Proceedingsofthe2ndMarineEnergyTechnologySymposiumMETS2014

April15‐18,2014,Seattle,WA

FIELD MEASUREMENT TEST PLAN TO DETERMINE EFFECTS OF HYDROKINETIC TURBINE DEPLOYMENT ON CANAL TEST SITE IN 

YAKIMA, WA, USA 

BudiGunawan1,VincentNeary,JesseRoberts,AnnDallmanSandiaNationalLaboratories

Albuquerque,NM,U.S.

ShaneGrovueInstreamEnergySystems

Vancouver,BC,CA

JoshMortensen,BryanHeinerUSBureauofReclamation

Denver,CO,U.S.

1Correspondingauthor:budi.gunawan@sandia.govABSTRACT  The primary goal of the Department ofEnergy’s Water Power Program is to efficientlydevelop and utilize the country’s marinehydrokinetic (MHK) and conventionalhydropower (CH) resources. The program hasrecently identified theneed to better understandthepotentialforhydrokineticenergydevelopmentwithin existing canal systems that may alreadyhave integrated CH plants. Hydrokinetic (HK)turbine operation can alter water surfaceelevations and modify the flow in a canal.Significant water level alterations andhydrodynamic energy losses are generallyundesirable not only for CH plan operations, butalso for irrigation and flood managementoperations. Thegoalofthisstudyistobetterunderstandthe effect of operating individual and arrays ofdevices on local water operations through fieldmeasurements and numerical modeling. Amethodology to study the effect of hydrokineticturbine deployment in a test site in Roza Canal,Yakima, WA, is presented. The methodologycomprises detailed water level and velocitymeasurements to characterize energy gradelineand inflow and wakeflow fields. Results from apreliminarytestingarealsodiscussed.INTRODUCTION   Marine hydrokinetic (MHK) energytechnology, often in the form of underwaterturbines, is receiving a growing interest. ThistechnologycanpotentiallybedeployedinexistingUS canal system, which comprises tens ofthousands of miles of canals. Canal deployment

hasitsownadvantages,suchastheavailabilityofaccurate flow and water level information fromthelocalirrigationdistrictthatmanagesthecanal.Despite having this advantage, HK turbineoperation can alter water surface elevations andmodifytheflowinacanal.Significantwaterlevelalterations and hydrodynamic energy losses aregenerally undesirable for conventionalhydropower, irrigation and flood managementoperations.Littleisknownaboutthedetailsofthemechanism that causes these alterations. Thislackofknowledgeaffectstheactionsofregulatoryagencies, theopinionsofstakeholdergroups,andthecommitmentofenergyprojectdevelopersandinvestors. Therefore, there is anurgentneed forpractical studies and accessible tools to helpindustryandregulators toevaluate the impactofHK device deployment, especially on wateroperationandenvironment,inordertobeabletoapply mitigation measures and to establish bestsitinganddesignpractices. TheRozaCanal,Yakima,WA,USA,ismanagedby the US Bureau of Reclamation (USBR). USBRalsooperatestheRozahydropowerplant, locatedfive miles downstream the test site, which usesthe Roza Canal water flow to drive hydropowerturbines. The effect of deploying one or moreMHK turbines at the test siteon thehydropowerplant performance and operation is not yetknown. Deployment of a significant number ofturbinescanpotentiallycauseundesirableeffects,suchasdecreasingthehydropowerplant’spowergeneration. In addition, turbine operation canraise the water surface above the limit set byUSBR, which requires the water surface in thecanal to be at least 1.8’ below the canal’s

