.4 SuggestedTestStructure.................................................................................................47
3.1.5 ApproximateCostofaDemonstrationProject............................................................47
3.1.6 PotentialPartnersandFunctions...................................................................................48
GLOSSARY..............................................................................................................................................50
REFERENCES..........................................................................................................................................52
APPENDIXA:PiezoelectricEnergyHarvestingDemonstrations...............................................A1
Innowattech........................................................................................................................................A1
PiezoPower,PiezoelectricFloors.....................................................................................................A1
POWERleapInc..................................................................................................................................A2
KCFTechnologies..............................................................................................................................A3
OregonDepartmentofTransportation...........................................................................................A3
ChannelTechnologiesGroup...........................................................................................................A3
VirginiaTech.......................................................................................................................................A4
Genziko................................................................................................................................................A4
APPENDIXB:LiteratureEvaluation.................................................................................................B1
APPENDIXC:DataSummary............................................................................................................C1
APPENDIXD:EvaluationCriteria...................................................................................................D1
WhatWouldanEvaluationoftheTechnologyLookLike?.........................................................D1
ModulePowerOutput..................................................................................................................D1
DurationofHit(PowerPulseDuration).................................................................................D2
v
EnergyTransferandVehicleWeight..........................................................................................D3
DurabilityandLifetimeofthePiezoelectricCeramic...............................................................D3
DurabilityofSystemandComponents.......................................................................................D4
TrafficVolume................................................................................................................................D5
VehicleSpeed..................................................................................................................................D6
InstallationMethods......................................................................................................................D6
ValueofAdditionalDataandAvoidedInspectionCosts........................................................D7
EnergyStorageVersusNetMetering..........................................................................................D7
APPENDIXE:Railways.......................................................................................................................E1
APPENDIXF:CalculationDetails.....................................................................................................F1
CompressionBasedHarvesters,VendorClaims..........................................................................F1
Case1:MaximumFiveYearLifetime.........................................................................................F2
Case2:MaximumTenYearLifetime..........................................................................................F3
Case3:MaximumThirtyYearLifetime.....................................................................................F5
TrafficBasedLCOETechnologyAgnostic..................................................................................F6
APPENDIXG:ADDENDUM:ReviewerandStakeholderCommentsandResponsesFrom
DraftReport..........................................................................................................................................G1

LIST OF FIGURES
Figure1:EnergyHarvestingofVibrationsIsPossibleTypicallyintheLowerFrequencyRange
From11,000Hz.........................................................................................................................................5
Figure2:PowerandEnergyScaleforPiezoelectrics,Photovoltaics,andConventionalPower....6
Figure3:EnergyDensityDependsonFootTrafficforthePiezoelectricFloor.................................7
Figure4:PiezoelectricDevicesHaveVoltageandPowerCapabilitySimilartoLiIonBatteries..9
Figure5:DifferenceinPerformanceCharacteristicsofCompressionBasedEnergyHarvesters
andCantileverEnergyHarvesters........................................................................................................10
Figure6:ConfigurationofStacksintheVirginiaTechPiezoelectricHarvester.............................11
Figure7:PowerOutputFromaSingleInnowattechEnergyHarvesterDuringtheVirginiaTech
Demonstration..........................................................................................................................................12
vi
Figure8:Ten(red)Outof222PiezoelectricUnitsCanBeExcitedatAnyGivenTimeWitha
SinglePassingTractorTraileratLowSpeeds......................................................................................15
Figure9:EstimationofPowerOutputasaFunctionofWeightontheVehicleWheel(Virginia
TechBasis).................................................................................................................................................16
Figure10:TheMajorityofVMTComposedofPassengerCarsandLightDuty(Noncommercial)
Trucks........................................................................................................................................................17
Figure11:VehiclesWithShorterWheelbasesAreAlsoLighter,butHaveHigherEstimated
CapacityFactors.......................................................................................................................................17
Figure12:GenzikoClaimsSignificantPowerDensityOverCompetingTechnologies................18
Figure13:ActualOutputper1kmofRoadwayVersusRelevantOutputs....................................24
Figure14:RegressionCoefficientsfortheNPV(at60months)fortheTrafficModelLCOE
Calculation................................................................................................................................................27
Figure15:SensitiveFactorsAffectingtheLCOE.................................................................................28
Figure16:CapitalCostoftheSystemBasedonStatedPowerCapabilityandCapitalCostsFrom
theLiteratureReview..............................................................................................................................29
Figure17:ComparisonofCapitalCostsforVariousEnergySources..............................................36
Figure18:EstimatedLCOEforVariousEnergySystems..................................................................36
Figure19:ElectricityCostsNearSacramento,California..................................................................37
Figure20:CapitalCostsforEnergySystemsComparedtothePiezoelectricSystem....................38
Figure21:ComparisonoftheLCOEforthePiezoelectricSystemComparedtoOtherEnergy
Systems......................................................................................................................................................39
Figure22:CrossSectionalDiagramoftheRoadwayInstallationofPiezoelectricEnergy
Harvesters.................................................................................................................................................40
Figure23:AboveGroundInstallationofPiezoelectricEnergyHarvestingDevicesforRailways
....................................................................................................................................................................41
Figure24:SchematicofanAcceleratedAgeingPlatformforaPiezoelectricEnergyHarvester
Application...............................................................................................................................................45
Figure25:EfficientRoadTestingJigDesignedtoCycleConcreteUnderRoadWearConditions
....................................................................................................................................................................45
Figure26:InstallationofFieldDemonstrationinVirginiaUsingInnowattechEnergyHarvesters
(left)andSchematicofInstallationWithDataCollection(right)......................................................46

vii
viii
LIST OF TABLES
Table1:ComparisonofCalculationResultsAcrossMultipleThirdPartyInvestigatorsIntothe
CompressionBasedEnergyHarvestingTechnology.........................................................................13
Table2:TwoDifferentEnergyHarvestingTechnologiesComparedfora600Vehicle/hrFlow
Rate.............................................................................................................................................................20
Table3:DataSummaryforPiezoelectricMaterialsandInstallations..............................................20
Table4:DataSourcesFromLessObjectiveReferences......................................................................21
Table5:CorrelationBetweenTrafficandHarvesterMetricsandSystemPowerOutput.............23
Table6:EstimationofKeyPerformanceCharacteristicsofPiezoelectricSystemsBasedonTraffic
ParametersWitha1020YearLife........................................................................................................25
Table7:ExternalFactorsThatAffectSystemLevelPerformance.....................................................26
Table8:SummaryoftheLCOEAnalysisforThreeCases.................................................................28
Table9:CrossCheckofTrafficModelAgainstWhatWasDeterminedFromInnowattech
Numbers....................................................................................................................................................30
Table10:LCOEfortheGenzikoTechnologyBasedonVendorInformation.................................32
Table11:CrossCheckofTrafficModelAgainstWhatWasDeterminedFromInnowattech
Numbers....................................................................................................................................................33
Table12:ATestofMutuallyExclusiveLCOEandCapitalCostsfortheInnowattechTechnology
....................................................................................................................................................................34
Table13:AStagedGateApproachtoEvaluateVariousPiezoelectricTechnologies.....................43
Table14:OutlineofSuggestedDemonstrationProject......................................................................47

EXECUTIVE SUMMARY
Introduction
TheCaliforniaEnergyCommissioncontractedDNVKEMAtoinvestigatetheviabilityof
deployingpiezoelectricmaterialsinCaliforniaroadwaysforthepurposeofharvestingelectrical
energy.Piezoelectricmaterialsgenerateelectricitywhensubjectedtomechanicalstress.The
purposeofthiseffortwastoassessthevalueofpiezoelectricbasedenergyharvesting
technologytodetermineiftheearlyresultsfromprototypedemonstrationswarrantamore
detaileddemonstrationinCalifornia.Thisreportdescribesthepresentstateoftheartin
piezoelectricmaterials,assessesdemonstrationdatathatarepresentlyavailable,andconductsa
technoeconomicanalysisofthesedatatoassessthecostofenergyforpiezoelectricenergy
harvestersinroadways.Inthesamewaythatahybridvehiclerecaptureskineticenergywhilea
vehicleslows,piezoelectricdevicesembeddedinroadwaysmayrecaptureenergyintheformof
vibrationandcompressionunderthevehicletires.DNVKEMAisanindependentthirdparty
consultancyanddoesnotendorsethetechnologiesdescribedinthisreport.Thefollowing
summarizesthereportfindings:
DNVKEMAreviewedthreevendors.Thevendorbasedlevelizedcostofenergy(LCOE)
rangesfrom$0.03$0.18/kilowatthour(kWh)forcompressionbasedsystemswitha
meanof$0.11/kWh.Acompressionbasedsystemhasastackofpiezoelectricmaterials
thatiscompressedtogenerateenergywitheachcompressioncycle(suchasthepassage
ofavehicletireoverthestack).Vibrationbasedsystemsclaim$0.06$0.08/kWh.A
vibrationbasedsystemhasanarrayofcantileverorbentbeampiezoelectricenergy
harvestervibratingasaresultofexternalstimulus(suchasthepassageofavehicletire
overhead).DNVKEMAindependentlyestimatestheLCOEtobebetween$0.07
$0.20/kWhwith90percentofvaluesat$0.20/kWhorless,providedthatinstalled
nameplatepowerdensitiesgreaterthan300Watts/squarefoot(W/ft
2
)arepossible(a
criticalvalidationmetric).
AUCBerkeleyreviewofthetechnologywasmorecriticalthanotherreviews.The
Berkeleycritiquemakesacompellingargumenttoshowthatenergydensityisa
challengeforthistechnology,butitdidnotconsiderrecentengineeringinnovationon
mechanismstomaximizepowerdensity.Inthesamewaythatphotovoltaic(PV)
systemscanemploymirrorsandlensestoconcentratelightonactivemodules,the
piezoelectricmodulescanusemechanicaladvantageandnovelpackingofmaterialsto
thesameeffect.Forexample,ademonstrationatVirginiaTechhasmeasuredpower
outputgreaterthantheUCBerkeleyestimation.
Usingacomputationaltrafficmodel,itwasfoundthatsomeperformancemetrics
appeartobemutuallyexclusiveandrequirefurthervalidation.Poweroutputper
moduleisthepreferredmetrictovalidatethevendorclaims.
Importantfactorsthatdeterminethecosteffectivenessofthepiezoelectricroadway
energyharvestingsystemare(1)poweroutputperinstalledmodule,(2)lifetimeofthe
1
2
system,and(3)totalinstalledcost.Powerdensity(W/m
2
)andtrafficflowrate
(vehicles/hr)affecttotalpower.Systempoweroutputdependsonvehicleweightand
powerpulseduration.
Powerdensityisincreasedbymeansofhighdensitypackingofmaterialsand
mechanicaldesign.Piezoelectricenergyharvestingsystemsarebestplacedinareaswith
hightrafficflowratestomaximizeenergygeneration.Trafficflowconditionssuchas
vehicleweightandspacingbetweenvehiclesarelocationdependent(interstatehighway
versuscountyroad)andcriticaltosystemeconomicviability.
Thetechnologyisinearlystagesofproductdevelopment.Mostvendorsquotesystem
levelmetricssuchaskW/kmwhichcontaincontingencydatathataredifficultto
compareacrossregions.Publicdemonstrationstodatelackdataforcommercialdesigns.
Duetotheintermittencyofthepowergenerated,thereisaneedforenergystorageor
netmetering.Onlyoneofthethreevendorsevaluatedacknowledgesenergystorage
costs.
TheresultsofthisresearchindicateafuturedemonstrationinCaliforniaiswarrantedto
assessactualenergyprovidedinaroadwayorrailwaydemonstration.
Furtherconsiderationofrailwaysisneeded.Theinstalledcostofrailwayharvestersis
likelylessthanroadwayharvesters.Therailwayharvesterscanbeinstalledbetweenthe
railtieandthesteelrailanddoesnotrequirerailwayexcavation,whichlowersthelabor
requiredforinstallationandoverallefficiencyandcostofenergycanbeimprovedin
thisapplication.However,thereislittlepublicdataonrailwayinstallation.
Intheeventofanindependentevaluation,theassessmentshouldinclude:
o Independentverificationofpoweroutputperpiezoelectricmodule.
o Lifetimeanddurabilityasafunctionofunevenwearinthesystem,and
downtimeassociatedwithfailingmodules,maintenance,andreplacement.
o Costsassociatedwithinstallationmethods,downtimeassociatedwith
installation,andmaintenanceofthesystem,aswellasverificationofother
balanceofsystemcostssuchasenergystorageornetmetering.
o Addedvalueandpotentialmarketingrevenuefromadditionalinformation
collectedfromthepiezoelectricsystem,andpotentialrealtimestatusofthe
roadwayorrailwaycondition,thusavoidingunnecessaryinspectioncosts.

CHAPTER 1:
Basics of Piezoelectric Materials
1.1 What Are Piezoelectric Materials?
Simplystated,piezoelectricmaterialsarecrystalsthatgenerateelectricitywhencompressedor
vibrated.Theyhavetheuniqueoppositepropertyofgeneratingastresswhenvoltageisapplied
tothem.
Piezoelectricmaterialsfallwithinaclassofmultiplesolidstatematerialsthatcangenerate
electricitywiththeapplicationofsomestimulussuchasheat,stress,orlight.Photovoltaic
materialsgenerateelectricitywiththeapplicationoflight(thebasisofsolarpanels),and
thermoelectricmaterialsgenerateelectricitywiththeapplicationofheat.Piezoelectricmaterials
generateelectricitywiththeapplicationofstress.Thesematerialsareallsemiconductors,
meaningtheyaremuchlikeconventionalelectronics,generallyconstructedofSilicon(Si)or
Germanium(Ge)withadditionalelements.
Usingpiezoelectricstoharvestvibrationenergyfromhumanswalking,machineryvibrating,or
carsmovingonaroadwayisanareaofgreatinterest,becausethisvibrationenergyisotherwise
untapped.Sincemovementiseverywhere,theabilitytocapturethisenergycheaplywouldbea
significantadvancementtowardgreaterefficiencyandcleanerenergyproduction.
1.2 Benefits of Energy Harvesting
Inthesamewaythatahybridvehiclerecaptureskineticenergywhileavehicleslows,
piezoelectricdevicesembeddedinroadwaysmayrecaptureenergyintheformofvibrationand
compressionunderthevehicletires.
Thetopicofenergyharvestinggenerallyreferstothecaptureandstorageordirectuseof
ambientenergyforhumanpurposes.Assolarpanelsharvesttheenergycontainedinsunlight
andconvertittoelectricalenergy,otherformsofenergyharvestingalsocaptureambient
energy,usuallyintheformofvibrationorheat,andconvertittoausefulenergymediumsuch
asmechanicalorelectricalenergy.
Energyharvestingmayormaynotcapturerenewableenergy.Inthecaseofsunlight,theenergy
isrenewablebecauseitissourcedfromthesun,asourceofnearlyinfiniteenergyfortheplanet
andthesolarsystem.Wasteheatinanindustrialfacilitymaynotberenewablesincethe
processesgeneratingthewasteheatmaynotberenewable,however,wasteheatmaybea
significantsourceofenergytobeharvested.Generally,thetermrenewabletendstobepaired
asinexhaustibleinthecontextofenergy,sotheclassificationofharvestedenergydependson
thisdefinition.Inthesensethatallprocessesareinherentlyinefficient(asstatedinthesecond
lawofthermodynamics),thereistheoreticallyaninexhaustiblesupplyofwasteenergyand
fractionsofitmaybeharvestedfrominefficientprocesses.
Generally,ifwasteenergyinasystemisharvested,theoverallefficiencyofthesystemis
increased.Wastevibrationenergymaycomefromrotatingmachinery,manufacturingprocesses
3
withhydraulicmachinerypresses,conveyorbelts,electricmotors,orengineswhichmayor
maynotbefueledbyrenewableenergy.Inthiscaseofthisstudy,wastevibrationenergyin
roadwaysisthedesiredenergytobeharvested.
ThesubjectofthisinvestigationistoexaminetheharvestingofenergyfromCalifornia
roadways.Informationaboutrailwaysissparse,butitisprovidedwithinthisreportifitis
available.Vehiclesdrivingalongthehighwayorcitystreetgeneratevibrationasthevehicle
treadencountersthetextureofthepavementandthevehiclesuspensionundulatesfrom
variationsinheightalongtheroadway.Thekineticenergycontainedinthesemovementsgoes
unusedonasystemlevel,althoughtheseprocessesarepartofthephysicsincreatinga
comfortableandfunctionalrideinavehicleandmaintainingtraction.Themainsourceof
energyinapiezoelectrichighwayenergyharvestingeventistheimpactofthevehicletire(and
theweightitbears)asittransitionsoverapiezoelectricdevice.Thisenergyiskineticenergy
thatgoesotherwiseunused,andisanacceptedinefficiencythatcomeswiththeuseofvehicles
asatransportationmode.Harvestingafractionofthisenergymaybeasourceforincreasingthe
overallefficiencyoftransportationinfrastructure.
1.3 What are piezoelectric materials commonly used for?
Themajorityofliteratureforpiezoelectricmaterialsisdirectedtowardvibration,ultrasonic
sensors,andtransducers.Apiezoelectricdevicerequiresastresstofunction,suchas
compressionfromoutsideforces.Thefirstapplicationofstresswillgeneratevoltageand
current(power)withinthematerial,butthestressmustberelaxedinorderforthematerialto
generatepoweragain.Inthisway,piezoelectricmaterialsrequirealternatingstressor
vibrationtofunctionpseudocontinuously.
Thefieldofacousticsdeployspiezoelectricmaterialsforawiderangeofsensorsandactuators.
Thesesensorsvibrateatveryhighfrequenciesabovetherangeofhumanhearingwhichallows
themtobeverysensitivetomicroscalephysicalfeatures.Ultrasonicacousticpiezoelectric
transducersareoftenusedtoinspectweldsinpipelines,forexample,tolookforvoids,cracks,
orotherdefectsthatmightmaketheweldincapableofholdingpressure.Inthecontextof
energyharvesting,piezoelectricsareoftenconsideredassmallscaleenergyharvestingsources
topowerasensornetworkorsensorarray.
Forthepurposeofenergyharvesting,thevibrationfrequenciesaretypicallymuchlowerthan
whatisrequiredforultrasonics.Recallthatpiezoelectricmaterialscanbestressedor
compressedtocreatecurrent,orcurrentcanbeappliedtogenerateastress.Ahumanwalking,
forexampleisalowfrequencyeventthatcanbecapturedintheformofstressonapiezoelectric
platform.Apersonwalkingacrossaroommaycomplete12stepspersecond.Eachstep
introducesastressintheflooroftheroom,andthefrequencyofthat alternating stresswouldbe
about12vibrationspersecond,andthiswastevibrationalenergycanbeharvested.
Vibrationspersecondareameasureoffrequency,oftenstatedinHertz(Hz).Onevibrationper
secondisequalto1Hz.Twovibrationspersecondareequalto2Hz.ThecommonUnited
Stateshouseholdselectricalcircuitcarrieselectricityoscillatingat60cyclespersecond,or60
Hz,whichisevidencedbythelowfrequencybuzzofanelectricshaver.Anultrasonicsensor,
4
however,mayvibrateatthousandsortensofthousandsofcyclespersecondandthismaybe
abovetherangeofhumanhearingwhichrangesfrom20Hzto20,000Hz.Ultrasonicsvibrate
abovethehumanrangeofhearing,ashighas100,000Hz.Oneofthemostcommonultrasonic
technologiesthatmostpeopleexperienceistheultrasoundusedtoimageunbornbabiesin
thewomb.Thescanisperformedbyanultrasonicdevicewhichprocessesreflectionofthe
soundwavestoproduceanimageofthebaby.Thisultrasoundequipmentoperatesaround1
millionHzto18millionHz(MHz).TherangeofthesefrequenciesisillustratedinFigure1.
Whyisfrequencyimportant?Oneneedstounderstandthatthetypeofpiezoelectricthatbest
harvestsenergyshouldhaveafrequencyresponsesuitableforthevibration.Piezoelectrics
designedforultrasonicswouldhaveverylittlesuccessiftheywereputintoanapplicationto
harvestenergyfromwalking.Theywouldalsolikelybetooexpensive.Thepiezoelectric
materialsrelevanttothisapplicationaredesignforlowfrequencyvibrationharvestinginthe
rangeof100120Hz
1
.

Figure 1: Energy Harvesting of Vibrations is Possible typically in the Lower Frequency Range
from 1-1,000 Hz

Source: DNV KEMA

Harvestingthevibrationenergyfromhumanswalkinghasbeenapasttargetofpiezoelectrics.
Therehavebeenstudiesfocusedonfutureenergyefficientcitiesthathavesolarpanelsonthe
rooftopsofbuildingsandpiezoelectricsinthesidewalkstoharvestenergyfromfoottraffic.
1.4 What are some relevant cost and energy metrics?
Todeterminehowmuchenergypiezoelectricscanproduce,afewmetricsneedtobedefined
priortodiscussion.

1
CookChennault.PiezoelectricEnergyHarvesting:AGreenandCleanAlternativeforSustainedPower
Production.BulletinofScience,Technology,&Society,Vol28,No6Dec2008pp496509.
5
Thefirstispower.PowerisdefinedinWatts(W),whichisdefinedasunitsofenergyper
second.Powerisanindicationofhowquicklyenergycanbedelivered.Apowerfulair
conditionercancoolaroomquickly,whereasaweaklypoweredheatermayrequirealongtime
toheataroom.Otherexamplesincludeasolarpanelwhichmayberatedat200Winpeak
sunlightatnooninthemiddleofasummerCaliforniaday.Anaturalgaspowerplantmay
produceasmuchas200millionwatts(megawatts,orMW)topoweracityanditssurrounding
neighborhoods,onemilliontimesmorepowerfulthanasinglesolarpanel.
Thesecondmetricisenergy.Energyisdefinedinmanyunits.Instandardunits,energyis
statedinJoules(J),butforelectricityitisoftenmostusefultodefineenergyintermsofwatt
hours(Wh),forexample,howmanywattsareproducedinanhour.Intheexamplesabove,the
solarpanelwouldproduce200Whfromnoonto1PM.Thenaturalgaspowerplantwould
produce200millionwatthours(200megawatthours,orMWh)inthesamehour.Again,the
twoexamplesaredifferentbyafactorofonemillion.
Whendiscussingpowerandenergysystems,itishelpfultotalkabouthowmuchpoweror
energycanbemadewithinafootprint(area)orbox(volume).Thesearestatedasmetricssuch
aspowerdensityandenergydensity.Forsystemsthatareflatlikesolarpanels,powerdensity
mightbeinunitsofwattspersquarefoot(orsquaremeter).Considerthesolarpanelexample
fromabove,producing200Wor200Whinanhour.Atypicalsolarpanelmightmeasure2ftx3
ft,orsixsquareft(6ft
2
).Itspowerdensitywouldthenbe200wattsinsixsquarefeet,or
200/6=33W/ft
2
.Thenaturalgaspowerplantmightoccupyaspaceof100,000squarefeet,
perhapsfencedoffinaremoteplaceoutsideofthecity.Itspowerdensitywouldbe200million
wattsin100,000squarefeet,or2,000W/ft
2
.Similarly,theenergydensityofthesesystemsis33
Wh/ft
2
forthesolarpaneland2,000Wh/ft
2
forthepowerplant.Thesemetricsaredescribedona
scalebarinFigure2below.

Figure 2: Power and Energy Scale for Piezoelectrics, Photovoltaics, and Conventional Power

Source: DNV KEMA

6
Thesenumbersbecomeinterestingwhenappliedtopiezoelectrics,particularlyinpowerand
energydensity.ConsidertheDigitalSafariGreenbizCompanyproduct.Itestimatesthata3x5
feetpanelwillgenerate17.5wattsperstep.Humanfoottrafficoverthispaneloccursat
approximatelytwostepspersecond(2Hz).However,notethatthepoweroutputisnot
continuous,becausepowerisgeneratedonlywhenahumanstepsonthepanel.Thepower
densityoftheflooringproductis17.5Wper15ft
2
,or1.2W/ft
2
,about30timessmallerthana
solarpanel.Theenergydensityisdifferentbecauseitdependsonhowoftenpeopleare
steppingonthepanel.Atbestthereisnearlycontinuousfoottrafficonthepanelresultingina
nearlycontinuous17.5Wgeneratedwhichwouldresultin17.5Wheveryhour,oranenergy
densityof1.2Wh/ft
2
.Therefore17.5Wisthemaximumpowerthepanelgenerates,anditwill
belessiffoottrafficisless.Infact,energydensityislinearlyproportionaltothefoottraffic,as
showninFigure3.Thesameruleappliestoasolarpanel;itsenergydensityvarieswith
sunlightandithaszeroenergydensityatnight.
Anothermetricworthdiscussingiscapacityfactor.Therelationshipoftrafficvolumeto
capacityfactorisimportantfortheconsiderationofpoweroutputforaroadwayenergy
harvestingsystem.Powerequipmentusuallyhasanameplateratinglikethegasturbine
mentionedabove,forexample:a1.5MWwindturbine,a200Wattsolarpanel,ora100kWgas
microturbine.Thesenameplateratingscarryunspokenqualifiersassociatedwithpeakoutput
andhaveanimpactontheenergygenerated.Inreality,thewindturbinelikelyspinsabout30
percentofthetimewhenwindconditionsarefavorable.Thereforeitmayberatedat1.5MW,
butthisdoesnotmeanitproduces1.5MWhperhour.Instead,itproduces0.3*1.5MWh/h,or
about0.5MWh/h.Thefractionoftimethatthepowerequipmentproducespoweristhe
capacityfactor.
Figure 3: Energy Density Depends on Foot Traffic for the Piezoelectric Floor

Source: DNV KEMA

ThelessonlearnedfromFigure3appliestoaroadwaymodelalso,andimmediatelyimplies
thatanenergyharvestingsystemwillbenefitthemostfromroadswithhightrafficvolumesin
7
thesamewaythatapiezoelectricfloorwillbenefitfromhighfoottraffic.Amajorfocusofpast
researchandanessentialpartofanyUnitedStatesbaseddemonstrationswillbeastudyof
energyproductionversustrafficvolume,averagevehiclespeed,andevenvehicleweight.This
willbeexplainedfurtherinChapter3:ConclusionsandRecommendations.Energyandpower
densityareexpressedbyvendorsintermsofkWh/kmandkW/km,respectively.Howeveritis
alsousefultodiscusspowerdensityintermsofW/moduleorW/ft
2
ofdevices.
Thecostofthesesystemsisdisparateacrosstheliterature,buttwocostmetricsareofuse.Floor
basedpiezoenergyharvestingsystemssuchasthoseadvertisedbyPiezoPoweruseaRochelle
saltpanelmeasuring1500ft
2
inareafor$2,250,or$1.50/ft
22
.TheInnowattechroadwaysystem
hasbeenquotedtocost$650,000fortheinstallationofonekmofroadway,whichgenerates100
kW.Thiscorrespondstoanovernightcapitalcost
3
of$6,500/kW.Thesetwotechnologiesplacea
widerangeonthecostpersquarefootandthiscanbeattributedtotheinstallationneeds.
Innowattechhasoptimisticallystatedinpressreleasesthattheinstalledcostcanbecutbytwo
thirds.
Efficienciesofpiezoelectricmaterialscanrangefrom2030percentforsomedevicesandaslow
as 1015 percent for low cost devices. These roadway piezoelectric devices are engineered
towardlowcostandthereforehaveefficiencyatthelowerendofthespectrum.
Becausetheapplicationofpiezoelectricmaterialsasbulkenergyproducingdeviceshasonly
recentlybeendemonstrated,thereisawiderangeofliteratureconcerningpowerandenergy
densitycharacteristicsofthesematerials.Somematerialsarelikelybettersuitedformicro
harvestingapplications(suchasinsensornetworks)whileothersaremoresuitableforbulk
powerharvestingandproduction.AscanbeseeninFigure4,piezoelectricdevicescanbe
comparedtothepowerdensityofLibatteriesinsomecases,buttherangeofpowerdensitiesis
quitelarge
4
.Thisilluminatesthefactthatthetechnicaldatabaseonthesematerialsislargeand
diverseandthereforeastudyoftheirapplicationtowardroadwaysrequiresafocusedstudyof
thespecificmaterialsanddevicesavailable.

2
Walsh,etal.PiezoPower.DigitalSafariGreenbizzCompanyBusinessPlanCompetition,2011.
3
Overnightcapitalcostisanestimateforthematerialsandinstallationforenergysystems,anddoes
notincludethesometimesimmeasurablecostsofpermitting,constructiondelays,andotherdelayswhich
addtothecostofconstructionprojectsthatarespecifictothelocation,thecontractors,andthe
technology.Overnightcostsareagenerallyacceptedcomparisonforenergysystemsandareoftenquoted
inthisfashioninDOE,EIA,andIEAdocuments.
4
CookChennault.PiezoelectricEnergyHarvesting:AGreenandCleanAlternativeforSustainedPower
Production.BulletinofScience,Technology,&Society,Vol28,No6Dec2008pp496509.
8
Figure 4: Piezoelectric Devices have Voltage and Power Capability Similar to Li-ion Batteries

Source: Cook-Chennault, et al
5

1.5 Two Ways to Harvest Energy with Piezoelectric Materials

Asmentionedaboveinthesectiontitled1.1WhatarePiezoelectricMaterials?,apiezoelectric
materialgeneratesenergywhenitisstressedandstrained.Thisprimarymechanicalstress
generatesavoltageandcurrentpulse,theproductofwhichispower(Watts).Thedetailsof
roadwaypiezoelectricenergyharvestingsystemsaregiveninAppendixA.
Therearetwotechnologiesreviewedinthisreportthatharvestenergyintwodifferentways,
butbothrelyonstressingapiezoelectricelement.Inthefirstcase,astackofpiezoelectric
materialsiscompressedtogenerateenergyandapulseofpowerisregisteredwitheach
compressioncycle(suchasthepassageofavehicletireoverthestack).Inthesecondcase
basedonwhatcanbesimplifiedandinferredfromproductliteraturetheenergyharvesting
mechanismappearstobeanarrayofcantileverorbentbeampiezoelectricenergyharvesters
vibratingasaresultofexternalstimulus(suchasthepassageofavehicletireoverhead).The
powerprofilecontinuestogeneratepowerforalongerdurationasthevibrationsdecay.An

5
CookChennault.PiezoelectricEnergyHarvesting:AGreenandCleanAlternativeforSustainedPower
Production.BulletinofScience,Technology,&Society,Vol28,No6Dec2008pp496509.
9
exampleofavibratingdeviceisshowninFigure5incomparisontoacompressionbased
energyharvestingdevice.
Anarrayofmanydeviceswithinasmallvolumeleadstohigherenergyorpowerdensity.
However,thearrayconfigurationanddurationofoutputwilldeterminetheseproperties.

Figure 5: Difference in Performance Characteristics of Compression-based Energy Harvesters and

Cantilever Energy Harvesters

Source: DNV KEMA

1.5.1 Energy Density of a Compression-based System

Acriticalassessmentofthecompressionmechanismforharvestingenergywasprovidedby
UniversityofCalifornia,Berkeley
6
(Berkeley).Thecalculationexplainswithmathematical
justificationthatsimplecompressionofagivenvolumeofspacedoesnotitselfactualize
significantenergy.However,thecalculationneglectspossibleengineeringinnovationsinthe
piezoelectricmodulewhichcanincreaseenergydensityandamplifytheeffect.Thisis
highlightedbyademonstrationfromVirginiaTech,whichusesalabdesigned,simplified
piezoelectricpowermodulethatgenerates100timesmorepowerthanwhatBerkeleyconcluded
ispossiblewithoutanysignificantengineeringotherthanoptimizedplacementofstacks.

6
Waterbury,Wright
10
Berkeleycalculatedthatthemaximumamountofenergyimpartedduringacompressionof0.08
m(3)is6.6x10
5
J(lessthan1mW).VirginiaTechhasbuiltaprototypeenergyharvesterthatis
explainedindetailinAppendixA.TheVirginiaTechprototypeenergyharvesterhas
demonstratedanoutputof0.080.14Wfromthesamefootprint,sothereisadiscrepancyin
whatBerkeleyhascalculatedversuswhathasbeendemonstrated.Someofthisdiscrepancyis
intheassumptionsforthecalculation,andothersourcesofdiscrepancyarisefromengineering
modificationstoincreasepowerdensity.Commercialcompressionbasedenergyharvesters
containmechanicallinkagestomultiplyforcestoincreasetheirpowerdensity.VirginiaTech
modeledaplateoverpillardesigntodistributeloadacrosssmallpiezoelectricstacks(Figure6),
whichhasinherentlylowerenergydensitythanwhatisimpliedbycommercialdesigns.The
VirginiaTechunitdoesnotincludeanyforcemultiplyingmechanismorleverconfiguration,
yetithasdemonstratedgreaterpoweroutputthanwhatBerkeleycalculated.

Figure 6: Configuration of Stacks in the Virginia Tech Piezoelectric Harvester

Source: Virginia Tech

ThetargettrucksintheVirginiaTechstudyweretractortrailerswhichareaClass8weight
ratingat33,000poundsormore.Tractortrailershavefiveaxles:twoonthetrailer,tworear
axlesonthetractor,andonefrontaxle
7
.ThecharacteristicsoftheBerkeleytruckdonotmatch
anyUnitedStatesvehicleclass,sothemodifiedcalculationusesClass8vehiclesasthebasis
whichdescribesthesametractortrailersdemonstratedintheVirginiaTechstudy.Usingthis
estimate,theloadshouldbeincreasedperwheelto14,700Newton(N),oraboutfivetimesthe
Berkeleyassumption.Usingthefiveaxletruckcorrectsthenumberofcyclesto2.9x10
7
.Using
thesimilarestimatesforthedimensionsoftheharvestersandaccountingforthecorrectionsfor

7
Berkeleyestimated8axles.
11
weightandaxles,theBerkeleycalculationmethodyieldsabout0.01kW/kmratherthanthe
0.0018kW/kmthatwasestimatedinthepaper.Nevertheless,thisisstillmuchlessthanwhat
vendorshaveclaimed.ItisalsolessthanwhathasactuallybeendemonstratedbyVirginiaTech
(Figure7).
DatasharedwithVirginiaTechbyInnowattechindicatedthata4mmthickstackof15mm
diametercouldgenerate0.25Wwith64percentefficiencyat100MPaofload.Thisisequivalent
to64,516N/in
2
.WhencomparingFigure8andFigure9itcanbeseenthatsuchloadsarenear
expectedvehicleloadsatthetire.Intermsofpowerdensity,thisis0.25W/.001ft
2
,orabout
250W/ft
2
.Thisiswhatispossiblefromthematerialsbutitisacknowledgedthatthisisnot
necessarilywhatispossibleinpractice.Thereareseveralinefficienciesinvolvedinpackaging
andconvertingthispowertousefulenergy.Thisdata,inadditiontowhathasbeen
demonstratedbyVirginiaTech,impliesthatitispossibletoharvestpoweratlevelshigherthan
whatBerkeleycalculated.
ThedemonstrationfromVirginiaTechmeasured0.080.14Wforasinglecompressionevent
(exampleinFigure7)whichis100xlargerthanthe~1mWoutputcalculatedbyBerkeley.
Withina1kmstretchofhighway,thewheelbaseandshadowfootprintofthevehicles
occupyingthespacewilldeterminehowmanydevicescanbesimultaneouslystimulated.
Becausethecompressivebasedenergyharvestersgeneratepowerinsharp,discretepulses,
thereisverylittleoverlapbetweenexcitedharvestersandinactiveharvesters(see1.5.2 Effect
ofWheelbaseonCapacityFactor).ThecomparisonbetweentheBerkeleycalculation,the
correctedestimateusingUStractortrailersasthebasis,andtheVirginiaTechdemonstration
areshowninTable1.

Figure 7: Power Output from a Single Innowattech Energy Harvester during the Virginia Tech
Demonstration

Source: Virginia Tech

12
Table 1: Comparison of Calculation Results across Multiple Third-Party Investigators into the
Compression-Based Energy Harvesting Technology
Berkeley Result Modified Berkeley Calculation Virginia Tech Demonstration
600 trucks per hour 600 trucks per hour
8
As low as 167vehicles per hour
45,000 N truck weight (10,0000
lbs)
147,000 N truck weight, tractor
trailer (33,000 lbs, Class 8)
Tractor Trailer
8 axles per truck Tractor trailer, 5 axles 5 axles
0.0018 kW/km output, <1 mW at
the module
0.01 kW/km, 0.017 W/module 0.08-0.14 W per module
Source: DNV KEMA

TheBerkeleycritiquemakesacompellingargumenttoshowthatenergydensityisachallenge
forthistechnology,butitignoresengineeringinnovationonmechanismstomaximizepower
density.Inthesamewaythatphotovoltaic(PV)systemscanemploymirrorsandlensesto
concentratelightonactivemodules,thepiezoelectricmodulescanusemechanicaladvantage
andnovelpackingofmaterialstothesameeffect.Whileconservative,thedemonstrationdata
showsthatmoreispossiblethanwhatBerkeleycalculated,yetnotenoughtobefinancially
compellingwithasimpledesign.Commercialdesignsarefurtheralongtheproduct
developmentcyclethantheVirginiaTechdesign,andhavepresumablyovercomesomeof
thesechallenges.Thereforethereisreasontoshowcarefulconsiderationtoinvestmentin
demonstrationwhileaccountingforthenovelaspectsofcommercialproducts.Itisproposedin
therecommendationssectionthatifanevaluationpathischosen,itbemanagedinastagedgate
fashiontominimizeriskandinvestmentinanR&Dendeavorbyfirstperformingindependent
testsofcommercialpowermodulestoverifytheirclaimedoutput.Anyevaluationshouldstart
ataminimumwithalaboratoryindependentconfirmationofthemodulepoweroutputfrom
eachofthevendors.Suchtestingcanvalidateclaimswithoutaskingvendorstorevealtheir
intellectualproperty,aswellasconservativelyaddresstheconcernembodiedintheBerkeley
calculation.

1.5.2 Effect of Wheelbase on Capacity Factor on kW/km

Recallthediscussionofcapacityfactoronpage7.Themaintakeawayfromthecapacityfactor
discussionisthis:becausethepiezoelectricsystemisdistributedoverawidearea(forexample,
a1kmstripofroadway),thesystemischallengedtobe100percentactive,thatis,itisunlikely
thatallmodulescanbegeneratingatthesametime.Inordertounderstandthefactorsthat
contributetothecapacityfactorofapiezoelectricsystem,theVirginiaTechdemoisusedto
considerasimplewalkthroughofhowtheenergyharvestingsystemworks.

8
600vehiclesperhourisacommonlyquotedvendormetric.Moredetailisrequiredinorderto
accuratelyassesstrafficcharacteristics,asexplainedinthetrafficmodelbelow.
13
A1kmstripofhighwaysitsempty.Imbeddedinitspavementaretwoparallelrowsof
energyharvesters,eachnumbering4900unitsforatotalof9800units.Eachofthem,
whencompressed,generates0.1W.
Asingletruckcomesfromthedistanceandentersthis1kmstripofhighway.
Asthesetofwheelscrossesthefirsttwoenergyharvesters,afrozenframesnapshotin
timerevealsthateachharvestergenerates0.1W,foranadditivepoweroutputof0.2W.
Asthefirstaxlecrossesintothesecondrowofenergyharvesters,thenexttwo
harvestersareexcitedtoproduce0.2W,andthefirstrowisalreadyrelaxingtoan
uncompressedstate.Atthispoint,thenetpoweroutputremainsat0.2W.
Bythetimethelastaxleofthetruckhascompletelyenteredthepiezoelectricenvelopeof
theroad,itwillhavefiveaxlestouchingharvestersatanygiventime.Ifeachunit
generates0.1W,thismeansatanygiventimethetruckisproducing1W.
Thelengthofatypicaltractortrailerisabout74feet(888inches).Whilethistractor
traileroccupiesspace,noothervehicleshalloccupythatspace.Thefiveaxlewheelbase
ofatractortrailervariesbutitcanbeestimatedat68feet.
IftheVirginiaTechdevicesarelaidendtoendincontactwithoneanothertomaximize
theirdensitywithintheroadway,tworowsof111unitseachcanfitwithintheshadow
ofatractortrailer.Withintheshadowofthetrailer,only10unitsoutof222arebeing
excitedatanygiventime.Outofapotential22W(nameplatecapacity),1Wisbeing
generated.
ThethoughtexperimentisillustratedinFigure8whereitisshownthat10energyharvestersare
activated(showninred)withinthefootprintofatractortrailer.Bythisillustration,inonehour
thenameplatecapacityoftheharvesterswouldindicate22Wh,butthenetproductionwould
onlybe1Wh.Thisrepresentsacapacityfactorof4.5percent.
Within1kilometer,44truckscanfitiflaidendtoend.Thatmeansthatwithin1km,440
piezoelectricunitsareproducingpoweroutof9800,orabout4.5percent.Thenameplate
capacityofthe1kmstripwouldbe980W,butitsactualoutputwouldbe44W,orlessthanan
incandescentlightbulb
9
.Thisistheminimumcapacityfactorlimit.
Becausethepulseofpowerinthecompressionbasedpiezoelectricunitisrelativelyfast(~0.1s),
onemightbeabletoimaginetraffictravelinginfinitelyfastsuchthatthedurationbetween
pulseswouldbesoshortthatallunitswouldbeenergizedcontinuously.Thisimpliesthatthe
harvestersshouldexperienceanimpactataminimumofevery0.1sinordertobenearly
continuouslyon.Highwaytrafficspeedscaninduceanearlyconstantactiveconditionforthe
harvestersdependingonspeedandvehiclewheelbase.Giventhelengthoftheaverage5axle
tractortrailerof68feet,forsimplicityitmaybeassumedthatanaveragedistancebetweenaxles

9
TheseassumptionsarebasedontheVirginiaTechdemounit,whichisprototypicalandnoncommercial
andgenerateslesspowerthanwhatcommercialproductshaveclaimed.
14
of13.6feet.Avehiclespeedof65mphcorrespondsto95ft/s,whichindicatesthatimpactsoccur
onaverageevery0.13secondswhichbeginstoapproachthepoweroutputduration.Recallin
previoussectionsthatfrequencywasdefinedinHz.Apowerpulseevery0.13secondswould
correspondtoacompressionfrequencyof7Hz.
Acontinuouslineoftractortrailersconnectedatthebumpersmovingat65mphwould
producea0.1spowerpulse,followedbytheremaining0.03secondgapuntilthenexttire
impactstheunitproducinganother0.1spowerpulse.Thisidealizedcalculationwouldestimate
acapacityfactorof0.1s/0.13swhichis~76percent.Inreality,trafficisnotevenlyspacedandthe
powerdeliveryisnotflat,andthedensityoftractortrailersisnotperfect.Asaruleofthumb,
traffictendstobespacedatleastonetotwovehiclelengthsbetweeneachvehiclewhich
indicatesthatonlyoneoutofeverythreespacesforvehiclesareoccupied(cuttingmaximum
capacityfactorestimatesby1/3).Therefore,onecouldseehowcapacityfactorsof~2030percent
maybepossibleinhighspeedanddensetraffic.Nonetheless,theVirginiaTechdemonstration
illustratesamaximumlimitofabout980continuouswatts,andestimatingcapacityfactorsof
4.550percentcorrespondstoapoweroutputof44440Watts/km.Notethatthisiswattsnot
kilowatts.
Thelastargumentconcerningcapacityfactoralsoillustrateshowcapacityfactorsfortrainscan
beincreasedsincetrainsarecloselyspaced.Howeverthereisdowntimebetweentrains.
Piezoelectricsystemsonbusytracklineswouldhavemaximizedcapacityfactors.

Figure 8: Ten (red) out of 222 Piezoelectric Units can be Excited at Any Given Time with a Single
Passing Tractor Trailer at Low Speeds

Source: DNV KEMA

Themajorityofvehiclesontheroadarenottractortrailers.Usinginformationfromthe
TransportationEnergyDataBook
10
,onecanseethatthemajorityofvehiclesontheroadareof
thetwoaxlevehicle(cars)andlightdutytruckvarieties.UsingtheVirginiaTechdataagainasa

10
Davis,Stacy;Diegel,Susan;Boundy,Robert.TransportationEnergyDataBookEd.31.July2012,Oak
RidgeNationalLabsORNL6987
15
standardoutputmetric,ifoneassumesthataClass8tractortrailergeneratesamaximumof
0.14Wperwheelimpact,onecanscalethepoweroutputlinearlywithweighttoestimatethe
netoutputpervehicletype
11
.Figure10revealsthatthosevehicleswiththegreatestenergy
harvestingpotentialarethefewestontheroad.Ofallvehiclemilestravelled(VMT),trucksand
bussescomprisesmallfractions.Thereforethereisaneedtoplaceenergyharvesterswherethey
canbeoptimizedforpoweroutputandcapacityfactortotakeadvantageofhightrafficflow
rateswithamaximizedcrosssectionofheavyvehicles.ThisisperhapsthereasonwhyVirginia
Techchoseatruckweighstationonahighway,asitwouldtargetheavyvehicleswhile
attemptingtocaptureahightrafficflowrate.
TheestimationofpoweroutputversusvehicleweightwaslinearizedfromtheVirginiaTech
demonstration(calculationshowninFigure9).Shorterwheelbasesleadtohighercapacity
factors,butlargerwheelbasestendtobeattributedtoheaviervehicles.Becauseheavyvehicles
tendtorequiremoreaxlestodistributeweight,wheelbasedoesnotgrowlinearlyasvehicle
weightgrows,buttendstoleveloutforheaviervehicles.Forexample,atractortrailerthat
measures68feetlongwithfiveaxleshasanaveragedistancebetweenaxlesof13.6feetandit
willbecarryingloads>33,000pounds.Abuscanhavesimilarloadrequirementsbutmayhave
ashorterwheelbaseof1519feetwithonlytwoaxles.Becauselightervehicleshaveshorter
wheelbases,theytendtohavehighercapacityfactors,yetlightervehiclesgeneratelesspower.
Thevehicletypeversustheestimatedcapacityfactorat40and65mph,respectively,isshownin
Figure11.ComparisonofFigure10toFigure11showstheengineeringcompromiseinroadway
energyharvesting;heavyvehiclesgeneratethemostpower,buttheyarelessfrequent,while
smallervehicleshavethehighestcapacityfactoryettheygeneratetheleastpower.

Figure 9: Estimation of Power Output as a Function of Weight on the Vehicle Wheel (Virginia Tech
Basis)

Source: DNV KEMA

11
Itisnotknownatthistimewhetherpowerscaleslinearlywithweightbutitcanbeapproximatedfor
thisreport.
16
Figure 10: The Majority of VMT Composed of Passenger Cars and Light-Duty (Noncommercial)
Trucks
0.00
0.05
0.10
0.15
0.20
0.25
0% 10% 20% 30% 40% 50% 60%
P
o
w
e
r

p
e
r

I
m
p
a
c
t

(
W
)
%ofVMT
0.30
LargeTrucks
Buses
OtherSingle
UnitTrucks
LightDuty
Trucks
Motorcycles
Cars

Source: DNV KEMA

Figure 11: Vehicles with Shorter Wheelbases are also Lighter, but Have Higher Estimated Capacity
Factors
0%
5%
10%
15%
20%
25%
30%
35%
40mphCapacityFactor
65mphCapacityFactor

Source: DNV KEMA

1.5.3 Increased Capacity Factor through Longer Power Duration

Thepreviousdiscussionindicatesthatcapacityfactorisdependentonwheelbaseandvehicle
speed.However,itisalsodependentonthelengthofthepowerpulsefromtheenergy
harvester.Amanufacturer(Genziko)claimssuchadvantages(seeAppendixA).TheGenziko
productliteraturedisplaysdatabasedontrafficvolumesbutalsomakesanefforttoquantifya
capacityfactorbasedonvibrationfrequencyharvestedfromtheroadway.At65mph,tractor
17
trailerimpactswouldoccurat7Hz.Therefore,theremustbeanothervibrationharvested,
perhapsassociatedwithroadtexture,tiretexture,vehiclespeed,andvehicleweight.The
Genzikodesignharvestsvibrationalmodesusinganarrayofmicroscalepiezoelectricmaterials
inordertoextendthepowerpulse.Theextentofthispowerpulseiscriticaltowhetherthe
systemcangenerateprofitableenergy.

Figure 12: Genziko Claims Significant Power Density over Competing Technologies

Source: Genziko

18
CHAPTER 2:
Cost Analysis of a Piezoelectric Roadway Energy
Harvesting System
Bothroadwayandrailwaysystemsareconsideredinthisreportbuttherearemorereference
materialsavailableforaroadwaysystem.Therefore,itisnotpossibletoprovideasmuch
informationaboutthecostofenergyforarailwaysystem,althoughitwouldappearthatthe
capitalandinstallationcostsofrailwaysystemsarelessthanthecostsforroadwaysystems.The
followingsectionsestimatethecostofroadwayenergyharvestingsystemsusingthe
evaluationsofvendorclaimsaswellassimplifiedtrafficmodels.
Ananalysisofthecostmetricsindicatesthataroadwayorrailwaymodelwouldcomprisea
numberofkeyfactorsforconsideration:
Maximumpoweroutputofmodule(itsratedpowerdensity)
Durationofpowerpulsefrommodule
Lifetimeofthesystemanditscomponents
Trafficvolume
Trafficwheelbase
Weightofvehicles
Averagespeedoftraffic
Capitalcostsoftechnologyandinstallation
Maintenanceandotheroperationalcosts
2.1 Vendor Claims and Demo Data
InTable2,twodifferentvendortechnologiesarecompared.Itcanbeseenthatfromthesame
trafficflowrate,verydifferentpowerlevelsareclaimed.ThetableshowsthattheGenziko
productclaims90timesmorepowerwiththesametrafficflowrate,50percenthighervehicle
speeds,and80percentlessmodules.Disparitiesareobservedinthecategoriesoftrafficspeed,
poweroutput,thenumberofunits,andthecostperkm.Thefollowinganalyseswillseparate
outthefactorsthatleadtothesedifferencesandidentifythekeymetricsthatdifferentiateone
technologyfromanother.Whilethecapitalcostishigh,theLCOEhasanopportunitytobelow
becauseofincreasedenergydensity.Thisisexaminedinthefollowingsections.

19
Table 2: Two Different Energy Harvesting Technologies Compared for a 600 Vehicle/hr Flow Rate

Innowattech
Numbers
Genziko
Numbers
Difference - Genziko
vs. Innowattech
Vehicles per hour 600 600 0.0
Vehicle speed (mph) 45 65 0.5
Claimed power generated, 1 km (kW) 150 13,600 89.7
Number of harvesters, 1 km 9,800 2,037 -0.8
Cost per km $650,250 $27,200,000
12
40.8
Source: DNV KEMA

AsummaryoftheknownliteratureforpiezoelectricsisshowninTable3.Muchofthedatais
presentedindifferentsourcesandthereforedifferentunits.Aconsolidationofthedatainto
comparativeunitsissummarizedinTable3.

Table 3: Data Summary for Piezoelectric Materials and Installations

Parameter
Low
Estimate
High
Estimate
Source
Objectivity
Ranking (1=low,
3=high)
Optimal vibration frequencies (Hz) 100 120
Cook-
Chennault
3
Tested wheel speeds (mph) 7.5 15 Virginia Tech 3
Voltages (V) 400 700 Virginia Tech 3
Amperage (mA) 0.2 0.35 Virginia Tech 3
Power Duration (s) 0.1 0.2 Virginia Tech 3
Maximum measured power per event,
(W)
0.08 0.14 Virginia Tech 3
Virginia Tech Traffic Flow speed (mph) 40 Virginia Tech 3
Virginia Tech Traffic Flow rate (vehicles
per day)
4000 Virginia Tech 3
Oregon DOT Traffic Flow Rate (vehicles
per hour)
600 Oregon DOT 3
Energy Generated for 1.0 km, Oregon
(kWh/month)
350,000 Oregon DOT 3
Number of harvesters, Oregon DOT 6,000 Oregon DOT 3
Energy harvested for bridge mounted
devices, per vibration (microJ )
18 S.F. Ali, et al 3
Vehicle speed for micro harvesters (m/s) 25 S.F. Ali, et al 3
kW per km 0.0018 Berkeley 3

12
Basedon$2,000/kWcapitalcostprovidedbyvendor,includingenergystorage
20
Parameter
Low
Estimate
High
Estimate
Source
Objectivity
Ranking (1=low,
3=high)
units per km 10,000 Berkeley 3
Axles per vehicle 2 8
Berkeley,
Oregon
3
Source: DNV KEMA

Table 4: Data Sources from Less Objective References

Parameter
Low
Estimate
High
Estimate
Source
Objectivity
Ranking (1=low,
3=high)
Target Cost of piezoelectric material (per
unit)
$1
Channel
Technologies
2
Energy generated in 1 km stretch of road
(kWh)
400 600
POWERleap,
Treevolt
1
Time span of energy measurement (hr) 16
POWERleap,
Treevolt
1
Traffic flow rate, POWERleap
(vehicles/hr)
12.5 25
POWERleap,
Treevolt
1
Vehicles per hour 600 POWERleap 1
Power rating (kW) 720 POWERleap 1
Length of energy harvesting section (km) 1 POWERleap 1
Number of harvesters per 1 km 6,000 POWERleap 1
Power per unit per impact (W) 10 POWERleap 1
Power generated per sq ft, foot traffic
(W/ft^2)
1.13 Piezo Power 1
Cost per square foot, foot traffic ($/ft^2) $1.50 Piezo Power 1
Power rating, 1.0 km (kW) 200 Innowattech 1
Power rating, train (kW) 120 Innowattech 1
Traffic flow rate (vehicles per hour) 600 Innowattech 1
Vehicle speed (kph) 72 Innowattech 1
Train speed (wagons/hr) 300 Innowattech 1
Size of each unit (ft^2) 1 Virginia Tech 3
Power per km (kW) 100
Innowattech,
Haaretz
article
1
Cost per km ($) $215,400 $650,000
Innowattech,
Haaretz
article
1
LCOE ($/kWh) 0.06 0.08 Genziko 1
Lifetime (y) 20 Genziko 1
21
Parameter
Low
Estimate
High
Estimate
Source
Objectivity
Ranking (1=low,
3=high)
Installation cost ($/W) 0.4 Genziko 1
Capacity Factor 0.32 0.4 Genziko 1
Vehicles per hour 600 Genziko 1
Power Density (kW/km) 13,600 Genziko 1
Long dimension of unit (m) 0.45 Genziko 1
Short dimension of unit (m) 0.3 Genziko 1
Genziko Units per km 2,222 3333
Calculated
from Genziko
1
Number of harvesters 9,800
Calculated
from Virginia
Tech and
Berkeley
1
Source: DNV KEMA

2.2 Relationship between Traffic Parameters and Harvester

Characteristics
Theimportanceoftrafficdatasuchasvehicleweightandwheelbasewasshownin1.5.2Effect
ofWheelbaseonCapacityFactoronpage13.Inaddition,thedatashowninFigure13shows
thefactorsthatcontributetonetpoweroutputarevehicleweight,vehiclespacing,powerpulse
width,andwheelbase.Inordertoprioritizetheimportanceofthesefactors,theycanbe
analyzedwitharangeofvaluesforeachparameterinacalculationtomodelnetpoweroutput
oreconomicperformance.Then,regressionanalysiscanbeperformedtounderstandtheeffect
ofeachparameteronthecalculation.
Regressionanalysisisawayofobservinghowdependentvariableschangewhenindependent
variablesarevaried.Forexample,itisusefultoseehowtheLCOEofthesystem(adependent
variable)changeswhenthelifetimeofthesystemisvaried(anindependentvariable).The
regressioncoefficientisameasureoftherelativeinfluenceeachvariablehasontheLCOE.A
negativeregressioncoefficientcorrespondstoanegativeinfluenceontheLCOE,andapositive
coefficientcorrespondstoapositiveinfluenceontheLCOE.Asaresult,theregression
coefficientsindicatethesensitivityoftheLCOEtotheinputparametersandareariskranking
system.Acoefficientvalueofzeroindicatesthatthereisnorelationshipbetweentheinputand
theoutput.Avalueof+1or1indicatesa1or1standarddeviationchangeintheoutputfora
changeintheinputof1standarddeviation.
Thechartshowsthatthreeofthemainfactorsthataffectsystemlevelpoweroutputaretraffic
dependent.Theonlyfactorthatcanbecontrolledbythetechnology(besidesitsabilityto
maximizeenergyharvested)isthedurationofitspoweroutput.
22
Table 5: Correlation between Traffic and Harvester Metrics and System Power Output
Regression Coefficient Affected Power Metric
Vehicle Weight Maximum energy harvested power
output (W) at harvester
Vehicle Spacing Capacity Factor
Power Pulse Width Capacity Factor
Wheelbase Capacity Factor
Source: DNV KEMA

TheregressioncoefficientsbelowwerecomputedbyusingtrafficdatafromtheTransportation
EnergyDataBooktocalculatecapacityfactorbasedontrafficspeedandvehiclewheelbase.In
addition,distributionsofvehicleweightsfromthesamedataandassumptionsaboutvehicle
spacingwereincluded.Thesenumberswereusedtofairlyestimatethevehiclecharacteristicsof
atypicalUnitedStatesroadway,andthenthesenumberswereadjustedtomatchthenumbers
quotedbymanufacturers.Forexample,atroadwayspeedsnear65mphandavehiclespacing
rateof0.06,thetrafficflowrateapproaches600vehiclesperhourandthesystemcapacityfactor
approaches20percent.
Capacityfactoriscomputedbythetimebetweenvehicleaxlehitsdividedbythepowerpulse
width.Ifthetimebetweenaxlehitsislessthanthepulseduration,capacityfactoris100percent.
However,thereisaneedforascalingfactortoaccountforthefactthatvehiclesdonottravel
bumpertobumperandsomespacingbetweenthemispermitted,whichiscalledthevehicle
spacingoccupationfraction.
Uncertaintyinthedurationofthepoweroutput(powerpulse)isconstructedaroundthedata
fromVirginiaTech,whichindicateda~0.1secondpulsewidth.Sincemanufacturersdonot
quotetheirunitoutputdirectly,thiswasestimatedandtheinputparameterswerevariedin
ordertoapproachcostofenergyestimationssimilartomatureadvancedenergytechnologies.
Byapproachingtheproblemthisway,atrafficinferredestimationcanbeusedtocrossvalidate
vendorclaimsandrevealtheperformancerequirementsinordertoachieveclaimedcostof
energy.
Thelessonslearnedfromthisanalysisare:
1. Poweroutputisincreasedforheaviervehicles
2. Trafficspacing,wheelbaseofvehicles,andoutputpowerpulsewidthdictatescapacity
factor
3. Poweroutputcanbeincreasedinharvesterpulsewidthisincreased.
Thelastpoint(#3)illustratestheadvantagethatatechnologylikeGenzikomighthavein
comparisontoInnowattech,providedthattheoutputisverifiedandthevehicleweight
23
influenceissimilar.Ifpoweroutputissustainedforlongerdurationswithalongerpower
pulse,capacityfactorisincreased.
AsshowninFigure13,thesystemoutputfor1kmofroadwayislargelydependentonthe
weightofthevehicleanditswheelbaseinadditiontothepulsewidthoftheharvesterandthe
vehiclespacing.

Figure 13: Actual Output per 1 km of Roadway versus Relevant Outputs

-
0
.
6
-
0
.
4
-
0
.
2
0
.
0
0
.
2
0
.
4
0
.
6
0
.
8
Source: DNV KEMA

2.3 What is the Power Output Required per Harvester?

AnanalysisofthetrafficbasedmodelthatgeneratedtheregressioncoefficientsinFigure13can
beusedtotesttheeffectofinputsoncriticaloutputssuchascapitalcostandcostofenergy.This
analysisreversestheanalysisdonein2.4 CostofElectricityofaCompressionbased
PiezoelectricRoadwayEnergyHarvestingSystemVendorClaims.Inthatanalysis,vendor
claimsareusedtoestimatethecostofenergy.Inthetrafficmodel,thetechnologycharacteristics
areinterrogatedtodeterminewhatcombinationsareneededtocreateareasonablecostof
energy;itisequivalenttofixingthecostofenergyandbackcalculatingtheinputs.
Thegoalofthefollowingcalculationistodeterminewhichmetricsanidealpiezoelectric
roadwayenergyharvestingsystemwouldhaveinordertoachieveareasonablecostofenergy.
24
Thefollowingextractsusefulmetricsthatpermitdirectcomparisonacrossmanufacturersby
decouplingperformancefromtrafficbaseddata.Thetwotablesrepresentcasesforlifetimesof
15yearsand1020years,respectively.Thetableswerecalculatedwiththefollowing
assumptions:
1. Totalinstalledcost($650,000$1,000,000/kmperInnowattechassumptions)
2. Harvestersize(~8x8perInnowattechassumptions)
3. Pulsewidth(0.1s)
4. Vehiclecharacteristicsforweightperwheelandwheelbase(previoussections)
5. Vehicleflowrate(600vehicles/hr,65mph)
6. ElectricitysalepricesequivalenttothosenearSacramento,California(Figure19),
Sincethismodelisgeneralizedtoshowthesumofmeasuredenergyoutput,thecapitalcost
implicitlyassumestotalsystemcostincludingBOS,whichmayincludeenergystorageand
powerelectronicstooutputgridreadyenergy.InTable6,thesystemrequirementsforalonger
lifetime(1020years)areshown.Alifetimeof20yearsisacceptablefortypicalroad
maintenanceschedules.

Table 6: Estimation of Key Performance Characteristics of Piezoelectric Systems based on Traffic

Parameters with a 10-20-Year Life
Power Output
per Harvester
(W)
50
th
Percentile
Net Present
Value (NPV) at 5
Years
LCOE ($/kWh)
Capital Cost of
System ($/kW)
Nameplate
Power
Density
(W/ft^2)
Actual
kW/km
79 -$451,000 $0.19 $17,100 179 38
132 -$313,900 $0.11 $10,200 298 64
265 $30,190 $0.06 $5,100 596 128
Source: DNV KEMA

Thelessonslearnedfromtheabovetablesarethatanidealsystemwillhavecharacteristics
similartothefollowinglist:
1. Powerdensity>300W/ft
2
(inthiscaseamoduleoutput>150W)
2. A1020yearlifetime
3. Capitalcosts<$10,000/kW
4. ActualkW/km>100
25
Theseparameterscanbeusedasacoarsequalificationlisttodeterminethefeasibilityof
technology,andwhenpowerdensityandoutputisverifiedbytesting,themethodsshownin
thisreportcanbeusedforabetterestimationoftheLCOE.
Giventhesefactors,thesensitivityofthefiveyearNPVoftheinvestmentisdominatedbythe
weightatthevehiclewheel.Whenthisvalueisincreased,theNPVincreases.Itisusefultonote
thatadecreasingwheelbasereducesthereturnbecausethistendstocorrelatewithlighter
weightvehicleseventhoughshorterwheelbasesincreasecapacityfactor(forexample,small
carsandmotorcycles).Highercapitalcostsobviouslyincreasethecostofenergyanddelaythe
returnoninvestment.Forscenarioswithlowerpowerdensity,thecapitalcostbecomesan
increasinglystrongernegativeinfluenceonNPV.Increasingtheharvesterpulsewidthand
increasingthevehiclespacingfactor(forexample,moredensetraffic)willincreasethepayback
andtheamountearnedperkWhsoldtothegrid.Increasingtrafficspeedwillalsoincreasethe
returnrate(seeFigure14).Theexternalfactors,suchasTrafficandEconomicsthataffectsystem
levelperformance,aresummarizedinTable7.

Table 7: External Factors that Affect System Level Performance

Parameter Influence on Power System
Weight at vehicle wheel Power per module, total system power
and nameplate rating
Wheelbase Capacity factor
Cost per km Capital cost, ROI, NPV, LCOE
Energy sale price ROI, NPV
Vehicle spacing factor Capacity factor
Harvester pulse width Capacity factor, actual system output
Speed (traffic speed) Capacity factor
Source: DNV KEMA

Theabovetablesandchartsillustratetheimportantfactorsthatdeterminetheeffectivenessof
anypiezoelectricenergyharvestingtechnology.Sinceeachtechnologyhasspecific
requirements,thedeterministicvalueswillvarybutshouldgenerallyfollowthetrendsabove
andshouldproducecomparablepowerdensitymetrics.

26
Figure 14: Regression Coefficients for the NPV (at 60 months) for the Traffic Model LCOE
Calculation

-
0
.
6
-
0
.
4
-
0
.
2
0
.
0
0
.
2
0
.
4
0
.
6
0
.
8
Source: DNV KEMA

2.4 Cost of Electricity of a Compression-based Piezoelectric

Roadway Energy Harvesting System Vendor Claims
Ifthevendorclaimsareconsideredindependentlyofthetrafficmodel,thenLCOEcanbe
estimated,butitwillnotbreakdowninformationbytraits,suchastrafficcharacteristicsand
powerpulseduration.
Usingvendorsuppliedinformation,thecostofelectricitydependsstronglyonthelifetimeof
thesystemandtheassociatedmaintenancecoststoprolongitslife.Thereforethisanalysisis
basedonsomeassumptionsaboutthelifetimeofthesystemandisdividedintothreescenarios.
Inthefirst,arelativelyshortlifetimeoffiveyearsisassumed.Inthesecond,alongermaximum
lifeof10yearsisassumed.Inthethird,alifetimeof30yearsisassumed.Inallcases,theLCOE
assumesadiscountedvalueforfuturecostsovertimeandsumsthosecostsover240months(20
yearscorrespondingtotheexpectedusefullifeofroadmaterials)intoanetpresentvalue.It
alsosumstheenergygeneratedoverthattimeanddividesthediscountedinvestmenttotalby
thetotalenergygeneratedtodeterminetheLCOE.DetailsoftheLCOEcalculationareprovided
inAppendixE:CalculationDetails.
Factorsthathavenotbeenaccountedforinthisanalysisaredowntimeassociatedwith
maintenance,reliabilityoftheenergygeneratedorfailureratesofthepiezoelectricdevices,
individuallysortedcostsforinverters(assumedtobelumpedintothequoteoftheinstalled
27
capitalcost)andanyadditionalmaintenancerequiredfortheroadwayduringsystem
operation.AsummaryofthecostresultsisgiveninTable8.
Itshouldbenotedthatthiscostanalysisaddressesroadwayenergyharvestingspecifically
usingvendorclaims,whichhavebeenshowntobemuchhigherthanwhatisactually
demonstrated.Lackofdataabouttheinstallationandcapitalcostsforrailwaysystemspresents
significantuncertaintyintoanestimateofarailwaysystemcost.

Table 8: Summary of the LCOE Analysis for Three Cases

Case
Minimum
LCOE (/kWh)
Maximum
LCOE (/kWh)
Mean LCOE
(/kWh)
Standard
Deviation, LCOE
(/kWh)
Case 1: Maximum 5
Year Lifetime
$0.027 $1.15 $0.18 $0.14
Case 2: Maximum 10
Year Lifetime
$0.014 $0.41 $0.08 $0.05
Case 3: Maximum 30
Year Lifetime
$0.004 $0.20 $0.03 $0.02
Source: DNV KEMA

AsensitivityanalysisoftheLCOEfactorsisshowninFigure15.Thefigureisthesensitivityplot
forCase1,butitshowsthesametrendinallcases.Thesensitivityanalysisreflectsregression
coefficients.

Figure 15: Sensitive Factors Affecting the LCOE

Source: DNV KEMA

28
Theregressionanalysisindicatesthatinallcases,thelargestuncertaintyfactoristheclaimed
powergenerated,whichisdirectlydependentontrafficfactorsandthereliableoutputofthe
system.ThisfactorisnegativelycorrelatedwiththeLCOE,meaningthatifthepowergenerated
ishigher,theLCOEisdrivenlower.Thesecondlargestuncertaintyfactoristhelifetime.Itis
alsonegativelycorrelatedanditindicatesthatifthelifetimeislonger,theLCOEisdriven
lower.
Therearedegradationfactorsthatwilldecreasepoweroutputovertime.Sincethesensitivity
analysisindicatesthatpoweroutputisthemostimportantfactor,followedbylifetime,itcanbe
seenthatthetwofactorsarecoupledandareparamounttothesuccessfulperformanceofthe
system.
Thecapitalcost($/kW)isbuiltfromtheliteraturesourcesandisshowninFigure16andis
calculatedfromtherangeofpowervaluesandinstalledcostsextractedfromthedatasummary
fora1kminstallation(recallTable18).Themeanvalueofthecapitalcostis$2,300/kWwitha
maximumnear$10,400/kWandaminimumof$600/kW,andastandarddeviationof$1,600/kW.

Figure 16: Capital Cost of the System Based on Stated Power Capability and Capital Costs from
the Literature Review

Source: DNV KEMA

Duetouncertaintyconcerningthecapitalandinstallationcostsforrailwaypiezoelectricenergy
harvestingdevices,itisnotyetpossibletoprovideausefulLCOEprediction.However,itmay
bepossibletoassumethattheLCOEissimilartoorlessthanwhathasbeenestimatedinthis
reportduetotheassumptionthatthecapitalandinstallationcostsareless.Itmayalsobe
possibletoassumethattheregressioncoefficientswouldbesimilar,forexample,thereturnon
29
investment(ROI)wouldbesimilarlydependentontrafficvolume,lifetimeofthesystem,and
costperkilometerormile.
2.4.1 Comparison with Traffic Model
Usingthetrafficmodel,thefollowingcompressionbasedharvesterinputsareused:
Harvesterspacing:8
Harvesterpulsewidth:0.1s
Lifetime:1020years
Lengthofinstallation:1km
CostofInstallation:$650,000
Trafficspeed:45mph
Vehiclesperhour:600
UsingtrafficmetricssuchasthoseshowninFigure41andtheTransportationEnergyData
Booksexplanationinthesectiontitled
1.5.1 EnergyDensityofaCompressionbasedSystem,trafficmodelsweretestedintwocases
toestablishiftheoutputsmatchtheInnowattechclaims.First,theparametersweretunedto
achievethecostofenergythatwasdeterminedinthesection2.4 CostofElectricityofa
CompressionbasedPiezoelectricRoadwayEnergyHarvestingSystemVendorClaimsand
comparedtowhatisknownaboutInnowattechsclaims.Second,theparametersweretunedto
achievethepowerratingthatInnowattechclaimedandthencheckedforagreement.Theresults
ofthiscomparisonarefoundinTable9.Notethatbothsystemshavepowerdensitiesnearor
above300W/ft
2
.

Table 9: Cross Check of Traffic Model Against what was Determined from Innowattech Numbers
Fixed Parameter LCOE = $0.11/kWh kW/km = 100-200 Vendor Claims
Reference Traffic Model Traffic Model Innowattech
Power per unit (W) 132 265 Not given
Capacity factor 0.05 0.05 Not given
Capital cost $10,200/kW $5,100/kW Mean $2,300/kW, see
Figure 16 max is
$10,400/kW
Nameplate Power
Density (W/ft^2)
298 596 Not given
Actual capacity factor
adjusted output (kW/km)
64 128 100-200
30
Fixed Parameter LCOE = $0.11/kWh kW/km = 100-200 Vendor Claims
Reference Traffic Model Traffic Model Innowattech
Nameplate system power
(kW/km)
1,303 2,607 Not given
LCOE ($/kWh) $0.11/kWh $0.06/kWh Calculated in previous
section
Source: DNV KEMA

Table9illustratesthatifthecostofenergyistobeachieved,itisdifficulttomeetthecapital
costsquotedbyInnowattechorthepoweroutput(kW/km)whileholdingalloftheother
requirements.Alternatively,inordertoobtainthepowerratingsquotedbyInnowattech,the
capitalcostsappeartobehigherthandesiredandtheLCOEbecomeslowerthanwhatwas
calculated.Theimplicationisthatthequotesfromthevendorsmaybefrommutuallyexclusive
conditions,ortheconditionsaredifferentfromwhatwouldbeexpectedwithUnitedStates
traffic.
Refinementstothetrafficmodelmaybeneededtoclosethegapbetweenthesedisparities.
However,thepurposeofTable9istodiscovertheinherentcompromisesintechnicalsystems,
forexample,achievingalowLCOEisdifficultwhenincreasingpowerdensityaccompanies
increasedcapitalcosts.Thetrafficmodelmaybeillustratingthatvendorquotedmetricsmaybe
thebestresultsformutuallyexclusivescenarios.Itisbeneficialtoexplorethesepossibilitiesand
identifyobjectivemetrics,suchaspowerpermodule,totestinathirdpartymannerinorderto
removethecontingenciesandqualifiersassociatedwithmetricssuchaskW/km.
2.5 Cost of Electricity from Vibration-based Roadway Energy
Harvesting System Based on Vendor Claims
TheGenzikoproducthassubstantiallydifferentclaimsthanthecompressionbased
(Innowattech)device.Fora600vehicleperhourtrafficflowrate,theGenzikosalespresentation
claims$2/W($2,000/kW)installedcostincluding1MWhofstorageperMWofinstalledpower,
withperformanceof13.6MW/km.TheGenzikoliteraturediffersfromPOWERleapand
Innowattechdiscussionsinthatitspecificallymentionsenergystorage.Thereisonlyonesource
fortheGenzikodata;therefore,therearenotenoughcrossreferencedvaluestoplaceboundson
theuncertainty,andtheresultisthattheLCOEcalculationissomewhatdeterministic.Thisis
whythevaluesinTable10haveverylittlevariationastheestimatedlifetimeincreases.

31
Table 10: LCOE for the Genziko Technology based on Vendor Information
Case
Minimum
LCOE (/kWh)
Maximum
LCOE (/kWh)
Mean
LCOE
(/kWh)
Standard
Deviation, LCOE
(/kWh)
Case 1: Maximum 5
Year Lifetime
0.05 0.22 0.10 0.03
Case 2: Maximum 10
Year Lifetime
0.03 0.04 0.03 0.004
Case 3: Maximum 30
Year Lifetime
0.01 0.01 0.01 0
Source: DNV KEMA

2.5.1 Comparison with Traffic Model

TheGenzikounithasdimensionsof0.3mx0.45mx0.25cm.Theunitfootprintis~11.8x17.7
(average15).Inordertoobtainthecostofenergythatthemanufacturerclaims($0.06
0.08/kWh),itisestimatedthatfor65mphtrafficwithUnitedStatesbasedvehiclepopulations,
thefollowingisassumed:
Harvesterspacing:24
Harvesterpulsewidth:variedbetween0.21.0s
Lifetime:1020years
Lengthofinstallation:1km
CostofInstallation:$27,200,000
Trafficspeed:65mph
Vehiclesperhour:600
Theresultofthiscalculationisthatthereisanassumed3280unitsperkm.SincetheGenziko
marketingpresentationscontainalotofinformationwithdifferentclaims,thebestattemptat
identifyingasetofselfconsistentclaimswasattemptedinthelastcolumnofTable11.Inthis
table,fourfactorsaretestedwiththetrafficmodelinordertofindagreementwiththeGenziko
claims:(1)LCOE,(2)claimedpowergenerated(kW/km),(3)capacityfactor,and(4)capitalcost
($/kW).TheresultsareshowninTable11,andtheconclusionisthat,basedontheinformation
provided,itappearsthatsomeparametersaremutuallyexclusiveuntilfurtherclarificationis
provided.Forexample,itappearsdifficulttoachieve$0.06$0.08/kWhandthehighpower
outputof13.6MW/kmclaimedwhenholdingtheconstraintsonUnitedStatestraffic
characteristics.Inordertoachievehighpoweroutputsperkm,thepowerdensityoftheunits
mustbeverylarge,inthe1213kWrange.However,atthesepoweroutputlevelstheLCOE
becomesverylow(lowerthanwhatGenzikoclaims).Themainpointisthetableshowsthat
withthecharacteristicsofUnitedStatestrafficandwithwhatisinferredaboutthepower
32
outputoftheunits,thesehighpoweroutputswithanLCOEof$0.06/kWhappearstobe
mutuallyexclusive.Eitherthepoweroutputislessorthecostofenergyismore.
AnattempttomatchthemajorityoftheGenzikoclaimswasmadebymatchingthenameplate
powercapacitytothestatedpowercapacity.Ifthecapacityfactoradjustpowerissetnear13.6
MW/km,itisfoundthatthecapitalcostsapproach$2,000/kW,andtheLCOEapproaches
$0.02/kWh.Inthiscase,thecapitalcost,actualpoweroutputandcapacityfactormatchthe
claims,buttheLCOEismuchlowerthanwhatGenzikoclaims.Thisisonlypossiblewithvery
highmodulepowerratingsnear13kW.

Table 11: Cross Check of Traffic Model against what was Determined from Innowattech Numbers
Fixed
Parameter
LCOE =
$0.06-
0.08/kWh
kW/km =
13,600
Capacity
Factor = 42%
Capital Cost =
$2,000/kW
Vendor
Claims
Reference Traffic Model Traffic Model Traffic Model Traffic Model Genziko
Power per unit
(W)
3,973 12,714 2,649 13,243
Not given at
time of report
13

Capacity factor 0.32 0.32 41%
14
0.32 32-42%
Capital cost $6,521 $2,038 $7,744 $2,065/kW $2,000/kW
Nameplate
Power Density
(W/ft^2)
993 3,178 662 3,311 Not given
Actual capacity
factor adjusted
output (kW/km)
4,201 13,444 3,538 13,267 13.6 MW/km
Nameplate
system power
(kW/km)
13,035 41,712 8,690 43,450
Not given at
time of report
15

LCOE ($/kWh) $0.07/kWh $0.02/kWh $0.09/kWh $0.02/kWh
$0.06-
$0.08/kWh
Source: DNV KEMA

13
SeeAddendumattheendofthereport.
14
Notepowerpulsedurationincreasedto0.51.2seconds,mean0.8seconds.Thiswasmodifiedinorder
tokeepthetrafficflowratethesameat611vehicles/hr.
15
SeeAddendumattheendofthereport
33
2.6 DNV KEMA Estimates Based on Restricted Assumptions
TheInnowattechevaluationdeterminedtheLCOEtobe$0.11/kWhwithanaveragedcapital
costrangingfrom$2,300$10,400perkW(mean$4,000/kW).Withthetrafficmodel,two
scenarioscanberuntotestthemutualexclusivityoftheLCOEandcapitalcost.Usingthe
parametersdescribedinTable12,parametersweretunedtoeitherachieveanLCOEof
$0.11/kWhoracapitalcostof$4,000/kW,andtheresultingvalueswereachieved.Thecapital
costforatargetLCOEof$0.11/kWhisnear$10,000/kW,andtheLCOEforatargetcapitalcost
of$4,000/kWisbelow$0.06/kWh(showninTable12).Tobeconsistent,theparametersthat
wereoftenquotedintheproductliteraturesuchasvehicleflowrate(600vehiclesperhour)and
vehiclespeed(near65mph)weremaintained.Otherassumptionssuchasharvesterspacing
werekeptconsistentwiththeBerkeleyevaluation.Thetrafficdatasuchasvehicleweight
distributionandvehiclewheelbasedistributionweretakenfromtheTransportationEnergy
DataBook.EstimationsofpowerpulselengthwerebasedontheVirginiaTechdemoand
lengthened(seeTable12)byassumptionthatcommercialunitshavemechanismstodoso.Note
thatineithercase,thepowerdensityisabove300W/ft
2
.

Table 12: A Test of Mutually Exclusive LCOE and Capital Costs for the Innowattech Technology
Fixed: $0.11/kWh Fixed: $4,000/kW
LCOE $0.11/kWh $0.04/kWh
Capital Cost ($/kW) $9,615/kW $4,172/kW
Capacity Factor 0.09 0.13
Vehicle Flow Rate (vehicles/hr) 611 611
Vehicle Weight Distribution
(N/wheel)
26,486 26,486
Power Per Unit (W) 143 185
Unit Spacing (in) 8 8
Nameplate Power Density (W/ft
2
) 322 417
Nameplate Power System Rating
(kW/km)
1,408 1,825
Actual System Output (kW/km) 107 149
Units per km 9,843 9,843
Power Pulse Length (s) 0.1-0.2 0.1-0.5
Average vehicle wheelbase (ft) 11.24 11.24
Vehicle Speed (mph) 60-70 60-70
Cost per km ($/km) $600,000 - $1,500,000 $600,000 - $1,500,000
Source: DNV KEMA
34
2.7 Comparison of the Projected LCOE to Distributed Renewable
Energy Generation Sources
Fromtheaboveanalysis,ithasbeenestimatedthattheLCOEforcompressionbasedroad
applicationsrangesfrom$0.03$0.18/kWh.Themeanofthisrangeis$0.11/kWh.TheLCOEfor
vibrationbasedroadapplicationsisclaimedtobe$0.06$0.08/kWh.Usingtheestimatesfrom
TheInnowattechevaluationdeterminedtheLCOEtobe$0.11/kWhwithanaveragedcapital
costrangingfrom$2,300$10,400perkW(mean$4,000/kW).Withthetrafficmodel,two
scenarioscanberuntotestthemutualexclusivityoftheLCOEandcapitalcost.Usingthe
parametersdescribedinTable12,parametersweretunedtoeitherachieveanLCOEof
$0.11/kWhoracapitalcostof$4,000/kW,andtheresultingvalueswereachieved.Thecapital
costforatargetLCOEof$0.11/kWhisnear$10,000/kW,andtheLCOEforatargetcapitalcost
of$4,000/kWisbelow$0.06/kWh(showninTable12).Tobeconsistent,theparametersthat
wereoftenquotedintheproductliteraturesuchasvehicleflowrate(600vehiclesperhour)and
vehiclespeed(near65mph)weremaintained.Otherassumptionssuchasharvesterspacing
werekeptconsistentwiththeBerkeleyevaluation.Thetrafficdatasuchasvehicleweight
distributionandvehiclewheelbasedistributionweretakenfromtheTransportationEnergy
DataBook.EstimationsofpowerpulselengthwerebasedontheVirginiaTechdemoand
lengthened(seeTable12)byassumptionthatcommercialunitshavemechanismstodoso.Note
thatineithercase,thepowerdensityisabove300W/ft2.
Table12,theDNVKEMAestimatesfortheLCOE(at$4,000/kW)andcapitalcost(atLCOE
$0.11/kWh)areincludedtoplaceboundsontheestimates.DNVKEMAestimatestheLCOE
coststorangefrom$0.08$0.33/kWh.
ItcanbeseeninFigure17thatextensiveanalysishasbeenperformedbyNationalRenewable
EnergyLaboratory(NREL)tocomparethecostsofvariousenergysources
16
.TheOpenEI
databaseisanopendataplatformdevelopedbyNRELwiththeDepartmentofEnergy
(DOE)supportthatcatalyzestheworldsenergyinformationandprovideslinkedopendata
aboutthecostofenergyformultipletechnologiesandregions.Energyinformationanddataare
availabletouse,edit,add,download,andbuildintoanalyses,tools,anddecisions.
TheovernightcapitalcostisshowninFigure17.Ashasbeenshownelsewhere,the
conventionaltechnologiessuchascoalandnaturalgasbasedpowergenerationconsistently
enjoycapitalcostslessthan$2,000/kW.Renewabletechnologiesthatcomeclosesttocompeting
withtheselowercapitalcostsareonshorewind,hydropower,andtoalesserextentbiomass
andgeothermaltechnologies.Itisworthremarkingthatdistributedgenerationisaconsolidated
categorywithovernightcapitalcoststhatcanapproach$1,500/kW,comingclosethecostof
gridscaleconventionalpowersystems.Thiscategoryincludescombinedheatandpower
(CHP),distributedwind,residentialsolar,andothertechnologies.Similarly,theLCOEfor
variousenergysystemsisdepictedinFigure18.Notethatcostofelectricitycanrangefrom

16
OpenEI:OpenEnergyInfo.http://en.openei.org/wiki/Transparent_Cost_Database.AccessedJanuary1,
2012.NationalRenewableEnergyLaboratory(NREL),OpenGovernmentInitiative,USDepartmentof
Energy.
35
$0.05/kWhforconventionaltechnologiestoasmuchas$0.35/kWhandhigherfor
unconventionaltechnologies.PartofthedatabasethatisusedtocomputetheNRELLCOE
modelswascreatedbyCapstonetoestimatethecostofenergybasedonlocality.For
comparison,estimatedenergycostsfortheSacramento,CaliforniaregionareshowninFigure
19.

Figure 17: Comparison of Capital Costs for Various Energy Sources

Source: NREL

Figure 18: Estimated LCOE for Various Energy Systems

Source: NREL

36
Figure 19: Electricity Costs near Sacramento, California

Source: Capstone, Open EI

ThisanalysisestimatestherangeoftheLCOEforthepiezoelectricsystemtobebetween$0.08
$0.18/kWh.ThechartsinFigure20andFigure21indicatethatthecompressionbased
piezoelectricsystemmayhaveanLCOElessthanthatofsmallhydroandfuelcellsandcapital
costssimilartothosefordistributedgenerationsystems.Byvendorestimates,thecapitalcost
appearstobelessthansolarPVandfuelcells.TheDNVKEMAestimatefortheLCOEhas
significantvariationbutcoincideswithcostssimilartoconcentratedsolarpower,distributed
generation,andoceantechnologies.ComparedtotheNRELdefinitionofdistributedgeneration,
thepiezoelectricsystemappearstohavehighercapitalcoststhantheaveragedistributed
generationsystem(comprisingmainlycombinedheatandpowersystems)andanLCOEwitha
similarspread.Tolendmorecertaintytothesecalculations,anindependentverificationofthe
moduleoutputisrequired.Onlythencanonebegintoascertainwhichconditionsproducethe
mostfavorablepoweroutputandLCOE.
UsingtheestimatesfromTheInnowattechevaluationdeterminedtheLCOEtobe$0.11/kWh
withanaveragedcapitalcostrangingfrom$2,300$10,400perkW(mean$4,000/kW).Withthe
trafficmodel,twoscenarioscanberuntotestthemutualexclusivityoftheLCOEandcapital
cost.UsingtheparametersdescribedinTable12,parametersweretunedtoeitherachievean
LCOEof$0.11/kWhoracapitalcostof$4,000/kW,andtheresultingvalueswereachieved.The
capitalcostforatargetLCOEof$0.11/kWhisnear$10,000/kW,andtheLCOEforatarget
capitalcostof$4,000/kWisbelow$0.06/kWh(showninTable12).Tobeconsistent,the
parametersthatwereoftenquotedintheproductliteraturesuchasvehicleflowrate(600
vehiclesperhour)andvehiclespeed(near65mph)weremaintained.Otherassumptionssuch
asharvesterspacingwerekeptconsistentwiththeBerkeleyevaluation.Thetrafficdatasuchas
vehicleweightdistributionandvehiclewheelbasedistributionweretakenfromthe
TransportationEnergyDataBook.Estimationsofpowerpulselengthwerebasedonthe
VirginiaTechdemoandlengthened(seeTable12)byassumptionthatcommercialunitshave
mechanismstodoso.Notethatineithercase,thepowerdensityisabove300W/ft
2
.
37
Table12,theDNVKEMAestimatesfortheLCOE(at$4,000/kW)andcapitalcost(atLCOE
$0.11/kWh)areincludedtoplaceboundsontheestimates.InFigure20themeancapitalcosts
areshown.ThefigureshowsthatthemutualexclusivitythatwasfoundwiththeLCOE
estimatesandcapitalcostsisagainapparentinthiscomparison.Whileminimumcapitalcosts
couldbeaslowas$2,100/kW,90percentofthevaluesarebetween$3,700and$36,000(Figure
42).InFigure21themostlikelyand90
th
percentilevaluesoftheLCOEareshownandranked.
TheDNVKEMAestimatesfortheLCOEonaveragearehigherthanvendorclaims,ranging
fromameanof$0.07/kWhand90percentofthevaluesarelessthan$0.20/kWh.Itcanbeseen
inFigure43that90percentofthevaluesarelessthan$0.20/kWh.Thesensitivityfactorsforthis
estimateareweightatthevehiclewheelandharvesterpulsewidth.
(Figure44)andthesensitivityrankingsareidenticalforthecapitalcostestimate.Sincethese
parametersarenotexplicitlyadvertisedinproductliterature,significantuncertaintyisplaced
ontheseestimatesbuttheycanbeindirectlyderivedfromUnitedStatestrafficdata.InFigure
20,onlythemostlikelyormedianvaluesareshown,sincethemaximumvalueoftheDNV
KEMAestimateiscloseto$90,000/kWduetosignificantuncertaintyinthepoweroutput
characteristicsofthetechnology.

Figure 20: Capital Costs for Energy Systems Compared to the Piezoelectric System

Source: DNV KEMA

38
Figure 21: Comparison of the LCOE for the Piezoelectric System Compared to Other Energy
Systems

Source: DNV KEMA

2.8 Added Value: Data and Reduced Inspection Costs

Whiletheenergyharvestingdevicesmaygenerateenergy,thereisalsothepotentialtogenerate
vastamountsofdata.Thevalueofthisdatamaybedifficulttoquantify,butitcouldbe
exploredforthefollowingapplications:
1. Realtimemonitoringoftrafficflowpatternsthatcanbeusedintrafficstudiesand
evaluationoftheneedforconstructionorroadwayupdates.
2. Datatofeedintotrafficalertsandcongestionreportsduringrushhour.
3. Datatomonitortheintegrityandhealthoftheroadwaytoinformmaintenance
schedulesandperhapssaveinspectioncosts.
4. Specificdataaboutvehicleweightwhichmaybeincorporatedintotrafficweighstations.
Asisthecasemostoftenwithmonitoringsystems,theopportunityisthatthesystemprovides
datathatmaymitigatetheneedforinspectioncosts.DNVKEMAperformedacostassessment
ofStructuralHealthMonitoring(SHM)monitoringsystemsforwindturbinebladesin2011,
andfoundthelargestcomponentofthesavingswasinreducedinspectioncosts
17
.Theremaybe
asimilaropportunityforusingpiezoelectricsinrailways.
2.9 Roadways versus Railways
Theroadwayinstallationrequiresmorelaborandmaterialtoinstallthanarailwayapplication.
Theunitsareinstalledintheroadbedandepoxyisusedasafillertopermanentlyaffixthe

17
Hill,D.ReturnonInvestmentforStructuralHealthMonitoringSystemsinWindTurbineBlades.
DNVReportNo.20109509,December,2010.
39
unitsinplace,suchaswhatwasdemonstratedbyVirginiaTech.Approximately2(5cm)of
asphaltisoverlaid.Sawcutchannelstorunelectricconnectionsarerequiredtoroutewiringto
theroadside,andthesemustalsobeburied.Becausetheasphaltlayerandroadbedarenot
perfectlyrigid,someenergyisabsorbedbythesurroundinglayersofmaterialandthereforea
largerunitwithforcemultiplyingcomponentsisrequiredtoharvestsignificantenergy.This
affectsthecostofinstallationandoperationbecausegreatercapital,time,andlaborisneededto
installtheunits,andtheunitsaremadewithgreatermaterialvolumestomakethemrobustfor
theharshconditionsintheroadway.Asaresult,greaterlaborcostsarerequiredforinstallation,
andgreaterdifficultiesareencounteredwithmaintenance.Alsointhisconfiguration,aless
efficienttransformerisusedwhichmayleadto30percentlosses(70percentpassthrough
efficiency)inconversionofthepowersignaltousablepower.
18
AstudyfundedbytheGerman
FederalHighwayResearchInstituteexaminesthesepropertiesindetail.

Figure 22: Cros-Sectional Diagram of the Roadway Installation of Piezoelectric Energy Harvesters

Source: DNV KEMA

Therailwayapplicationimpliestheuseofathinnerunitfortworeasons:(1)thegeometryofthe
installationrequiresathinunit,and(2)therearelessinelasticforcesinactioninthisapplication
andfewerdiscsareneededperunittoharvestusefulenergy.Thereareanumberofcostsaving
opportunitiesinthisinstallation.Theunitisthinner,soitrequiresfewerpiezoelectricdiscs,

18
Milgrom,Charles.Innowattech.Phoneconversation,1/24/2012.
40
thusloweringitscapitalcosts.Theunitcanbeinstalledbetweentherailtieandthesteelrail
anddoesnotrequireexcavation,whichlowersthelaborrequiredforinstallation(andlowers
installationcost)asshowninFigure24.Inaddition,becausetherailisrigid,itislesselasticthan
aroadwayandthereforeimpartsmoreforceperunitareaonthepiezoelectricunit,which
improvesefficiency.Forthatreason,theseunitshavehigherenergydensity.Thepassingrail
carsimpartmoredirectenergytotheunitsandconsequentlythereislessdependenceonspeed.
Becausetherailcarsareonrails,thereis100percentcertaintythatapassingcarwillcontactthe
piezoelectricdeviceandasaresult,theefficiency(energy/compressionevent)isgreaterthana
roadway.Lastly,becausetherailsystemisusuallyprivatelyownedandthereisdirectaccessto
nearbypower,permittingproceduresarelessintensive,andconnectiontonearbypower
sourcesismoreaccessible.Theunitsareeasilyaccessibleformaintenance.Accordingto
Innowattech,convertingthesignaltoauseablepowersignalis>90percentefficientbecausea
moreefficienttransformerdesignispossibleinthisconfiguration.

Figure 23: Above-ground Installation of Piezoelectric Energy Harvesting Devices for Railways

Source: DNV KEMA

41
CHAPTER 3:
Conclusions and Recommendations
Asaresultofthisanalysis,anumberofconclusionsandrecommendationscanbemade.
ThisanalysisusesvendordatatoestimatetherangeoftheLCOEforthepiezoelectric
systemtobebetween$0.08$0.18/kWh.Thisvalueisstronglydependentontraffic
conditionsandvehiclecharacteristics.Usingbestestimationsfromvehicledata,the
DNVKEMAestimateisthatthemaximumfortheLCOEiscloserto$0.20/kWh.
Atrafficmodelusedapproximationstoderivetrafficcharacteristicsinordertocalculate
capacityfactorandvehicleweightsforUnitedStatesroadways.Thisdatawasthenused
inconjunctionwithknowndataaboutpiezoelectricdemonstrationsinroadwaysto
assesskeyparameterssuchastheLCOEandcapitalcost.Vendorclaimshavebeen
foundtobemutuallyexclusive,likelyindicatingthepresentationofbestvaluesfrom
mutuallyexclusiveconditions.
Thirdpartyvalidationofpoweroutputpermodulewouldgreatlyreduceuncertaintyin
theseestimates.Untilthepoweroutputpermoduleistransparentlyquantified,with
specificconditionsunderwhichitcanbereplicated,costofenergyestimateswillcontain
inherentuncertainty.Atthispointintimewiththeinformationavailable,itwould
appearthatpowerdensitiesof300W/ft
2
ormoreareneededtoapproachtheeconomic
viabilityclaimedbyvendors.
Thelifetimeofthesystemneedstobebetterquantifiedviademonstration.Present
demonstrationislimitedtotwoyears.Acceleratedtestscanevaluatelifetimeinamore
costeffectivemannerthananactualdemonstration.
3.1 Stage-Gate Evaluation
Ademonstrationandthoroughevaluationofthetechnologyshouldattempttoquantifythe
poweroutput,durability,andlifetimeofthesysteminadditiontoitsperformanceasafunction
oftrafficvolume.Detailsofhowandwhytheseevaluationsshouldoccurareprovidedinthe
Appendix(AppendixD:EvaluationCriteria).
ItisrecommendedthatanyresearchintheareabestagedwithGo/NoGogatessuchthatrisk
fortheprojectfundsismitigated.
Presentlytherearefourpotentialproductsforevaluation:
1) InnowattechRoadwayharvester
2) InnowattechRailwayharvester
3) POWERleaproadwayharvester
4) Genzikoroadwayharvester
42
Itisrecommendedthatifanevaluationpathisdesired,eachoftheproductsbeevaluated
againstoneanotherinanobjective,labscaleevaluationfirst.Ineachphaseofresearch,the
minimuminvestmentrequiredtoanswerfundamentalquestionsaboutpoweroutputand
lifetimeshouldbeconsidered.Forexample,inPhaseI,suchtestingcanbeperformedina
modifiedstressframewithduplicateunitsfromeachmanufacturer.InPhaseII,theaccelerated
testscanbeperformedinasimilarlabenvironmentwithminimizedcostsinordertoverifythat
promisingtechnologiesarealsodurableintheirexpectedenvironment.OnlyinPhaseIIIwould
afieldtestbeconsidered,andeveninthiscaseadownselectedproductgroupwouldbe
consideredandmayonlybeisolatedtoasinglelocationforverificationpurposes.The
suggestedscopeisshowninTable13.Furtherdetailsaboutthetestingphasesareincludedin
Table13.

Table 13: A Staged Gate Approach to Evaluate Various Piezoelectric Technologies

Phase Description Expected Result/Outcome Pass/Fail Criteria
Validation of Power
Output per module
Tentatively, we calculate a
power output of 300 W/ft
2
is
required to make the system
viable. If power output is
promising or if any vendor
claims are verified, proceed
to Phase II. Determine top
performer, select pathway
for implementation (road or
rail)
Using calculation approaches
in this report, verify that power
output matches the needed
levels for payback to reach the
targeted power densities or
power per km metrics. If it
does not, it shall not proceed
to Phase II.
Accelerated Tests Identify decay mechanisms
and durability issues.
Reduced list of products
from Phase I will be tested.
If durability and failure
modes are acceptable,
proceed to Phase II.
For products that have made it
to Phase II, they shall show a
cycle life equivalent to critical
lifetime, such as 10-20 years.
Should account for weathering
and other abuse factors.
Field Demonstration For durable products that
have shown acceptable
power output, a field
demonstration in an
appropriate environment
should be chosen.
Actual use data should verify
the needed power output and
durability requirements.
Source: DNV KEMA

3.1.1 Phase I: Lab Scale Tests

Loadingcyclingduringthesetestsmaybeconsideredprovidedthatthecontrolvariablesare
complimentarytothefieldtests.Loadingandcyclingtoverifypoweroutputshouldbedonein
acontrolledfashion.Loadshouldbecalculatedbasedonsimulatedvehicle(ortrain)loads.
Poweroutputshouldbemeasuredandpresentedintheformofwattspercyclesuchthatthis
43
datacanbetranslatedtoroadwayorrailwayperformance.Effectofsubstratelayers(asphalt,
concrete)shouldbeconfirmedeitherexperimentallyorwithfiniteelementanalysis(FEA)
models.StressframetestsandFEAmodelingexamplesareprovidedinFigure32.
Fromtheaboveassessments,desiredpoweroutputshouldequatetovaluesthatgenerate
favorableLCOEestimatesascalculatedinthisreport.Productsthatdonotpassthesecriteria
wouldnotbeconsideredforfuturephasesofwork.
Considerationsoftheotherstandardsmentionedinthisdocument(MIL1376andASTMC627)
shouldalsobeconsideredwhileaccountingforuniquefeaturesatthesystemlevel.
3.1.2 Phase II: Accelerated Tests
Thecostofenergycalculationinvolvesanassumedproductionofenergyoverthelifeofthe
system.Presently,sincenodataisknownaboutthedecayrateofenergyproductionovertime,
itisassumedtobeconstant.Itisunlikelythatthisisthecase,sothisdecayrateshouldbe
quantified.
Tothisend,ademonstrationshoulduseacceleratedtestingtoidentifymaterialsdegradation
factors.Thesetestsshouldbedesignedtoidentifythedominantdecaymechanism,suchas
moisture,stress,fatigue,ortemperature.
Theloadingofthediscsandtheirexposuretotemperatureandmoisturemaybeofinterest.
Examiningthepoweroutputofthesystemascomparedagainstcommonageingtestssuchas
saltfog(ASTMB117)canbeconsidered.
Forexample,controlledloadingconditionsovertimewithoutweatheringormoisturecouldbe
comparedagainsttheacceleratedtestsinPhaseIItodetermineifthecyclingduetoloadingis
themaindegradationfactor,oriftheaddedeffectsofweatheringcausethedegradation.
Acceleratedtestswouldneedtocapturethedegradationfactorsthataremostharmfultothe
systemlife.AsuggestedschematicofsuchadegradationsystemisshowninFigure24.A
deviceusedtosimulatearoadtirewearingonthesurfacewithvariableloadcanbeusedto
simulatethepassingweightandwearofvehicles.Loadcanbevariedtosimulatevariancesin
vehicleweight.Therepeatedactionofrollingthewheelacrossthesurfacewillacceleratethe
numberofcyclesontheinstallationanddecreasethetimeneededtoobservedegradation
indicators.Ifpossible,weatheringcanbeaddedbyintroducingheat,cold,rain,orUVtothe
systemtosimulateageingconditionsbyweather.Poweroutputandsystemperformance
shouldbemonitoredasafunctionofcontrolparametersandcycles.Amodificationtothe
designinFigure23:AbovegroundInstallationofPiezoelectricEnergyHarvestingDevicesfor
Railwayscouldbemadeforrailapplications,byusingasectionofrail,atie,andaloadedwheel
ontherail.

44
Figure 24: Schematic of an Accelerated Ageing Platform for a Piezoelectric Energy Harvester
Application

Source: DNV KEMA

Thesystemshowninthefigurehasbeendesignedinvariousiterations,asmentionedina
NorwegiantestprogramdevelopedintheNorwegianMasonryandConcreteResearchInstitute
inNorway(seeFigure25).
Carefulconsiderationoftheefficientuseoftimeaswellasadesignofexperiments(DOE)based
approachtoisolateeachofthevariablesandquantifytheireffectwouldberequired.Itshould
bepossibletoextracttheexpectedcyclelifeandtesttothatparameterintheacceleratedtests.
Forexample,iftrafficpatternsindicateanaverageof4,000vehiclesperday,itmightbe
expectedthatthenumberofhitsorcycleswouldbe4,000vehiclesperdaymultipliedby
twotirespervehicle,multipliedby365daysperyear,multipliedby20years,or58million
cycles.Inthisexample,oneyearrepresentsabouttwomillioncycles.Weatheringpatternscould
becontrolledaccordingtocyclelengthtosimulateanacceleratedseasonalwearpatternforthe
system.

Figure 25: Efficient Road-Testing Jig Designed to Cycle Concrete under Road Wear Conditions

Source: Norwegian Masonry and Concrete Research Institute

45
3.1.3 Phase III: Field Tests
Fieldtestingcanbeusedtoquantifyinstallationcostsandinformationaboutrealworld
performance(suchasvehiclevolumeandadditionaldataproduced).Thecombinationof
acceleratedtestsandfieldtestswouldthenplaceboundsonthelifetimeanddurability
uncertaintiesidentifiedinthisassessment.Assumingproductshavemadeitthroughthefirst
evaluationphases,thefieldtestwouldbethefinalconfirmationthattheproductfunctionsas
intendedandhaspotentialtomeettargetLCOEandcapitalcostmetrics.
Asotherdemonstrationshaveshown,afieldtestshouldinvolveactualinstallationofthe
systeminanactualroadway.Thepurposeofthesetestsistocompareactualperformancetolab
andacceleratedtestperformanceandisolatevariablesassociatedwithrealworldapplication,
suchasinstallationoftheelectronics,associateddifficultieswiththeroadsurface,andother
wearfactors.Unlikethepriorinstallations,itshouldbefeasibletoexaminetheperformanceof
thesysteminasphaltandconcreteforcomparison.TheVirginiaTechfielddemonstrationis
showninFigure26.

Figure 26: Installation of Field Demonstration in Virginia Using Innowattech Energy Harvesters
(left) and Schematic of Installation with Data Collection (right)

Source: Virginia Tech

46
3.1.4 Suggested Test Structure

Table 14: Outline of Suggested Demonstration Project

Test Phase or Task Purpose How the data should be
used
Phase I: Lab Tests Isolate power output as well as
controlled tests to verify
performance as a function of load
or frequency
Make first data-validated revision
to LCOE calculation.
These tests should determine
whether a product passes or
fails performance and merits
investigation in further Tasks.
If LCOE is greater than an
unacceptable value (that is,
$0.20/kWh), the product does
not proceed to next phase.

Phase II: Accelerated Testing Quantify ageing factors, energy
output decay rate, failure rates.
Identify dominant variables
associated with decay factors to
add confidence to findings from
accelerated tests,
Overlay energy output
degradation on the cost of
energy assessment and
determine how the economics
are affected.
Support and inform the design
of the field test and will be
used to isolate influencing
factors in the accelerated
tests. Products that do not
show needed durability or
cycle life shall not proceed to
Phase III.
Phase III: Field Test Show system performance in real
world, isolate real world factors
that may influence system
performance
Confirm field factors that
affect system performance
that are not seen in the
accelerated or field tests.
Phase IV (continuous throughout
project): Revised LCOE and
Evaluation
Use degradation factors and
energy output data from first 3
tasks to reassess cost of energy.
Refine the calculations
presented in this report and
give increased confidence to
the assumptions used in the
assessment of piezoelectric
energy harvesting.
Source: DNV KEMA

3.1.5 Approximate Cost of a Demonstration Project

TheVirginiaTechprojectappearstobea$600,000$1,000,000projectexecutedoveratwoto
threeyeartimeframe.Theprojectinvolvessomesystemmodificationanddesignaswellas
installationofthefieldtest.
47
Thedemonstrationseemstolackcriticalassessmentssuchasmultiproductcomparison,lab
controlledpoweroutputmeasurements,acceleratedtestingandafinalizedcostofenergy
assessmentbasedonthetestfindings.Amoreextensivetestingprogramisneededtoaddthese
dimensions.ThefieldtestfortheVirginiaTechprojectisnotlongenoughtodeterminecritical
ageingfactors.
Inastagedapproach,investmentcanbeminimizedbymeansofshorttermcommitmentsto
eachresearchphase.Alaboratoryinvestigationofmultipleproductscouldbeaccomplishedfor
anestimatedcostof$50,000$100,000.Ifnoproductspassthisstage,theprojectwouldbe
terminated.
Ifproductsmakeittothesecondphase,thecomplexityofacceleratedtestingwouldaddtothe
costbuttestscouldberunforarangeof$100,000$200,000.Ifnoproductspassthisphase,the
projectwouldbeterminated.
Thefielddemonstrationwouldlikelyrequireaninstallationsiteandcontinuedmonitoringand
processingofdata.Thiswouldbethemostexpensiveportionoftheprojectandmaycost
$300,000$600,000.Ateachstage,therevisedLCOEwouldbeprovidedbasedonfindings.
Thetotalprojectcostwouldrangefrom$450,000$900,000,whichwouldcostlessthanthe
VirginiaTechdemonstrationwithmorevalueaddeddataoverathreeyearprojectduration.
Thevaluefromsuchademonstrationwoulddetermineifpiezoelectricmaterialsareasuitable
technologytogeneratecleanenergyfromroadwaysandimprovetheoverallefficiencyof
transportationonhighways.
3.1.6 Potential Partners and Functions
PotentialpartnersfortestingcanincludeUniversityofCalifornia,DavisandCaltrans.Caltrans
hasthecapabilityofofferingfieldtestsitesandtestbeds.Caltranscouldbeausefulpartnerina
fielddemonstrationphase.
Thelaboratoryfacilities,suchastheDNVofficeinColumbus,Ohio,arecapableofperforming
labscaleandacceleratedmaterialstests.IfatestingpartnerinCaliforniaisdesired,the
laboratorywillbeabletospecifytheequipmentandexpertiseneededandworkwitha
subcontractedpartnertocompletethesetests.
Intheeventofarailwaydemonstration,Caltransmaystillbeabletooffertestingfunctionsor
capabilitytothecommonfactorsbetweenrailandroadway,buttheremaybeaneedtoidentify
awillingrailpartnerforguidanceonsometestsand/orfieldapplication.
UCDavishasbothlaboratoryequipmentassociatedwithvehicletestingaswellas
demonstrationcapabilities.AfacilitysuchasUCDavismaybecapableofsupportingan
acceleratedtestjigsuchastheonedescribedinFigure25viaadaptationofexistingequipment
orconstructionofnewequipment.
ForthepurposeoftheLCOEevaluation,proceduressuchasthoseshowninthisreportcanbe
used.DNVisonesuchpartnercapableofmanagingsuchaproject,providingtechnical
48
leadershiptowardthetestgoals,andaggregatingthedataforthepurposeoftechnology
qualificationandevaluation.
49
GLOSSARY
Abbreviation/Acronym Definition
AC Alternating Current
ASTM American Society for Testing and Materials
BOS Balance of System
CHP Combined Heat and Power
DC Direct Current
DNV KEMA DNV KEMA Energy and Sustainability
DOE Department of Energy
DOT Department of Transportation
EIA Energy Information Administration
FEA Finite Element Analysis
FHWA Federal Highway Administration
ft Foot
Ge Germanium
Hz Hertz
IEA International Energy Agency
J J oules
km Kilometer
kph Kilometer per Hour
kW Kilowatt
kWh Kilowatt Hour
Lbs Pounds
LCOE Levelized Cost of Energy
mA Milliampere
MHz Mega Hertz
MEMS micro-electromechanical systems
MJ Microjoules
Mm Millimeter
MPa Megapascal
mph Miles per Hour
MW Megawatt
50
Abbreviation/Acronym Definition
MWh Megawatt-Hours
NPV Net Present Value
NREL National Renewable Energy Laboratory
ODOT Oregon Department of Transportation
psi Pounds per square inch
PV Photovoltaic
R&D Research and Development
RMS Root Mean Square
ROI Return on Investment
SHM Structural Health Monitoring
Si Silicon
V Voltage
VMT Vehicle Miles Travelled
W Watts
Wh Watt-hour

51
REFERENCES
Ali,etal.AnalysisofEnergyHarvestersforHighwayBridges.JournalofIntelligentMaterial
SystemsandStructures22(16)19291938.2011.
Anton.MultifunctionalPiezoelectricEnergyHarvestingConcepts.DoctorofPhilosophy
Thesis,VirginiaPolytechnicInstituteandStateUniversity,April2011.
Cattermole,Tannith.MobilePhonesChargedbythePowerofSpeech.September20,2010.
Gizmag.http://www.gizmag.com/mobilespoweredbyconversation/16417/Accessed
12/12/2012.
CookChennault.PiezoelectricEnergyHarvesting:AGreenandCleanAlternativefor
SustainedPowerProduction.BulletinofScience,Technology,&Society,Vol28,No6
Dec2008pp496509.
Davis,Stacy;Diegel,Susan;Boundy,Robert.TransportationEnergyDataBookEd.31.July
2012,OakRidgeNationalLabsORNL6987.
EderyAzulay,Lucy.Innowattech:HarvestingEnergyandData;AStandaloneTechnology.
IsraelNationalRoadsCompany,Ltd.FirstInternationalSymposium.TheHighwayto
Innovation,13November2010.TelAviv,Israel.
EvaluationofEnergyHarvestingTechnologyforUseonOregonRoadways.FY20145
ResearchProblemStatement.ODOTResearchSection,SalemOR.
Fourie,D.ShoeMountedPVDFPiezoelectricTransducerforEnergyHarvesting.MIT.2009.
GenzikoRPGproductbrochure,technicaldata.2012.
Guigon,etal.HarvestingRaindropEnergy:Theory.SmartMater.Struct.17(2008)015038
(8pp).
Hemsel,T.,Wallaschek,J.SurveyofthePresentStateoftheArtofPiezoelectricLinearMotors.
Ultrasonics,38(2000)3740.
Hill,D.ReturnonInvestmentforStructuralHealthMonitoringSystemsinWindTurbine
Blades.DNVReportNo.20109509,December,2010.
Horowitz,etal.AMEMSAcousticEnergyHarvester.J.Micromech.Microeng16(2006)S174
S181.
Innowattech:HarvestingMechanicalEnergyfromRoadwaystoProduceElectricity.Kurzweil,
AcceleratingIntelligence.October25,2011.http://www.kurzweilai.net/innowattech
harvestsmechanicalenergyfromroadwaysAccessed12/10/2012
InnowattechWebsite.www.innowattech.com.Accessed1/3/2013.
Kuntsal,Ender.ChannelTechnologies.Verbaldiscussion1/14/2012.
52
IsraeliEnergyStartupTurnsTrafficintoSourceofElectricity.Reuters,Mar102009.Haaretz
online.http://www.haaretz.com/news/israelienergystartupturnstrafficintosource
ofelectricity1.271824.
Lazaru,Jon.OregonDepartmentofTransportation.Phoneconversation1/17/2013.
Li,etal.DevelopmentofaBiomechanicalEnergyHarvester.JournalofNeuroEngineering
andRehabilitation20096:22.DOI:10.1186/17430003622.
Lim,etal.DesignandPerformanceofMiniaturizedPiezoelectricStepDownTransformer.
JournalofElectroceramics.Vol13,Issue13July2004.
Milgrom,Charles.Innowattech.Phoneconversation,1/24/2012.
MILSTD1376B,MilitaryStandard:PiezoelectricCeramicMaterialandMeasurements
GuidelinesforSonarTransducers.February24,1995.[NOS/SDOCUMENT].
Moskowitz,Ira.IsraelsInnowattechtoProvideRenewableEnergyforHighwaySignsin
Italy.http://www.greenprophet.com/2010/05/israel%E2%80%99sinnowattechto
providerenewableenergyforhighwaysignsinitalyAssessedon1/29/2013.
OpenEI:OpenEnergyInfo.http://en.openei.org/wiki/Transparent_Cost_Database.Accessed
January1,2012.NationalRenewableEnergyLaboratory(NREL),OpenGovernment
Initiative,USDepartmentofEnergy.
POWERleapTechnologiesproductbrochure,technicaldata.2012.
Priya,Shashank.AdvancesinEnergyHarvestingUsingLowProfilePiezoelectricTransducers.
J.Electroceram(2007)19:165182.
Sathiyamoorthy,Bharathi.HybridEnergyHarvestingusingPiezoelectricMaterials,Automatic
RotationalSolarPanel,VerticalAxisWindTurbine.ProcediaEngineering38(2012)843
852.
Sodano,etal.AReviewofPowerHarvestingfromVibrationusingPiezoElectricMaterials.
ShockandVibrationDigestVol36,No.3May2004.
Sodano,etal.EstimationofElectricChargeoutputforPiezoelectricEnergyHarvesting.LA
UR042449,StrainJournal,40(2),4958,2004.
TreevoltPiezoelectricMembraneSystem.http://www.treevolt.com/en/treevoltpiezoelectric
membranesystem.htmlAccessed12/13/2012.
Tveter,AcceleratedWearTestingofHighStrengthRoadConcreteinNorway.Norwegian
MasonryandConcreteResearchInstitute.1993.
Uchino.PiezoelectricActuators2006:ExpansionfromIT/roboticstoecologicalandenergy
applications.JElectroceram(2008)20:301311.
Verbaldiscussions,ChannelTechnologies.
53
54
Vestel,LeoraBroydo.HarvestingCleanEnergyAlongtheRoad.NewYorkTimes,July23,
2009.http://green.blogs.nytimes.com/2009/07/23/harvestingcleanenergyalongthe
road/Accessed12/10/12.
Vullers,etal.MicropowerEnergyHarvesting.SolidStateElectronics53(2009)684693.
Walsh,etal.PiezoPower.DigitalSafariGreenbizzCompanyBusinessPlanCompetition,
2011.
Willoughby,Kim.PostConstruction&PerformanceReport,ExperimentalFeaturesWA0304,
0401,and0504,StuddedTireWearResistanceofPCCPavements,Contract6620I90
ArgonneRoadtoSullivanRoadMP286.91to292.38.WashingtonStateDepartmentof
Transportation.2007.
Wischke,etal.VibrationHarvestinginTrafficTunnelstoPowerWirelessSensorNodes.
SmartMater.Struct.20(2011)085014(8pp).
Xiong,etal.NewTechnologiesforDevelopmentofRenewableEnergyinthePublicRightof
Way.DTFH6110C00016.FHWA9thQuarterlyReport,VirginiaTech.October2012.
Xu,etal.Preparation,Properties,andApplicationResearchofPiezoelectricTrafficSensors.
ProcediaEngineering27(2012)269280.
Waterbury,Wright.EvaluationoftheConsultantReport:AssessmentofPiezoelectricMaterials
forRoadwayEnergyHarvesting,CostofEnergyandDemonstrationRoadmap.
January31,2013.

APPENDIX A:
Piezoelectric Energy Harvesting Demonstrations
Innowattech
Innowattechisthemostnotablecompanytoimplementpiezoelectricbasedenergyharvesting
onabulk,macroscale.
http://www.innowattech.co.il/
ThesystemisinstalledbyInnowattechinIsrael.Thepiezoharvestersareimbedded5cmbelow
thesurfaceoftheroad.Itisprojectedthatincreasingthesystemsizeto1kmwouldproduce200
kWhwhileafourlanehighwaywouldproduceaboutoneMWh.TrafficstudiesfortheAyalon
Highway,coastalhighway,andTransIsraelHighwayexaminedtheenergypotential.The
Israelitestwasconductedin2009.
InnowattechwasselectedbyImpregiloSpA,anItalianinfrastructureandcivilengineering
contractorsenergyproviderforlightingroadsignsontheVeniceTriestehighwayinItaly.This
contractispartofa225millionupgradeofthehighwaythatbeganin2010andisexpectedto
becompletedin2013.ThegeneratorsdevelopedatInnowattechwillbeplacedbeneaththe
highwaysupperasphaltlayer.Theelectricalenergygeneratedbythetechnologyiscreated
duringthemovementofvehiclesontheroadandisstoredviadedicatedelectricalsystems.This
willsupplyelectricalenergyforlightingVariableMessageSystemsigns.Driverswillread
trafficreportsonelectronicsignswhichwillbepoweredbyelectricityfromthedriversown
vehicles
19
.
PiezoPower, Piezoelectric Floors
LondonClubSuryaandRotterdamWatthavepiezofloorstoharvestenergyfromdancers.The
floorsystemisengineeredwithspringsandaseriesofcrystalandceramicblocks.Intheclubs,
thiscansupplyupto60percentoftheclubsenergyneeds.Eachpersoncanproducebetween5
20W.TheEastJapanRailwayCompanyworkedinconjunctionwithKeioUniversitytoimbed
piezointhefloorofterminalsandtrainstations.Thisisalsoanopportunityforhealthand
fitnessclubs.DigitalSafariGreenbizzCompanyisaimingtocapitalizeonthetechnologyby
buildingpiezoelectricfloorsandquotesTimeMagazinebyindicating1wattperbreath,70
wattspersteparepossible.TheproductiscalledElectroturf.PiezoPoweristhecompanythat
sellstheproduct,estimating1500ft
2
for$2250,or$1.50/ft
2
.Theproductisdesignedas
subflooringin3x5tiles.Itisestimatedthatabout25percentofthe70Winasinglestepis
captured(17.5W).ThepiezoelectricmaterialisRochellesalt.ThePiezoPowerbusinessplan
indicatesthatRochellesaltcosts$1permetrictonandissourcedfromPinhuangdaoBright
ChemicalCompany.

19
http://www.greenprophet.com/2010/05/israel%E2%80%99sinnowattechtoproviderenewableenergy
forhighwaysignsinitaly/.Assessedon1/29/2013.
A1
Figure 27: Piezo Floors

POWERleap Inc.
POWERleapInc.islocatedinSanFrancisco,California,andisinapartnershipwithTreevolt
locatedinColumbia.Forasphaltroadways,theharvestersaremembranebaseddesigns
containingrecycledbutylpropenewherefullsheetsofharvestingdevicesarefullyembedded
underthetoplayerofasphalt.Forconcrete,theharvestersareablockbaseddesignthatis
partiallyembeddedintotheroadwaysuchthatpressurefrompassingvehiclesisdirectly
transferredtothedevices.Themanufacturersliteratureonthedevicesindicatesthata1.0km
lengthofroadwaywith600vehiclesperhourand6,000Treevoltharvesterscanyield
approximately720kW.Thedevicescanalsobeequippedwithdatamonitoringcapabilitiesthat
cancollectroadwayandvehicledata.Thedevicescantransmitthedatathroughwireless
communicationstoacloudbasedplatformsothatthedatacanbeaccessedatanytimefrom
anyinternetenableddevice
20
.FloorbasedpowergenerationfromPOWERleapproductshave
shown10Wgeneratedfromafootstepwitha2.2voltpotentialacrossthefunctionalmaterial.
Thisimpliesapproximately4.5Ageneratedfromtheunit.

Figure 28: Treevolt Devices being Tested in Columbia (left) and POWERleap Data Harvesting
(right)

20
POWERleapTechnologiesproductbrochure,technicaldata.2012.
A2
KCF Technologies
Offtheshelfproductsaredesignedtosenseandharvestenergyfromvibrations,suchasthose
availablefromKCFTechnologies.KCFisdevelopingvibrationenergyharvestingdevicesto
powerwirelesssensornodes.Thedevicescavengesvibrationenergyfromahoststructure,
eliminatingbatteryreplacementforindustrialsensors.Theyhavealsodevelopedaself
poweredwirelesssensorkitthatsimultaneouslymonitorsvibrationlosseswhileusingthe
vibrationstopowerthedistributedsensornetwork.Freelyavailablevibrationenergyis
capturedandusedastheonlypowersource.
Oregon Department of Transportation
Inlate2008andearly2009,withthesuccessoftheOregonSolarhighway,theOregon
DepartmentofTransportation(ODOT)evaluatedharvestingenergyfromroadwayvibrations.
Vendorsclaimedtobeabletocaptureenergywithpiezoelectronicdevicesinstalledintothe
pavement.ODOTdidnotcommittoinstallingthedevicesbecausetherewasnoUnitedStates
basedvendoratthetime.Othervendorsofferedenergyharvestingfromacombinationofsolar
andspeedbumpdeviceswhichdepresswiththevehicleweight.ODOTelectedtopursuea
solutionwhichcouldbeinstalledandkeeptheroadsurfaceflatforhighwaytraffic
21
.
Sincethen,aUnitedStatescompanynamedPOWERleaphaspartneredwiththeColombia
basedcompanyTreevoltandenteredthemarket.In2012,OregonStateUniversitysubmitteda
newresearchproposaltotheODOTtostudypiezoelectricharvestersreliabilityand
maintenancerequirements
22
.Theapplicationwasrejectedinthefirstroundofevaluation
becausetheevaluationcommitteewasconcernedaboutthematurityoftheproduct,citing
FHWAsreportbyEricWeaverin2012.
Channel Technologies Group
ChannelTechnologiesGroup(SantaBarbara,California)manufacturestheceramicsthathave
beeninvestigatedbyInnowattechasaUSsupplierformaterials.ChannelTechnologieshada
lowvolumesupplyagreementwithInnowattechtoinvestigatetheirmaterialasthefunctional
elementfortheInnowattechdevices.Presently,InnowattechdoesnotuseChannelTechnologies
fortheirmainproductordevelopment.Ithasbeensaidthatthepiezoelectricdisccostshouldbe
targetedatabout$1/disc
23
.Ithasalsobeenstatedthatthelifetimeofthepiezoelectricmaterialis
expectedtobe30years,butthelifetimeoftheancillarypiecesoftheenergyharvestingdevice
mayormaynotreachthelifetimeofthediscitself.OneofthechallengesthatChannel

21
Lazarus,Jon.OregonDepartmentofTransportation.Phoneconversation1/17/2013.
22
Gambatese,John.FY2014ResearchProblemStatement:14047EvaluationofEnergyHarvesting
TechnologyforUseonOregonRoadways.OregonDepartmentofTransportation.
http://www.oregon.gov/ODOT/TD/TP_RES/docs/stage1/2014stageone/14_047_Energy_Harvesting.pdf.
Accessed12/11/12.
23
Verbaldiscussions,ChannelTechnologies.
A3
TechnologiesfacedearlyonindiscussionswithInnowattechwasmeetingthestrengthand
durabilityrequirementsfortheroadwayorrailway.
Virginia Tech
VirginiaTechismanagingathreeyearprojectwithacontractamountof$1million.
24
TheUS
DOTfundedprojectisinvestigatingtheuseofpiezoelectricmaterialsforroadwayenergy
harvesting.Theprojectjustcompletedits9
th
quarter.Presentlythedataindicatesthatthetotal
systemvoltagerangesfrom400700Vand0.20.35mAperunit,withpowerpulsesof0.10.2
seconds.Testedinheavytrucktrafficatabout40mph,thedemonstrationtargetsa4,000vehicle
perday(166vehicles/hr)trafficflowrate
25
.
ItisshownintheVirginiaTechreportsthattheharvesterdevelopedwasoftheirowndesign.
Thepoweroutputperunitislowcomparedtovendorclaims,andthelabbaseddesignmay
accountforthelowpoweroutput.Commercialdesignsmayincorporatelinkagesfor
mechanicaladvantagetoincreaseforceonthepiezoelectricstacksandimprovepoweroutput,
andthereforeitmaynotbefairtosaythattheVirginiaTechunitisrepresentativeof
commercialdesign.

Table 15: Summary of Known Demonstrations and Their Approximate Cost

Demonstration Estimated Project Cost
Innowattech, Israel $650,000/km
Virginia Tech, US FHWA $1 million

Genziko
26

VendorinformationfromGenzikoestimatesinstallationcostsat$0.40/WandLCOEof$0.06
$0.08/kWh.Theseestimatesareconsiderablymoreoptimisticthanwhathasbeenimpliedin
discussionswithothervendors.Thelifetimeestimateof20yearsisconsistentwithwhatis
expected,howeverthereappeartobenosignificantdifferencesinthesystemconstructionas
comparedtoothertechnologiesproposedhere,Thereforethesamesensitivityfactorswould
apply:1)claimedpowergenerated,2)lifetime,3)capitalcost.Thesensitivitytothefirstfactor,
however,islikelymuchgreaterduetothereasonsexplainedbelow.

24
Xiong,etal.NewTechnologiesforDevelopmentofRenewableEnergyinthePublicRightofWay.
DTFH6110C00016.FHWA9thQuarterlyReport,VirginiaTech.October2012.
25
Xiong,etal.NewTechnologiesforDevelopmentofRenewableEnergyinthePublicRightofWay.
DTFH6110C00016.FHWA9thQuarterlyReport,VirginiaTech.October2012.
26
GenzikoRPGproductbrochure,technicaldata.2012.
A4
ThelowLCOEquotedbyGenzikoislikelyattributedtohighpowerdensityclaimsanda
claimedlowercostthancompetingsources.Ona6lanefreewaywith2,250vehiclesperhour
perlane(about54,000vehiclesperday),theyclaimapoweroutputof51MWperkm.Theyalso
provideanestimateofcapacityfactornear3242percent.Sincethecapacityfactornumberis
contingentonroadtraffic,itwouldimplythattherearepowergenerationeventsoccurring
approximatelyevery3seconds,or20eventsperminute,or1200eventsperhour.Sinceevery
vehiclerepresentstwoevents(twoaxlespervehicle),thiswouldcorrespondto600vehiclesper
hourwhichisconsistentwiththelowestimateforcollectionroadsor2laneroadsinthe
Genzikomarketingdocuments.
Takingthiscalculationfurther,the600vehicleperhourestimatecorrespondstoa13.6MW/km
powerdensity.Assuming6000unitsperkm
27
,eachunitiscontributingacontinuouspower
outputof0.002MW(2kW).RecallthattheVirginiaTechdemorecordedabout0.080.14Wfor
everywheelimpactevent,orabout10,000timesless.Inorderforthisdisparitytobepossible,
somedifferentiatingcharacteristicoftheGenzikoproductmustharvestenergymoreefficiently
thanthecompetingtechnologies.
ThedifferentiatingcharacteristicoftheGenzikoproductappearstobeaclaimthattheunit
continuestogenerateelectricityaftertheinitialwheelimpact,indicatingaresonantor
persistentlyvibratingenergyharvestingtechnologythatisunlikethesingleimpulse
compressiveenergygenerationsystemdesignedbyInnowattech(Figure12).IntheGenziko
product,theinitialimpactgeneratesenergybutalastingresonanceordecayingvibration
continuestogenerate.Suchtechnologiesmaybebaseduponarraysofmicrocantileverbeams
similartowhatisusedinmicroelectromechanicalsystems(MEMS)andstructuralhealth
monitoring(SHM)systems.
ComparedtothedatafromODOTandthepressreleasesfromInnowattech,theGenzikotraffic
flowrate(2250vehiclesperhour)isabout313timesgreater.Howeverthepowerclaim(51
MW)is70500timesgreaterthanwhatisclaimedbyODOTandInnowattech,whicharemore
optimisticthanwhathasbeendemonstratedbyVirginiaTech.
Theinformationprovideddoesnotprovidemuchtechnicaldetailaboutthefunctional
piezoelectricorceramicmaterialsthatconvertvibrationorstressintoenergy,soitisdifficultto
assesstheexactintellectualpropertyortechnologyadvantagethatwouldleadtothese
performancemetrics,butitislikelyvibrationbased.
Thedifferenceinpowermetricsisillustratedinthecomparisontablebelow.

27
BasedonthePOWERLeap/ODOTestimates.
A5
Table 16: Power Metrics
Parameter Genziko ODOT Innowattech
Berkeley and
Virginia Tech
Power per km
(single lane)
13-51 MW 486 kW 100-200 kW 0.0018-0.5 kW
Vehicles per hour
(single lane)
600-2250 600 600 600
kW per km per
vehicle per hour
21.6-22.6 0.81 0.16-0.3 .000003
0.00083

Genzikohasnotprovidedinformationaboutrailways.
TheinterconnectbetweentheGenzikopowersystemandthegridimpliesagridtiedinverter
systemwhichislikelylowercostthanabatterybasedenergystoragesystem,thoughtitstillhas
provisionsforenergystorage.Thematurityofthesetechnologiesmaynotbeofftheshelf,as
indicatedbyTable17.

Table 17: Genziko product not Mature Material and Device Under Development
Technology Inventor(s) Date(s) Patent
Verification
Procedures
Material Near 1991-1999 To be
DoD-1376B(SH)
3

Nano Process
Dawson
1
/Swartz
1
/
Near
1988-1996
Patented +
Potential
DoD-1376B(SH)
3

Forming Kiggans
2
/Near 1998-2003 --
DoD-1376B(SH)
3

Processing Near 1987-1994 --
DoD-1376B(SH)
3

Configuration Near 1995-2009 Application
Actuation
4

Harvesting
5

RPG Near 1975-2011
6
To be --

A6
Figure 29: Genziko Estimates for Capital Costs Based on Storage Capacity

A7
APPENDIX B:
Literature Evaluation
1. Source:CookChennault.PiezoelectricEnergyHarvesting:AGreenandClean
AlternativeforSustainedPowerProduction.
SubjectMatterEvaluation:
Thepeerreviewedpaperconnectsthecapabilityofthetechnologywithstateand
federalinitiativestoproducecleanandrenewableenergy.Table3inthispaper
illustratesawiderangeofapplicablepiezoelectrictechnologiesthatcanbecross
referencedtotheroadwayapplicationinordertoverifythepowerdensity,energy
density,andperformancemetrics.Thedocumentalsodemonstratesdataconcerningthe
optimumharvestingofenergy.
Objectivity:Thisisapeerreviewedscientificarticleandisthereforeconsideredtobe
objective.
Persuasiveness:Thearticleispeerreviewedandnotintendedtobepersuasive,but
informative.
Value:Powerdensitymetrics,optimalvibrationfrequencies(between100120Hz),
dimensionsofpiezoelectricdevices,totalpowergenerated(asafunctionofsize),and
additionalenergymetricsareprovided.Thisispeerrevieweddatathatcanbe
immediatelycrossreferencedtothecommercialclaimsstatedinothersourcesforthe
purposeofvalidation.Thedocumentalsodescribescriticaldevelopmentneedsand
technicalchallengespreventingtheimmediateadoptionofpiezoelectricmaterialsfor
macroscaleenergyproduction.
2. Source:Xiong,etal.NewTechnologiesforDevelopmentofRenewableEnergyinthe
PublicRightofWay.
SubjectMatterEvaluation:
ThisreporttotheFederalHighwayAdministration(FHWA)concernsademonstration
ofpiezoelectricsinroadwaysinVirginia.Studyofthedurabilityofthematerialsis
included,andthestudymentionsInnowattech.Theharvesterappearstodegradein
performancewhenexposedtowaterandthedemonstrationmentionsdegradationofa
siliconelayerwhichalsoimpactstheeffectivenessofthedevicetoharvestmechanical
energy.Thereissomevaluabledatainthereport,suchastheuseofacontrolledtesting
apparatuswitha600lbloadandwheelspeedof7.5mphand15mphtogenerate
electricityandstorepowerinachargedcapacitor.Thecapacitorenergyisdirectly
proportionaltothevoltagegeneratedinthepiezoelectricdevice.Thereportconfirms
thatthereisaproportionalrelationshiptoloadtimeandpowergenerated.Thereport
confirmsademonstrationbeingperformedwitheightfielddevicesattheTroutville
weighstationoninterstateI81Ninthebypasslane.Thetrafficpatternismostlytrucks,
B1
estimatedinquantitiesof4,000perday,travelingat40mph.Patternswerecutoutto
installthedevices,indicatingthatatleastforthisdemonstration,removalofanentire
roadsectionwasnotrequired.Thedevicesmeasureapproximately1ft
2
andgenerate
approximately400700Vwitheachtruckimpactand0.20.35mA.Thetrucksgenerally
havefouraxlespassingoverthedevicesrepresentingshortpulsesofpowermeasuring
~0.1s.Therefore,theenergygeneratedis400V*0.2mA*0.1s=.008Whpertireimpact
perunit,ataminimum.Thepowerarrivesinpulsescorrespondingtowheelimpact
averaging.08Wormore.Thereportestimatesthatthemaximuminstantpoweris.14W.
Thecapacityfactorinthiscaseperunitoveran8secondintervalappearstobe~0.4/8=
0.05.
Objectivity:ThereporttotheFHWAiswrittenbythePrincipalInvestigatorandappears
tobeaphysicallydemonstratedevaluationofthetechnology.
Persuasiveness:Thearticleisastatementofaseriesoffactsandisthereforenotintended
tobepersuasive;however,thedataisdemonstrableandthereforevaluabletoan
objectiveassessmentofthetechnology.
Value:Durability,performance,load,traffic,andinstallationdataareallprovidedtoa
highlevelofdetailwithusefuldataandpowermetrics.
3. Source:EvaluationofEnergyHarvestingTechnologyforUseonOregonRoadways.
SubjectMatterEvaluation:
ThisinformationaldocumentexplainsstrategicresearchinitiativesguidedbytheODOT
andexplainstheapplicationofpiezoelectricsandalsonamessomecompaniesof
technologiesexplainedinthisreport,suchasPOWERleapInc.Thereisintriguingdata
inthesourcewhichcanbeusedforcrossvalidation:a1.0kmlengthofroadwitha
trafficflowof600vehiclesperhour(10perminute)with6,000Treevoltharvesterscan
yieldapproximately350,000kWhpermonth.
Objectivity:TheODOTevaluationisinherentlyintendedtobeobjectiveasitguides
investmentbythestate.However,theclaimsaboutthePOWERleapInc.technologyare
notclearlyidentifiedasthirdpartyvalidated.
Persuasiveness:Theexplanationofthetechnologyiscompellingandpersuasiveand
showspotential.
Value:ThedataconcerningPOWERleapInc.isvaluableandwillbeusedinthereportas
partoftheevaluation.ItalsodemonstratesthatODOThasintentionstobuilda1.0km
roadwaydemonstration.
4. Source:Ali,etal.AnalysisofEnergyHarvestersforHighwayBridges.
SubjectMatterEvaluation:
ThepaperisprimarilyconcernedwithStructuralHealthMonitoring(SHM)andusing
piezoelectricdevicestopowerthewirelesssensornetwork.ReferencestoSodanohave
B2
beenusedinotherDNVKEMAresearch.Thisappearstobeadefinitivereference.It
illustratesthatinabridge,theloadismovingwhichchangesthedeflectionand
resonancefrequencyoverthelengthofthebridge.Thepapercalculatesthe
displacementofthebeamandthefrequencyasafunctionofvehiclespeedandload.
Energyharvestingwithinductionbasedmechanismswasexploredinadditionto
piezoelectrics.Afterthederivationofthemathematicstoderivecriticalparameterssuch
asload,frequencyandoptimizedlocationoftheharvesters,anumericalexampleis
providedwhichcitessomeinterestingmetrics.First,atypicalvehiclespeedof25m/s
correspondstoafundamentalexcitationfrequencyfora25mbridgeof0.5Hz.
Harvesterlocationsareatintegralfractionsofbridgelength,forexample,L/2,L/3,and
L/4.Theenergyharvestedfora2.5gramenergyharvestingdeviceisupto18micro
joulespervibration.
Objectivity:TheauthorsarefromSwanseaUniversityandappeartobeobjectiveand
withoutcommercialbias.
Persuasiveness:Thepaperisnotintendedtobepersuasive,althoughithighlightsuseful
datawhichindicatestechnicalfeasibilityforusingvibrationbasedenergyharvestingto
powerSHMdevices.
Value:Thedescriptionofthemovingloadproblemisusefulbecauseitistransferableto
theproblemathand,forexample,imbeddingpiezoelectricmaterialsintoaroadway.
Themovingloadproblemhighlightsthatclaimedefficiencyforapiezoelectricpanel
deviceislargelydependentonitsorientationandpossibleaspectratioifharvesting
directvibrationalenergyfromaroadway.Thelowfrequencyvibrationsofthebridge
seemtobebestsuitedforaninductiontypeenergyharvesterandarelessrelevantfor
piezoelectrics.Thisisalsoarecentpaperwhichincludesdataformorerecent
technology.
5. Source:Priya,Shashank.AdvancesinEnergyHarvestingUsingLowProfile
PiezoelectricTransducers.
SubjectMatterEvaluation:
Thepaperexplainstheneedforharvestingenergyfromvibrationsandtheenergyneeds
ofvariouselectronicdevices.Itexplainshowvibrationenergycanpowerthesedevices
andhowtheycompareacrossanequivalentspectruminenergydensityandpower
densityterms.
Objectivity:Thepaperispeerreviewedandthereforerepresentsanevaluationwith
reasonableobjectivity.
Persuasiveness:Thepaperillustratestheviabilityofvibrationbasedenergyharvesting.
However,itdoesnotfullyaddresstheuseofvibrationenergyharvestingonamacro
scale.
B3
Value:Thereisalotofbackgroundinformationonenergydensityofenergystorageand
powerdevices,aswellasdataonenergygeneratedfromenergysourcessuchashuman
power,temperaturegradients,andpressurevibrations.Thepaperalsoillustratesthree
methodstoharvestenergyfromvibrations;piezoelectricmethodsisoneofthem.This
informationcanbeusedtocrossreferencetheefficiencyandenergygenerating
capabilitiesofpresenttechnologies,andmaybeusedasanobjectivereferenceto
validateotherwisecommercialclaims.
6. Source:Kurzweil.Innowattech:HarvestingMechanicalEnergyfromRoadwaysto
ProduceElectricity.
SubjectMatterEvaluation:
ThiswebarticlepresentstwocommercialvideosandanexplanationofInnowattechs
piezoelectrictechnology.Thevideosexplicitlyshowanimationsofvehicleson
highways.
Objectivity:ThearticleappearstobeapressreleaseforInnowattech,soitislikely
commerciallybiased.
Persuasiveness:Thetechnologyandtheclaimsarecompellingandappeartohave
technicalfeasibility,butverylittleactualdataisprovidedtoaidinthevalidationofthe
claims.
Value:Explanationofhowthetechnologyisproposedtoworkisusefulandhasdirect
relevancetothesubjectmatterofthisevaluation.
7. Source:TreevoltPiezoelectricMembraneSystem.
SubjectMatterEvaluation:
ThiswebsiteforthemanufacturerofthetechnologylicensedbyPOWERleapInc.shares
dataclaims.Claim:Theaverageenergygeneratedby1KmofinstalledPiezoelectric
membraneisintherangeof400600kWhforanestimated200to400vehiclesin16
hoursoftraffic.
Objectivity:Thisisavendorwebsitesothereisaconflictofinterestinthepersuasive
claims.
Persuasiveness:Theclaimsarepersuasive.
Value:Moredataisprovidedtocrossvalidatewithotherclaimsprovidedinthis
document.
8. Source:Walsh,etal.PiezoPower.
SubjectMatterEvaluation:
ThisdocumentisabusinessplanforacompanycalledPiezoPowerwhichintendsto
marketflooringmaterialstoharvestenergyfrompedestriansandpossiblymobility
B4
traffic.Thedocumentcontainsenergygenerationandcostdatafortheproductin
additiontoprojectionsofsalesoftheproductandgrowthofthecompany.
Objectivity:Thisdocumentisabusinessplanandthereforepresentsforwardlooking
statementsthatarenoteasilyverified.
Persuasiveness:Theaffordabilityoftheproductispersuasive,andiftranslatedtothe
roadwayproblem,presentsacompellingargumentforthecosteffectivenessof
piezoelectricenergyharvesting.
Value:Thebusinessplanmentionsthat70wattsperhumansteparegeneratedandthat
thesystemis25percentefficient,harvesting17wattsperstep.Costsareestimatedat
$2,250per1,500squarefeetofmaterial,using3x5tiles.Thesecostandefficiency
metricscanbecomparedtotheroadwayapplicationanddeterminetwothings:the
differenceincostbetweenthissimplerconstructionandroadwayconstruction,aswell
asvalidateclaimsaboutefficiencyandpowerdensity.
9. Source:EderyAzulay,Lucy.Innowattech:HarvestingEnergyandData;AStandalone
Technology.
SubjectMatterEvaluation:
ThepresentationisgivenbyInnowattechwithclaimsaboutperformanceand
demonstrationdata.
Objectivity:ThisisapresentationgivenbyInnowattechwhichisoneofthemostnotable
companiessellingpiezoelectricenergyharvestingtechnology.Thepresentationisaimed
atadvertisingfortheproduct.
Persuasiveness:Thedataispersuasive.
Value:Roadsapplicationclaim:200kWh/hfor600vehiclesperhourat72km/hspeed,
onekmsection.Railwayapplicationclaim:120kWh/hfortwogeneratorsinevery
sleeper,averagerailwaymovementof300loadedwagonsperhour.
10. Source:IsraeliEnergyStartupTurnsTrafficIntoSourceofElectricity.
SubjectMatterEvaluation:
Innowattech,anenergycompanyaffiliatedwithIsraelsTechnionInstituteof
Technology,saidspecialgeneratorsplacedunderroads,railwaysandrunwayscan
harvestenoughenergyfrompassingvehiclestomassproduceelectricity.
Article,pressreleaseforInnowattech.
Objectivity:ItisareviewarticlebutcontainsdataandinputfromInnowattechandthe
companiesinvolved,sotheremaybesomeconflictofinterestinthedata.
Persuasiveness:Articleispersuasivebutrevealshighcapitalcostforelectricity.
B5
B6
Value:UriAmit,chairmanofInnowattech,saidthecompanystechnologywillbethe
largestapplicationofpiezoelectricstodate,withasingle1km(halfmile)laneof
highwayprovidingupto100kWofelectricity,enoughtopowerabout40houses.
InnowattechchairmanAmitsaidthecurrentcostforfittingakilometer(halfmile)of
onelaneofhighwayisabout$650,000,withacostof$6,500perkilowatt.Hesaidwhen
massproductionbegins,thepricecoulddropbytwothirds,makingthesystemeven
cheaperthansolarenergysystems.
APPENDIX C:
Data Summary
Datafromtheliteraturesourceshasbeenextractedinordertoestablishrelevantbaseline
metricsandconsolidatethemintocommonunits.TherawdataisshowninTable18.An
additionalcolumnaddressingtheobjectivityofeachdatasourceisalsoincluded.Thereare
somecriticalparametersthatcanbeextractedfromthedataandconsolidatedintocommon
units.TheseareshowninTable19.

Table 18: Raw Data Extracted from Literature Review

Parameter
Low
Estimate
High
Estimate
Source
Objectivity
Ranking (1=low,
3=high)
Optimal vibration frequencies (Hz) 100 120
Cook-
Chennault
3
Tested wheel speeds (mph) 7.5 15 Virginia Tech 3
Voltages (V) 400 700 Virginia Tech 3
Amperage (mA) 0.2 0.35 Virginia Tech 3
Power Duration (s) 0.1 0.2 Virginia Tech 3
Maximum measured power per event,
(W)
0.08 0.14 Virginia Tech 3
Virginia Tech Traffic Flow speed (mph) 40 Virginia Tech 3
Virginia Tech Traffic Flow rate (vehicles
per day)
4000 Virginia Tech 3
Oregon DOT Traffic Flow Rate (vehicles
per hour)
600

Oregon DOT 3
Energy Generated for 1.0 km, Oregon
(kWh/month)
350,000

Oregon DOT 3
Number of harvesters, Oregon DOT 6000 Oregon DOT 3
Energy harvested for bridge mounted
devices, per vibration (microJ )
18

S.F. Ali, et al 3
Vehicle speed for micro harvesters (m/s) 25 S.F. Ali, et al 3
kW per km 0.0018 Berkeley 3
units per km 10000 Berkeley 3
Axles per vehicle 2 8
Berkeley,
Oregon
3
Target Cost of piezoelectric material (per
unit)
$1

Channel
Technologies
2
Energy generated in 1 km stretch of road
(kWh)
400 600
POWERleap,
Treevolt
1
C1
Parameter
Low
Estimate
High
Estimate
Source
Objectivity
Ranking (1=low,
3=high)
Time span of energy measurement (hr) 16
POWERleap,
Treevolt
1
Traffic flow rate, POWERleap
(vehicles/hr)
12.5 25
POWERleap,
Treevolt
1
Vehicles per hour 600 POWERleap 1
Power rating (kW) 720 POWERleap 1
Length of energy harvesting section (km) 1 POWERleap 1
Number of harvesters per 1 km 6,000 POWERleap 1
Power per unit per impact (W) 10 POWERleap 1
Power generated per sq ft, foot traffic
(W/ft^2)
1.13 Piezo Power 1
Cost per square foot, foot traffic ($/ft^2) $1.50 Piezo Power 1
Power rating, 1.0 km (kW) 200 Innowattech 1
Power rating, train (kW) 120 Innowattech 1
Traffic flow rate (vehicles per hour) 600 Innowattech 1
Vehicle speed (kph) 72 Innowattech 1
Train speed (wagons/hr) 300 Innowattech 1
Size of each unit (ft^2) 1 Virginia Tech 3
Power per km (kW) 100
Innowattech,
Haaretz
article
1
Cost per km ($) $215,400 $650,000
Innowattech,
Haaretz
article
1
LCOE ($/kWh) 0.06 0.08 Genziko 1
Lifetime (y) 20 Genziko 1
Installation cost ($/W) 0.4 Genziko 1
Capacity Factor 0.32 0.4 Genziko 1
Vehicles per hour 600 Genziko 1
Power Density (kW/km) 13,600 Genziko 1
Long dimension of unit (m) 0.45 Genziko 1
Short dimension of unit (m) 0.3 Genziko 1
Genziko Units per km 2,222 3,333
Calculated
from Genziko
1
Number of harvesters 9,800
Calculated
from Virginia
Tech
1

TheinformationinTable19sortsthenumbersbysource.Inthetablesshown,thedatais
organizedintoaprobabilitydistributionforaMonteCarlofinancialanalysis.Thefarright
C2
columnlabeledDistribution(meanshown)indicatesthatthevalueinthecolumnisthemean
ofadistributionofvalues.Thedistributionofvaluesisatriangularprobabilitydistribution
createdfromthespreadofvaluescollectedfromthedata,extractingtheminimum,average,
andmaximumvalue.AnexampleofoneoftheprobabilitydistributionsisshowninFigure30
(estimationoftheclaimedpowergeneration).

Figure 30: Example Probability Distribution Generated from the Range of Numbers Extracted from
the Literature Review

Theadvantageofcollectingthedatainthiswayistoaccountforaspanofuncertaintyand
incorporatethisuncertaintyinthefinancialoutcome.Thefinalcalculationofthecostofenergy
willincludescenariossuchaswhenalloftheminimumvaluesalign,forexample,orwhenallof
themaximumvaluesalign,andtheprobabilityofthatoccurringgiventheknowndatafromthe
literaturereview.Moreofthiswillbeexplainedinthecostofenergyassessment.

Table 19: Report Numbers Consolidated into Common Units

Va
Tech
Innowattech ODOT Ali POWERleap/Treevolt Distribution
Vehicles per
hour
166.67 600 600 12.5 600 336.11
Vehicle
speed (mph)
40 44.64 55.75

47.52
Claimed
power
generated, 1
km (kW)

200 100 486.11 25 720 350.41
Number of
harvesters, 1
km

6,000 6,000 6,000
Cost per km $214,500 $1,086,000 $650,250
C3
C4
ThedatainTable20iscalculatedfromTable19.Again,probabilitydistributionsareusedand
thesewillactasinputsforthecostofenergycalculationwhererelevant.Thereferencesare
notedandareconnected(bynumber)totheupdatedfinalliteraturelistinthisreport.

Table 20: Additional Parameters Estimated from the Literature Summary

Calculated
Parameters
Probability
Distribution
Mean
Calculation
Installed Cost per
harvester ($)
$108.38 Installed cost per km divided by
number of units per km
kW per harvester 0.06 kW per km divided by number of
units per km
kW per mph 7.37 kW divided by vehicle speed
Spacing Interval
(per m)
6.00 1000 meters (1 km) divided by
number of harvesters
Length of unit (m) 0.17 1 meter divided by number of
harvesters per meter
Units per km 6,000 Direct from literature
Length of
installation (km)
1 Direct from literature

Thisdatawillbeusedtocalculatethecostofenergy.Thecostofenergybeginswithasimple
proceduretoestimatethetimetopaybacktheoriginalinvestment.Forexample,if1kmof
roadwayisbuiltatacostof$650,000,thentheenergygeneratedperday(kWh)multipliedby
thesalepriceofenergyperkWh($/kWh)willberevenueusedtopaybackthatinvestment.The
timetopaybackwillbedeterminedbyhowmuchenergyisproducedandthesalepriceofthat
energy.Thisisasimplepaybackcalculationbutdoesnotincludeinterestrates.Anadditional
netpresentvaluecalculationwillalsobeprovidedwithappropriatediscountrateassumptions.
TheMonteCarlomethodwillprovideadditionalinformationsuchaswhattheminimum
tolerableelectricityrevenuesneedtobeinordertoprovidereasonablepayback,whatthe
lifetimeofthesystemshallbe,orwhatconditionsshallexistinordertoensureapaybackovera
certaintimeframe(threeorfiveyears,forexample).Inadditiontothisinformation,asensitivity
analysiswillprovidesensitivityindicatorswhichwillbethecriticalvariablesthatinfluencethe
calculation,suchasvehiclespeedorvehiclesperhour,forexample.
Theanalysiscanbeprogressivelymoredetailedbyusingthispaybacktechniquetocalculate
thenetpresentvalueoftheinvestmentwithrespecttoafuturetimeordate.
APPENDIX D:
Evaluation Criteria
What would an evaluation of the technology look like?
Anyevaluationshouldincludeananalysisofseveralcriticalparametersandanassessmentof
theimpactoftheseparametersontheperformanceoftheenergyharvestingsystem.The
analysiscanincludedemonstration,labscaleverification,andacceleratedlifetimeteststo
examinetheeffectofpiezoelectricmaterialsanddevicesonthelongevityormaintenanceofthe
roadway,maindegradationcharacteristicsofthecompositesystem,energygeneratedasa
functionofvehiclespeed,weight,andtrafficvolume,andthereliabilityofenergygenerated.
Thesecriticalparametersaredescribedinmoredetailbelow.
Manyofthesetestsandexaminationmethodshavecommonalitybetweenroadwayandrailway
installations,althoughtherearesomefactorsthataremoreapplicationspecificandtheywillbe
describedassuch.
Module Power Output
Firstandforemost,thepoweroutputofanindividualmodulemustbequantified.Themodule
outputintheuninstalledconditionwillbedifferentthanwhatitcanreplicateintheroadway.
Layersofasphaltandunderlyingsubstratewillaffecttheresponseofthemoduletostimulus.
Quantifyingtheseeffectscaneitherbedonedirectlyinasimulatedinstallation,ortheycanbe
modeledwithfiniteelementanalysis(FEA).Mostimportantly,thepoweroutputshouldbe
measuredinarealworldconditionthatis,onaloadcalibratedtobeequivalenttoagridor
energystorageconnectionsuchthatactualpoweroutputismeasured.
SuchanevaluationcanbedoneonastressframesimilartheInstrontestframeshowninFigure
31.Thedevicecanbeinstalledintheframeandloaded(compressedorcycled)tostimulatethe
powergenerationmechanism.Actualpoweroutputintheuninstalledconditioncanbe
comparedagainsttheinstalledconditiontoverifyperformanceandquantifythepoweroutput
permodule.
Testingthemoduleintheinstalledconditioncanbedonewithasimulatedinstallationor
withFEA.Theresultsfromtheuninstalledtestcanbemodeledinaninstalledconditionby
replicatedtheconditionsthatproducedaknownwattage,andthenmodelingmaterialoverlays
suchasasphalttodeterminetheireffectontheforcesfeltbythepowermodule.Thenetpower
outputcanbecalculatedinthisfashion.
TheInstrontestframeshownisinFigure31isequippedwithajigforperformingthreepointor
fourpointbends,andiscapableoffatiguecycles(locatedattheDNVTechnologyCenter,a
materialstestinglab).Similarmachinescanreplicatecyclingatfrequenciesnear40Hzandcan
performcompressionortensiontestsat10,000psiormore.Ajigforperformingtestson
piezoelectricmaterialscanbesimilarlyconstructedtoquantifypoweroutputasafunctionof
load.Adaptingamachinefortestingpiezoelectricmaterialsistechnicallyfeasiblewiththe
appropriatemodifications.
D1

Figure 31: Instron Test Frame at the DNV Technology Center

Figure 32: Finite Element Models of a Stressed Member with Layered Materials of Differing Moduli

Duration of Hit (Power Pulse Duration)

Relatedtovehiclespeed,therollingwheeloverapiezoelectricpanelrepresentsacontinuous
rollingstressthatisdifferentthanasinglehitlikeasteponfloorpanel.Thereforethereisan
elementofuncertaintyinthepowergenerationpotentialbasedonthespeedoftrafficasithasa
directimpactonthedurationwhichpiezoelectricdevicesarestressedandthusgenerating
power.
D2
Intheroadwayapplicationitappearsthereisastrongdependenceonvehiclespeed.Inthe
railwayapplicationthereislessdependenceonspeedsincethetrainrailactsasanefficient
forcetransfermechanism.
FortheGenzikoproduct,theproductliteratureindicatesadifferentenergyharvesting
mechanismlikelyrelatedtoresonanceatfundamentalfrequenciesassociatedwithtraffic,and
thereforethehithasalingering,decayingoutputnotunlikeadampedharmonicoscillator.
Energy Transfer and Vehicle Weight
Testingofthemodulepoweroutputwillinformwhatvehicleweightsareneededtomaximize
performance.Themodulepowerprovidesamaximumbaselineagainstwhichallother
performancemetricsmaybecompared.Inearlyphases,thiscanbeinferredbywayof
modelingorcalculation.Inlaterphasesitcanbeconfirmedbydemonstration.
Theenergyofthehitwilllikelyvarywithvehicleweightbecauseoftheenergytransferred
throughtheasphaltlayertothepiezoelectricsystem.Theasphaltisnotacompletelyrigid
materialmediumandthereforeheaviervehicles(suchastrucks)willlikelytransfermore
energytothepiezoelectricdevicesthanlightervehicles(suchascars).Thereismuchdata
suggestingthattrucksimpartmoreenergythancarsandtheVirginiaTechdemonstration
targetstrucksspecifically,presumablybecausemoreenergycanbeharvested.
Theangleatwhichavehicletireencountersthepiezoelectricdevicewillalsoaffectitsoutput.
Sincethereissignificantvariabilityinthetrackwidthofvehiclesandahighprobabilitythat
manyvehicleswillencounterthedevicesonalessthanidealtrajectory(suchaswhile
changinglanes),theoutputofthesystemshouldbeanalyzedwithregardtothisuncertainty.
Theroadwayharvestersemployaforcemultiplicationmechanismtoaccountforthis
uncertainty,butitwillhavelessrelevanceifthecarortruckchangeslanes(forexample)when
passingtheregionwheretheharvestersareinstalled.
Althoughtheidealpiezoelectricdevicehasalimitthresholdofthemaximumenergyitcan
generate,thenatureofthesurroundingroadwaymaterialswilllikelydampenitsresponseand
affectthetotalenergygenerated.Theperformanceofthesystemshouldbestudiedintheform
ofvehicleweightversuspowergenerated.Suchdatawouldinformtheoperatorwhetherthe
systemisbettersuitedfortrafficpatternswithheavyvehiclesversuspassengercars.
Durability and Lifetime of the Piezoelectric Ceramic
Piezoelectricmaterialsarelikeothersolidstatematerialsinthattheywilldegradeovertime,
resultinginreducedoutputandresponsetime.Thisdegradationisrelevanttobothroadway
andrailwayapplications.
Sinceademonstrationwouldtakesometimetoproduceresults,acceleratedtestmethodsare
recommendedtorankthelongevityofproductsinsimulatedenvironmentalconditions.
Factorsthatinfluencedegradationaretemperature,moisture,stressesgreaterthanthedesign
load,andunevenstresses(bendingmoments)thatcancrackandfracturethebrittlematerial.As
thepiezoelectricdegrades,theresponsefrequencywilldrift,thecapacitanceofthematerialwill
decrease,andthecouplingcoefficientwillchange.
D3
Sincelifetimeofthepiezoelectricdevicesisamajoruncertainty,thereshouldbeextensivestudy
ofthelifetimeanddurabilityofthesystemthroughacceleratedageingandweartests.
ModificationsofASTMC627,StandardTestMethodforEvaluatingCeramicFloorTileInstallation
SystemsUsingtheRobinsonTypeFloorTester,havebeencitedasanadequatetestingprocedure
forabrasionresistanceofpavements
28
.ThisASTMtestoramodifiedweartestshouldbeused
torankthewearresistanceofthecompositepiezoelectricpavementroadsystemagainstknown
roadwearcharacteristicsandload,asitwillhelpcharacterizethelifetimethathasadirect
impactonthecostofenergy.
StandardssuchasMILSTD1376describethedurabilitycharacteristicsofpiezoelectricceramic
materialsfordefenseapplications.Thisstandarddescribessixtypesofpiezoelectricceramic
materialsutilizedforsonartransducersfortheNavalservice.Thisstandardalsodescribesthe
propertiesoftheceramiccompositionsforthesesixtypesasmeasuredonstandardtest
specimens.Thedurabilityspecificationsandtestingrequirementswouldhaverelevancetoa
roadwayorrailwayapplication.Thesestandardsaddressthepropertiesoftheceramic,butnot
theageingbehaviorofthesystemasawhole.
Durability of System and Components
Thefunctionalcomponentoftheenergyharvestingsystemisthepiezoelectric,however,the
associatedelectronics,frame,andstructuralcomponentsofthepiezoelectricdevicesarealso
criticaltoitslongtermfunctionality.Inaroadwaychallengesmayexistsuchasisolatingthe
piezoelectricanditsassociatedelectronicsfromtemperature,moisture,andloadingeffectsthat
areinevitablylinkedtotheharshenvironmentoftheroadway.Theauxiliarycomponentsofthe
piezoelectricharvestingdevices,suchasforcemultiplyingmechanisms,wiring,circuitry,and
electronics,wouldrequireweatherhardeningandvalidationwheretheycanwithstandharsh
conditions.TestssuchasASTMB117,StandardPracticeforOperatingSaltSprayApparatus,have
beenmodifiedtotesttheweatheringperformanceofmaterialsusingageingaccelerationvectors
suchastemperature,saltwaterexposure,ultravioletlightexposure,andwet/drycycles.
Fortherailwayapplication,theexposuretoexternalweatherisapplicableandminor
modificationstoASTMB117wouldberequired.However,theburiedenvironmentofthe
roadwayimpliesnotonlywaterexposure,butcompressionforcesandmaterialsinteractionsof
adifferentnaturethantherailsystem.
Theroadwaydemonstrationdataindicatesthatapproximately5cm(~2in)ofasphaltis
installedoverthepiezoelectricsystems.Whilethesedemonstrationsarevaluable,theydolittle
toindicatethelongtermviabilityofthetechnology.VirginiaTechhasnotedthatwaterproofing
andwaterresistancearetechnicalchallengesfacedbythetechnology.Inaddition,thereisa
questionastowhetherthepiezoelectricsubstrateimpactsthedurabilityoftheroadsystemand
introducesunevenwearpatternsintheupperlayerofasphalt.Acriticalcomponentto
understandageinginthesetestsisnotjustwearandabrasionofasphalt,butthecombined

28
Willoughby,Kim.PostConstruction&PerformanceReport,ExperimentalFeaturesWA0304,0401,
and0504,StuddedTireWearResistanceofPCCPavements,Contract6620I90ArgonneRoadto
SullivanRoadMP286.91to292.38.WashingtonStateDepartmentofTransportation.2007.
D4
effectofwearandloadandtheripplingordimplingbehaviorofasphaltasaresultof
unevensubstratecompressionintheareaswherepiezoelectricenergyharvestingdevicesare
installed.Thethicknessoftheasphaltlayerlikelyhassomeindicationofitslifetimeandability
toendurewear.Thesubstrateoftheasphaltinstallationwillbeconstructedofmaterialswith
variablestiffnessranges,andtheimpactofthevaryingsubstratestiffnessshouldalsobe
investigatedfortheroadway.Forexample,ifthestiffnessofthepiezoelectricdeviceisequalto
thesubstrate,thennounevenwearwouldbeexpected.Ifthestiffnessofthepiezoelectricdevice
islessthanthesubstrate,thentheremaybeanindentationformingabovethedeviceovertime.
Ifthestiffnessofthepiezoelectricdeviceisgreaterthanthesubstrate,thentheremaybe
indentationsformingaround(butnotabove)thedeviceovertime.
Appropriateweathering,wear,andtribologyinvestigationsshouldbeconductedtoexamine
thelifetimeofthecompositeinstallationonboththeasphaltandthepiezoelectricdevices.
Failureratesofthedeviceswillhaveadirectimpactonthesystemsabilitytogenerateenergy.
Decaymechanismsinthepiezoelectricmaterialsthemselvesmayreducetheoutputofthe
system.Presumably,thepowerorenergygeneratedovertimewillbedecreasingasthesystem
hasacumulativeincreaseinindividualunitfailuresanddegradation.Anestimationofthis
failureratewouldrefinetheLCOEcalculation.
Finally,anydowntimeassociatedwithmaintenanceorreplacementshouldbeconsidered,as
thiswillalsoreducetheenergyoutputandaffectthereturnoninvestment.Downtimeinthe
railwaymayhaveadirectimpactonridership,whereasdowntimeinthehighwaycanincrease
laborcostsandstrainmaintenancebudgets.
Traffic Volume
Ashasbeenshowninallofthedemonstrationdatasofar,trafficvolumehasadirect
relationshiptothepowergeneratedforthesedevices.Ithasalreadybeenshownthatthevehicle
characteristicsandvehicleweightgreatlyinfluencetheperformanceofthesystem.Thetraffic
modeldevelopedforthisreportshouldbevalidatedbyactualvehicleassessmentsintheevent
ofademonstration.Onceconfirmationofthepowerpermoduleisobtained,theLCOE
estimationtechniquesemployedinthisreportcanberevisitedtoreassesstherequiredtraffic
characteristicstomakethesystemviable.
Dataintermsofvehiclesperdayorvehiclesperhourisnecessaryinordertoassessthecost
effectivenessofthesystem.Thetrafficvolumelikelyhasanimpactonthewearandtearofthe
system,soeffortstoquantifythisshouldalsobeconsidered.Trafficvolumeisdirectly
proportionaltotheequivalentcapacityfactorofthesystem.Theprofileoftrafficactivityisalso
ofimportanceasitwilldeterminewhentheenergyisavailable.Ifnoenergystorageisused,the
timingofenergyavailabilitymaybecritical.ModelingtherealLCOEwithavariabletraffic
volumeprofilewouldbebeneficial.
Vehiclesontheroadwayarefreetochangelanesatwill,andarealsoabletovarytheirposition
withinalanetosomedegree.Therefore,thereissomeprobabilitythatwhenenergyharvesters
areimplantedintheroad,passingvehicleswillnotimpartenergytothedevicesduetolane
changesorasymmetricalpositionwithinthelane.Therefore,thereisaneedtoquantifywhat
D5
numberofcarsandtruckswillactuallycontactthesystemwhenpassing.Ametricsuchas
numberofcarsperdayislikelygreaterthanthenumberofcarsthatwillactuallyproduce
energy.
Intherailwayapplication,virtuallyeverypassingtraincanensuretheharvestingofenergy
throughthepiezoelectricsystem.However,theoptimumnumberoftrainswouldneedtobe
calculatedtoensureanROI.
Vehicle Speed
Thedurationofavehicletireimpactwithabumpintheroadisaudiblynoticeable.Fastmoving
trafficcreatesabriefandloudreportwhentireshitaroadimperfection,suchasanexpansion
joint.Slowmovingtrafficcreatesaquieterreportwithalongerduration.Thesameistruefor
theimpactofthepiezoelectricdevices.Thesedeviceshavesomeextendedsurfaceareaand
length;atirerollingoverthemislikelycapableofproducingenergyoverafractionofthat
length.Thereforethereissomequestiontodetermineifalongerdurationofenergytransfer
(slowervehiclespeed)maycreatemorewattagefromthedevice.However,withdecreasing
vehiclespeed,thereislesskineticenergythatcanbetransferredtothedevice.Therefore,there
islikelyatradeoffbetweenvehiclespeedandpowergeneratedandthisshouldbedetermined.
Suchdatawouldindicatewhetherthetechnologyisbettersuitedforhighways,lowerspeed
mainarteries,orbusyintersections.
FortheGenzikoproduct,theliteratureillustratesveryhighenergydensities.Itisnotclearfrom
theliteraturewhatthemajorfunctionalcomponentsareinthetechnology,butitappearstobe
somethingsimilartoanarrayofcantileverpiezoelectricmaterialsthatvibrateorresonateat
frequenciesassociatedwithtrafficorpassingvehicles.Thisraisesthequestion,however,
concerningthesuitabilityofthisfrequencyrangeforalltrafficpatterns.Ifthereisadependence
onvehiclespeedforthistechnology,itmayshowthatthedevicehasverygoodenergydensity
atonlyspecificfrequenciesbutlowenergydensityatotherfrequencies(perhapslinkedto
vehiclespeedsoutsidethebestperformanceenvelope).Itisnotclearwhetherthisisaconcern
butthisshouldbevalidatedinastudy.
Asmentionedabove,therailwayapplicationappearstohavelessdependenceontrainspeed.
Installation Methods
ThedemonstrationwithVirginiaTechandthemediaproducedbyInnowattechhave
demonstratedthatthesesystemscanbeinstalledbyasawcutinstallation(cuttingorgrinding
sectionsofroadwayapproximatelythewidthofavehicletire,installingthepiezoelectric
devicesatauniformspacinginthetrench,connectingthedevicesandtrenchingthe
consolidatedoutputtotheroadside,encasingthearrayinconcreteorepoxy,andrepavingthe
installationinnewasphalt).Thecostanddowntimeassociatedwiththisinstallationshouldbe
investigatedinmoredetail.Inparticular,ifmaintenanceofthesystemistobecompletedat
regularintervals,thisdowntimeshouldbeincorporatedintothelifetimeenergyproduction
estimate(asitwillalsoaffecttheLCOE).
IthasbeenshownbyVirginiaTechthattheinstallationinvolvesasawcutintheroadfollowed
bychiselingtocreateapocketintowhichthedevicesareinstalled.Theyareepoxiedtothebase
D6
ofthisinstallation,andtheremainingareaaroundthemisfilledwithepoxyorconcrete.
Additionalchannelsarecuttorunwirestothesideoftheroad.Thesearealsofilledwith
concrete.Asphaltcanthenbelaidoverthetopofthisinstallation.
TheGenzikoproductisassumedtobeinstalledinasimilarfashion,withtheexceptionofspeed
rampswhichcanbelaidontopofaroadstructureforspeedcontrolandrequirelittle
modificationtotheroadway.
Therailwayapplicationappearstorequiremuchlessinstallationlaborthantheroadway
systemandthedevicescanbeinstalledaboveground,betweentherailandtherailtie.The
unitsfortherailapplicationarealsosmaller.Therefore,thecostofinstallationislikelylessfor
railwaysthanroadways.
Value of Additional Data and Avoided Inspection Costs
Thepiezoelectricsystemalsoprovidesalotofinformationaboutroadandtraffic.Thisdatamay
bepossibletoincorporateintoexistingmonitoringandstatisticssystemsneededbythe
DepartmentofTransportationortrafficsystems.Thetypeandqualityofthedatashouldbe
evaluatedtodetermineifthereisaddedvaluewhichcanbemonetizedinordertooffsetthe
costofenergy.
Inaddition,ifthisdataprovidesareductionininspectioncosts,thismayhavevaluetothe
overalloperationoftheroadway.Inthepast,DNVKEMAhasassessedthecostofstructural
healthmonitoringsystemsonwindturbinebladesandfoundthatavoideddowntimedueto
bladereplacementsandinspectionsamountedtosignificantcostsavingsoverthelifeofthe
system
29
.
Intherailwaycase,theenergygeneratedfrompiezoelectricdevicescanprovidedataaboutthe
trainweightsincethevoltageproducedisafunctionoftheforceimpartedonit.Overweight
trainscouldbetargetedandremovedfromservicebytherailoperator.
Energy Storage versus Net Metering
Generally,thesystemispresumedtocostlessifenergystorageisnotrequired.TheLCOE
calculationdoesnotseparatelyparseoutthecostsofinvertersorenergystorage,butonly
estimatesthecostbasedonquotedtotalinstalledcosts.Presentdemonstrationsimplythatno
energystoragewasemployed.Becausethepiezoelectricenergygenerationsystemisvariable
withroadtraffic,itisnotunlikerenewableenergysystemsinthatitscapacityfactormaybelow
anditsoutputmaybeintermittent.Thereareatleasttwo,ifnotthreecomponentsfor
renewableenergysystemsthatbringpowertothegridThesecomponentsareanenergystorage
orenergyconversionsystem(ifdesired),adirectcurrenttoalternatingcurrent(DCAC)
systemwhichistypicallyhandledbyinverters,andfinally,theoutputstagewhichiseither
constantoutputforafixedtimedurationoritisnetmetered,meaningthattheutilitymetercan

29
Willoughby,Kim.PostConstruction&PerformanceReport,ExperimentalFeaturesWA0304,0401,
and0504,StuddedTireWearResistanceofPCCPavements,Contract6620I90ArgonneRoadto
SullivanRoadMP286.91to292.38.WashingtonStateDepartmentofTransportation.2007.
D7
spinbackwardorforwardandthereforeaccountfornetenergyproductionbysubtracting
producedenergyfromgridenergy.Thesesubsystemsaredescribedinmoredetail:
EnergyStorage:Themostcommonformofenergystorageforrenewableenergysystemsislead
acidbatteries.Leadacidbatteriesarecosteffectiveandaresuitableforstationarystorage,such
ascommercialorresidentialbatterystorageforsolarPVinstallations.Leadacidenergystorage
systemsarecommonlyseenforsystemsinthe0100kWrange.Higherenergydensitybatteries
suchasLiion,NiCd,andNaSbatterieshavebeendemonstratedyetallhavehighercoststhan
leadacid.Appropriateevaluationoftheenergystoragetechnologyifrequiredisneededto
measurethebenefitsofcost.Alternativeenergyconversionsystemsmaybemorecomplicated
andcouldconsistofelectrolysissystemswhichproducehydrogenandthenmaybestoredand
laterimplementedintoafuelcell.Thesesystemsmaybemorecomplicatedthanbatteriesbut
dependingontheapplication,theymayhaveadvantagessuchasgreaterscalabilityorlonger
storagedurations.Asinthecaseofthebatteries,thesesystemsrequireacostandlifetime
studytoevaluatetheappropriatepairingofpiezoelectricandenergystoragesystems.
Inverters:Thereareawiderangeofinverters.Manyaresolidstatedevicesandtheircostis
directlyproportionaltotheirpowerrating.Therehavebeenrecentinnovationsininverter
technologysuchasthemicroinvertersofferedbyEnphasewhichcanattachdirectlytotheback
ofindividualsolarpanels.TheselowcostdevicesproduceanACpoweroutputdirectlyfrom
thepanelandminimizetheelectricalconnectionsneededtoconsolidatepoweroutputfroma
solararray.SincetheUnitedstatesgridoperateswitha60Hzalternatingcurrent,anyDC
sourcemustbeconvertedto60HzACattheappropriatevoltage(usually120or240V).
Innowattechhasimpliedthatatransformerhasbeendesignedtomatchtheimpedanceofthe
energyharvestertotheoutputsystem,andthistransformerismoreefficientinarailway
applicationthanroadwayapplications.
Output:Theneedsfortheoutputdependontherequiredcapacityfactorandwhetherornot
thesystemisgridconnected.Inamicrogridsituation,theremaybeaneedtostoretheenergy
sothatitcanbedeployedattherighttime.Forexample,iftheroadwaypiezoelectricsystemis
intendedtopoweradynamicbillboardduringrushhour,energyproducedbythepiezoelectric
systemcanbestoredintoanappropriatelysizedbatteryandthenthebatterysystemcanoutput
energyforafewhourstopowerthebillboard.Suchasystemwouldnotneedtobegrid
connected.However,ifthebillboardisintendedtobevisibleallday,anenergystoragesystem
maybeimpracticalandanetmeteringsystemmaybemorecosteffective.Ifthebillboardis
meteredonasinglemeter,thenthepiezoelectricsystemcanbeconnectedtothismeter;the
metercanbeconvertedtoanetmeteringsystem,andtheenergyconsumedbythebillboardcan
beoffsetbythepiezoelectricenergygenerated.
Netmeteringisoneofthemostcosteffectivewaystoimplementrenewableenergybecauseit
canbypassthecostofenergystorageanddirectlyaccountforproducedenergybyliterally
subtractingitfromtheenergymeterreading.

D8
Figure 33: Three Stages of Power Conditioning and Conversion to Deploy Intermittent DC Power
Sources

Source: DNV KEMA Energy and Sustainability

Thedeploymentofthesystemshouldoutlinetherequirementsandcostsofinvertersand/or
energystorageandshouldjustifytheappropriateenduseoftheenergysuchthatthesystemis
optimizedinthemostcosteffectivefashionpossible.TheGenzikoproductexplicitlyillustrates
thattheirsystemcanbeusedwithorwithoutenergystorage.
D9
APPENDIX E:
Railways
Innowattechhassomelimiteddataontheuseofpiezoelectricmaterialsforthepurposeof
harvestingvibrationenergyinrailways.Thetechnologyappearstobeincorporatedintothe
track.Thereisalsoanindicationofthedevicebeingincorporatedintotherailcarsthemselves,
asshowninFigure34.Thegenerallayoutfortherailwaypiezoelectricenergyharvesteris
shownin2.9 RoadwaysversusRailwaysonpage40.

Figure 34: Innowattechs Rail Applications

OnemajorinnovationappearstobeInnowattechsinstallationofrailwaydeviceswithlittleor
nointerferenceorreconstructionrequiredontherailwaytrack(Figure35).Thereisvalueadded
datafromtheInnowattechdevices,suchasthetransformationofmechanicalstressesinto
electricaloutput(voltage),whichshalldeterminethenumberofwheels,weightofeachwheel,
thewheelscapitationandwheelperimeterposition.Inadditionthespeedofthetrainandthe
wheeldiametercanbeconcludedviathefixeddistancebetweenpads.Theenergyisself
suppliedbythesystem
30
.

30
InnowattechWebsite.www.innowattech.comAccessed1/3/2013.
E1
Figure 35: Installation of Innowattech Devices on a Rail System

E2
APPENDIX F:
Calculation Details
Compression-based Harvesters, Vendor Claims
CapitalandcostofenergycostsinFigure17andFigure18weretakenfromtheminimum,
maximum,andaveragevaluesoftheNRELestimates
31
.
Thecostofenergycalculationinvolvesthefollowingparameterwithtimedependence(Table
21).

Table 21: Explanation of Calculations for the Model

Parameter [unit] Description Calculation
Months [integer] Time unit for model The lifetime is determined by
number of months or fraction of a
year.
Construction or Maintenance
Investment [$]
Costs incurred with the
installation or replacement of the
piezoelectric roadway system.
The probabilistic value ($/km) is
first incurred in the first month
(Month 1). The next date when
the cost is incurred will occur
when the lifetime expires.
If lifetime trigger =true, then the
replacement cost [$] =Cost per
km [$/km] * number of kilometers
[km]
Lifetime Trigger [conditional,
integer]
Determines if the lifetime of the
system has expired.
Generation of system =int
(month/12). Expiration and
replacement corresponds to the
lifetime trigger increasing from
one integer to the next.
kWh Generated [kWh] Number of kWh harvested from
roadway system per month.
Assumes no energy is generated
for the month of a replacement
event.
24 hours per day * 30 days per
month * claimed power
generated [kW]
Energy Prices [$/kWh] Uses energy prices from the
Sacramento, CA region.
Assumes the energy prices are
increasing over time at stated
rate, for example, compounded.
Price(n+1) =price(n)*(1+rate)

31
OpenEI:OpenEnergyInfo.http://en.openei.org/wiki/Transparent_Cost_Database.AccessedJanuary1,
2012.NationalRenewableEnergyLaboratory(NREL),OpenGovernmentInitiative,USDepartmentof
Energy.
F1
Parameter [unit] Description Calculation
Energy Revenue [$] Revenues earned by system due
to energy generation
Energy Revenue [$] =Energy
Prices ($/kWh) * kWh Generated
[kWh]
Year Time in years Month/12
Present Value of Investment [PV] The present value of future costs
and revenues.
(Installation and maintenance
costs +Energy Revenues) /
(1+discount rate/12)^year
Net Present Value The discounted total of all
investments and revenues at a
future time according to the
discount rate.
NPV(n) =sum(PV(0):PV(n)

Case 1: Maximum Five-Year Lifetime

Itcanbeseenthatthelifetimeofthesystemisassumedtobebetweenonetofiveyearswitha
mostlikelylifetimeoftwoyears(Table22).Presumablythesystemwilllastlongerthanthis,
however,theupperlimitontheliteraturereviewedonlydemonstratesatatwoyearlifetime
thusfar.Therefore,thisisincorporatedintotheprobabilisticestimationofthelifetime
correspondingtoageometricalmeanof2.67years.Asaresult,theLCOEiscalculatedtobe
$0.02/kWhataminimum,andabout$1.15/kWhatamaximum,withameanof$0.18/kWhanda
standarddeviationof$0.14/kWh(Figure36).ThefactorsthatmostgreatlyaffecttheLCOE
calculationareshowninFigure36.ThefactorsthatmoststronglylowertheLCOEarethe
powergeneratedandthelifetime.ThefactorthatmostincreasestheLCOEisthecapital
installationcost.

Table 22: Assumptions for the LCOE Model, Five-year Case

Min ML Max Dist
Discount Rate 0.05 0.06 0.1 0.07
Energy Inflation
Rate 0 0.01 0.03 0.013333

Lifetime of Unit (y) 1 2 5 2.67

F2
Figure 36: The LCOE Including Discounted Present Value of Future Investments

Case 2: Maximum Ten-Year Lifetime

Itcanbeseenthatthelifetimeofthesystemisassumedtobebetweenonetofiveyearswitha
mostlikelylifetimeoffiveyearsasshowninTable23.Therefore,thisisincorporatedintothe
probabilisticestimationofthelifetimecorrespondingtoageometricalmeanof5.67years.Asa
resulttheLCOEiscalculatedtobe$0.014/kWhataminimum,andabout$0.41/kWhata
maximum,withameanof$0.08/kWhandastandarddeviationof$0.05/kWh(Figure37).The
factorsthatmostgreatlyaffecttheLCOEcalculationareshowninFigure38.Asstated
previously,thefactorsthatmoststronglylowertheLCOEarethepowergeneratedandthe
lifetime.ThefactorthatmostincreasestheLCOEisthecapitalinstallationcost.

Table 23: Assumptions for the LCOE Model, Ten-year Case

Min ML Max Dist
Discount Rate 0.05 0.06 0.1 0.07
Energy Inflation
Rate 0 0.01 0.03 0.013333

Lifetime of Unit (y) 2 5 10 5.67

F3
Figure 37: The LCOE Including Discounted Present Value of Future Investments for the Ten Year
Case

Figure 38: Sensitive Factors Affecting the LCOE for the Ten-year Case

F4
Case 3: Maximum Thirty-Year Lifetime
Thethirtyyearlifetimecaseassumesamaximumlifetimeof30years(Table24)withamean
LCOEof$0.03/kwhwithastandarddeviationof$0.02/kWh(Figure39).Itcanbeseenthatwith
anincreaseinlifetime,thecostofenergyissignificantlyreduced,althoughthepowergenerated
andlifetimeremainstrongfactorsinreducingthecostofenergy(Figure40).

Table 24: Assumptions for the LCOE Model, Thirty-year Case

Min ML Max Dist
Discount Rate 0.05 0.06 0.1 0.07
Energy Inflation
Rate 0 0.01 0.03 0.013333

Lifetime of Unit (y) 2 10 30 14.00

Figure 39: The LCOE Including Discounted Present Value of Future Investments for the Thirty-year
Case

F5
Figure 40: Sensitive Factors Affecting the LCOE for the Thirty-year Case

Traffic-based LCOE Technology Agnostic

Amodelwasconstructedtodisregardthevendorspecificationsforeachtechnologyandinstead
estimatetheminimumenergydensityrequiredinordertotargetareasonablecostofenergy.
Themodelisthereforeacrosscheckingmechanism.Theinputsandcalculationsareasfollows.

Table 25: Installation Metrics for the Generalized Case

Parameter Value Justification
Length of installation (km)
1 All examples work with a 1 km installation
Discount Rate
0.05-0.1 Averages to 0.07
Lifetime of Unit (y)
Variable Same as previous cases estimating 1-20
years.
Price for electricity sold
($/kWh)

Wheel force multiplier
4x10
-6
- variable This coefficient was taken from the line fit in
Figure 9. It is the ratio for harvester power
output to vehicle weight at the wheel.
Assumes a linear relationship. This factor is
tuned to estimate the LCOE and power
output per module.
Energy Cost rate Increase
0.03 Assumes energy rates are increasing
harvester spacing (in)
8 Based on Innowattech dimensions, can be
adjusted to fit footprint of Genziko
F6
Parameter Value Justification
Cost per km
$650,000 or
$27,000,000
Innowattech estimated $650,000/km. Using
$2/W estimate from Genziko, $27,000,000 is
the estimate for a 13.6 MW system
harvester pulse width (s)
0.1 Virginia Tech Demo
Vehicle Spacing Factor
Typically 0.05-
0.07
Adjusted in order to calculate vehicles/hr

Table 26: Traffic Metrics for the Generalized Case

Parameter Value Justification
weight at vehicle wheel Distribution
Ranges from 2200 58,500 N, based
on variation across all vehicle types
from Transportation Energy Data
Book.
power per unit (W) 0.1 - variable
Virginia Tech demonstrated 0.08-0.14
W per unit. This value is a critical
indicator of total system performance.
This value can be varied to determine
the minimum required power output
to achieve a target LCOE (for
example, $0.10/kWh)
speed (mph) 45-65
Variable ranging between 45-65 mph
depending on conditions. In most
cases a distribution assuming +/- 5
mph is assumed.
Wheelbase (ft) Distribution, 11-13
Calculated from distribution of
vehicles form Transportation Energy
Data Book.
harvester pulse width (s) 0.1
From Virginia Tech demonstration
can also be tuned to investigate the
impact on the model. For the Genziko
product, values up to 1 s were
assumed
Energy Sale price ($/kWh)
$0.10 variable,
uncertain
Variable, using Sacramento, CA
prices ranging from $0.09-0.15.
Number of axles Distribution, 2-3
Distribution from traffic data, primarily
2 axles. Heavier vehicles increase
average to nearly 3. This value
determines time between hits
Vehicle speed (fps) ~95
Convert mph to feet per second
Time between axle hits (s) 0.12
Vehicle wheelbase / feet per second
Time between Vehicles (s)
6 seconds (corresponds
to 600 vehicles/hr)
3600 s per hr / vehicles per hr
Vehicles per day 14,656
Vehicles/hr * 24 hrs per day

F7
Table 27: Calculated Metrics
Parameter Value Justification
Units per km
~9,800 2*1000 m/km / harvester spacing (in
m)
Nameplate capacity per km
(kW)
Variable dependent on
wheel force multiplier
Units per km * power per unit
power per unit (W)
Based on wheel force
multiplier, taken from the
line fit in Figure 9.
Wheel force multiplier * weight at
vehicle wheel
Capacity factor
Conditional if time
between vehicle hits is
less than pulse duration,
capacity factor is 1
Vehicle spacing factor * power pulse
duration / time between vehicle hits
Vehicles/hr
Generally ~600 Vehicle spacing factor * 3600 s/hr /
(Number of axles/vehicle * s/axle hit)
Actual output per km (kW)
Variable capacity
factor dependent
Capacity factor * nameplate capacity
per km
Simple COE
Based on time series
calculation for 240
months (20 years)
Sum of all maintenance expenditures
(lifetime expiration) divided by sum of
all kWh produced
Discounted COE
Based on time series
calculation for 240
months (20 years)
Sum of present value of all
maintenance expenditures divided by
sum of all kWh produced
50th Percentile NPV at 5
Years
Based on electricity sale
price
Net present value of revenues costs
60 months into the future
Capital Cost ($/kW)
Includes capacity factor Cost per km / actual output per km
Cost per Harvester
Installed cost Capital cost ($/kW) * actual power
output (kW) / number of harvesters
Power Density (W/ft^2)
Estimate based on
harvester spacing, which
may include space
between harvesters if
they are not spaced end
to end
Power per unit / (harvester spacing in
square feet)
2

Installed Power Density
(W/ft2)
Dependent on actual
installed area
Actual output per km /
(harvesters/km)* (harvester spacing)
2

Table 28: Time Series Metrics

Parameter Description
Generated Energy per month
(kWh/month)
Actual output per km (kW) * 24 h/day * 30 days/month
Value of Generated Energy kWh/month * inflating Energy Sale Price $/kWh
Cumulative Value of Generated Energy

Revenue of previous month +revenue of current month
Construction or Maintenance Cost If lifetime is expired, cost per km * number of km is subtracted
from revenue for that month
F8
Parameter Description
Energy Prices Inflating over time, compounded from energy price inflation
rate
Cycle Fraction Month number / lifetime in months. When this is an integer
the lifetime is expired and a replacement cost is incurred. See
construction or maintenance cost above
Cumulative P/L Sum of this months costs and revenues added to last
months costs and revenues since time =0
Present Value (Installation and maintenance costs +Energy Revenues) /
(1+discount rate/12)^year
Net Present Value NPV(n) =sum(PV(0):PV(n)

Figure 41: Approximated Triangular Probability Distribution of Traffic Wheel Weight using
Statistical Weight for Each Car Segment to Determine Most Likely Average Vehicle Weight

0
1
0
,
0
0
0
2
0
,
0
0
0
3
0
,
0
0
0
4
0
,
0
0
0
5
0
,
0
0
0
6
0
,
0
0
0
V
a
l
u
e
s

x

1
0
^
-
5

F9
Figure 42: Capital costs with a Fixed LCOE of $0.11/kWh Indicates a Tailing Distribution with a
Maximum of $102,000/kW and a Mean of $14,391

Figure 43: Estimate of the LCOE with Capital Cost of $4,000/kW Calculates 90 Percent of the
Values to be less than $0.20/kWh
90.0% 5.0%
$3,793 $36,919
$
0
$
2
0
,
0
0
0
$
4
0
,
0
0
0
$
6
0
,
0
0
0
$
8
0
,
0
0
0
$
1
0
0
,
0
0
0
$
1
2
0
,
0
0
0
0.00
0.05
0.10
0.15
0.20
0.25
0.30
0.0%
16.7%
33.3%
50.0%
66.7%
83.3%
100.0%
Capital Cost ($/kW) / COST METRICS
Capital Cost ($/kW) / COST
METRICS
Minimum $2,164.16
Maximum $102,066.26
Mean $14,391.51
Std Dev $11,689.18
Values 500

F10
$
0
.
0
0
$
0
.
1
0
$
0
.
2
0
$
0
.
3
0
$
0
.
4
0
$
0
.
5
0
$
0
.
6
0
$
0
.
7
0
$
0
.
8
0
$
0
.
9
0
Figure 44: Weight at Vehicle Wheel and Harvester Pulse Width Will Drive the LCOE if the Capital
Cost is Fixed at $4,000/kW

-
0
.
6
-
0
.
5
-
0
.
4
-
0
.
3
-
0
.
2
-
0
.
1
0
.
0
0
.
1
0
.
2
0
.
3

F11
APPENDIX G:
ADDENDUM: Reviewer and Stakeholder Comments
and Responses from Draft Report
Priortothefinalpublicationofthereport,commentsweresolicitedfromthegeneralpublicand
stakeholders.Responsesareprovidedtoclarifyorcorrectelementsoftheoriginaldraft.
Comment#1:
The 2010 Highway Capacity Manual (HCM) is a well-known standard and could
perhaps be cited in the final report with reference to traffic volumes and truck
percentages. Specifically, perhaps the HCM could be used with regard to reviewing the
information in Table 26: Traffic Metrics for the Generalized Case
Response:
TheHighwayCapacityManual(HCM)andtheTransportationEnergyDatabook(TEDB)have
linkedreferenceandsourcebackgrounds.TheTEDBhasbeenpublishedsince1976bytheUS
DepartmentofEnergyandisproducedbyOakRidgeNationalLab.TheHCMhasbeen
publishedsince1950andismanagedbytheTransportationResearchBoardandtheFederal
HighwayAdministration(FHWA).TheHCMisalimitedaccessreferencedatabasefor
subscriptionorpurchase,andhasastrongemphasisonanalysismethodsinadditionto
ancillaryinformationabouttraffic,suchasairqualityandnoiseanalysis.TheTEDBisanopen
access,onlinetoolwithdownloadablespreadsheetsthatofferstatisticsconcerningvehiclemiles
travelled,crosssectionsofthevehiclepopulation,andexplicitbreakdownofvehiclestatistics
concerningclassificationandfuelconsumption.TheemphasisoftheTEDBisenergybutdueto
itspurposetoestimatetheenergydemandsassociatedwithtransportitcontainsawealthof
statisticalinformationaboutvehiclesandtraffic.TheTEDBhasbeenshowntousestatisticsthat
correlatewithEIAandDOTtrafficandenergydatabases.Becauseofitseaseofaccess,itis
commonlyreferencedinpeerreviewedpublicationsandpublicdocuments.
Ifademonstrationistobeheld,datafromtheHCMcanbeincorporatedintoupdatedtraffic
modelsbasedonthevalidatedpoweroutputfrompiezoelectricenergyharvestingsystems.

Comment#2:
Theremaybeinterestinapooledfundresearchfundingmechanismbetweenstatesorregions
asaprojectfordemonstrationandvalidationofthetechnology.
Response:
Therecommendationfordemonstrationorevaluationwouldstillapplyinthiscase,withthe
laterstagesandreassessmentoftheeconomicmodelspertainingtospecificregionaltraffic
considerationsinsuchademonstration.Apooledfundbetweenregionswouldreducetherisk
enduredbythefundingagencies.
G1
Comment #3:
Correctionaboutthecompanyname:ThecompletecompanynameisChannelTechnologies
GrouporCTGinshort;itiscurrentlystatedasChannelTechnologies,Inc.throughoutthe
report.CTGsinvolvementcanbedescribedasthefollowing.InadditiontoInnowattech,CTG
workedwithPOWERLeaponthedesignoffloorbasedenergyharvestersbasedonpiezoelectric
ceramics.Inthiseffort,CTGoptimizedtheharvesterdesignthroughasoftwareprogramto
maximizethepoweroutpututilizingceramicproperties,itsdimensions,placementandnumber
oftheminatile,usingweightdistributionandgaitcycleofahuman.
Response:
ChannelTechnologiesshouldbereferredtoasChannelTechnologiesGroup(orCTG),not
ChannelTechnologies,Inc.asithasbeeninthisreport.Correctionshavebeenmadewhere
applicable.
Comment#4:
ThecostquotedforGenzikoincludesthegenerators,storage,powerelectronics,inverters,
cabling,andinstallation(inotherwordstotheconnectiontothegrid).Theinformationfor
Genzikoshouldbe:
Assumptions:
vehiclesperhour600
speed4565mph
powercapacity51.03MW
powergeneration13.61MW
#generators1300
VendorSuppliedUpdates:
costgenerators:$2,600,000forpilotproduction,$739,700forfullproduction
storage:0.5MWh/MW(6.8MWh)$170,112
1.0MWh/MW(13.61MWh)$340,225
2.0MWh/MW(27.22MWh)$680,450
powerelectronics,powerinverters,plantandinstallation(shouldbesimilarforallsystems):
13.61MW$8,851,663
totals:0.5MWh/MW($0.7875/W)$10,332,000
1.0MWh/MW($0.9125/W)$11,972,000
2.0MWh/MW($1.1625/W)$15,252,000
GenzikoRPGspowerlevelsarehigherthanInnowattech13.61MWvs.0.15MW.(907times,
not89.7times)
GenzikoRPGsgeneratorsarepricedaboutthesameasInnowattech$740kvs.$650k.(increase
by14Percent,not40.8timeshigher)
G2
G3
Response:
Afterthepublicationofthefirstdraftofthisreport,Genzikoofferedadditionalinformation
concerningthecostandoutputofthesystem,completewithabreakdownoftheBOScostssuch
asinvertersandenergystorage.Notethattheestimatesincludeanassumptionof1,300
generatorsin1km.Intwolanesthiswouldcorrespondtoageneratorsizeofroughly1.5m,
with650generatorspervehicleside(twotracksof650harvestersrunninginparallel).This
informationisnewtoDNVKEMAafterthepublicationofthefirstdraftofthereport.An
additionalnewpieceofinformationistheacknowledgementofthenameplatecapacityat51.03
MW.Accordingtothisinformation,thepowerperunitwouldbe51.03MW/1,300units=0.039
MW/unitor39kWperunit.Iftheunitissquare,itsfootprintwouldbeapproximately2.25m
2
.
Thiscorrespondstoapowerdensityof39kW/unitdividedby2.25m
2
whichis17kW/m
2
or1.6
kW/ft
2
.
ThisinformationimpactsTable4andTable2regardingassumptionsaboutthenumberof
harvestersandpowerperunit.However,sincethepowerperkmisconsistentitdoesnot
changethemainlessonslearnedonpage25,forexample,anysystemshouldmeetthefollowing
criteria:
1. Powerdensity>300W/ft
2
(inthiscaseamoduleoutput>150W)
2. A1020yearlifetime
3. Capitalcosts<$10,000/kW
4. ActualkW/km>100
Itshouldbenotedthatthecostassessmentbasedonthevendorclaimsachievesthiscriteria.If
thevendorclaimsaretruetheGenzikounithasaveryhighpowerdensityandprovidedthe
costsareaccurateacompellingcasetomeetthelistedrequirementsabove.
Howeverthemutualexclusivityfoundintheseanalysesremains.RecallTable11onpage33.
TheupdatedcapitalcostswouldplacetheGenzikosysteminthe$10$15Mrangefora50MW
system(nameplate),correspondingtoaCAPEXof$0.30/Wor$300/kW(nameplate).DNVhas
shownthatforcapitalcoststhislow,thecorrespondingLCOEisdrivenlowerthan$0.02/kWh
whichconflictswiththeGenzikoclaimsthatremaintobeinthe$0.06$0.08/kWhrange.The
mutualexclusivityinthiscaseremainsbytheDNVassessment,untilfurthervalidationcanbe
provided.
Thisdiscrepancyagainhighlightstheneedtofirstverifythesepoweroutputs,astheentire
economicpicturedependsonit.Ifthesesystemsareindeedcapableofgenerating39kWper
unit,thisshouldbethirdpartytestedsuchthattheremainingeconomicscanbeinvestigatedin
moredetail.Itisthepowerdensitythatiscriticaltothesuccessofthetechnologyanditmust
firstbeverifiedbeforetheeconomicassessmentsarerevised.