freeboard,tominimizetheriskofflooding.Whenthewatersurfaceexceedsthefreeboardlevel,thecanal’s water intake will need to be reduced,whichinturnwillalsoreducetheamountofwaterto be allocated for irrigation and the Rozahydropowerplant. Thispaperdescribesafieldmeasurementtestplan to investigate the dynamics of water levelandvelocityintheRozaCanalHKtestsiteduetothe deployment and operation of a vertical axisturbine. The plan will be executed during thesummerof2014. Preliminaryvelocityandwaterlevel measurements at the site, conducted insummer2013,arealsopresented.SITE DESCRIPTION  The Roza Canal test site is locatedapproximately 1.5 kilometers downstream of theinlet of the canal, which diverts water from thenearby Yakima River. Since August 2013,InstreamEnergySystems(IES)hasbeenoperatinga25kW3‐bladeverticalaxisDarrieus turbineatthesite.Theturbinehasarotordiameter(DT)of3mandarotorheight(HT)of1.5m,andisdeployedatthemid‐sectionofthecanal(Figure1),at1.5mabovethebed.Theturbineismountedonalargecylindrical platform that can be rotated 90degrees, to enable the turbine to be takencompletely out of the water when not operating(Figure 2). The turbine is deployed at a straighttrapezoidal section of the canal. During thepreliminary measurement in 2013, the watersurface width and depth at the turbine locationwere 12.5m and 3.3m (Figure 3). The turbineblockage ratio was ~16%. Both canal sidewallshave a slope of 1.25:1 (or 39 degrees fromhorizontal).Thecanalbedandsidewallsaremadefromconcrete.Thelongitudinalslopeofthebedis0.0004. The Froude Number during themeasurementswas~0.4. The canal flows are ofinterestfortestingbetweenthespringandfalldueto the high flows typically occur during thisperiod.Atypicalmeanvelocityinthecanalduringthis period is ~1.9 m/s. The canal flow issignificantly reduced in the winter and is notsuitableforHKturbineoperation.METHODOLOGY  A set of water level and velocitymeasurementswillbeconductedinthevicinityofthe turbine in 2014. Pressure transducers wereinstalledin2013atseverallocationswithin50mof the turbine, approximately 1 km upstream oftheturbineand2.5kmdownstreamoftheturbine,tomonitorwater levels during themeasurementcampaign. These measurements serve severalpurposes, to: 1) calculate the water levels indifferent locations in the canal,when the turbine

isdeployedandnotdeployed,2)monitortheflowsteadinessduringthemeasurementperiod,and3)calibrate and validate numerical models. Theresults of (1) will directly apply in determiningchangesinenergygradeline(EGL)andhydraulicgrade line (HGL) due to HK turbine operation,whichmaybeanalyzedusingtheenergyequation:

∙

2∙

2

= Bed elevation at an upstream location,

relativetoadatum(m)=Watersurfaceelevationupstream(m)=Corioliscoefficientupstream(‐)=Meanvelocityupstream(m/s)

g=Gravitationalacceleration(m/s2)= Bed elevation at a downstream location,

relativetothedatumusedfor (m)=Watersurfaceelevationdownstream(m)=Corioliscoefficientdownstream(‐)=Meanvelocitydownstream(m/s)=Minorlosses(m)=Frictionlosses(m)=Energyextractedbyturbine(m)=Headlossduetoturbineblockage(m)

The EGL is a straight line created from twopoints,atotalenergyattheupstreamlocationanda total energy at thedownstream locationminusall the energy losses, with x axis corresponds tothedistancebetween the two locationsandtheyaxis corresponds to the amount of energy. TheHGL is a straight line created from the watersurfaceelevationsatthetwolocations,withxaxiscorresponds to the distance between the twolocationsandtheyaxiscorrespondstothewatersurfaceelevations. Inflow velocity will be measured using anacousticDopplercurrentprofiler(ADCP).Severalcross‐sections and transects along the turbinecenterline, upstream of the turbine, will bemeasured to determine the extent of the flowalteration caused by the presence of the turbine.Theinflowmeasurementscanbeusedtoestimateresource availability and calculate the powercoefficient of the turbine. TheADCPwill also beused tomeasurewakevelocity field, for studyingthe wake flow dynamics and wake recoverydistance, an important parameter for turbinearray design. Measurementswill be collected atsixcross‐sections(CS)downstreamoftheturbineand along the turbine centerline. A remotelyoperatedsurveyboat,equippedwithaRealTimeKinematic rover‐base GPS system (horizontalpositioningaccuracywithin0.1m),willbeusedto

deploy an ADCP and an echo sounder forcharacterizing velocity and bathymetry over alargeareawithina fewhundredmeters fromtheturbine. This information can be used forinvestigating the effect of geometry changes dueto channel meandering and constriction on localand global velocity distributions, which isimportant for determining the optimal sitinglocationofadditional turbines. Thevelocityandbathymetry measurements can also be used fornumerical model boundary condition andvalidation. In addition to the water level and velocitymeasurements, the thrust force acting on theturbinewillbederivedfromstrainmeasurements.A set of coil strain gageswill bemounted to therotating turbine shaft. Themeasured signal willbetransferredtoadataacquisitionsystemusingawirelesstelemetrysystem.Then,thethrustforceactingontherotorwillbederivedfromthestrainmeasurement, shaft material properties, shaftgeometry and shaft dimensions [1]. Themeasured thrust force can be used to derive athrust coefficient, a critical input parameter fornumericalmodels to simulate theeffectofanHKturbinedeployment.Inthenextphase,thethrustforcewillbeusedtosimulatetheeffectofmultipleHKdevices and array configurations on the localwateroperation.

FIGURE 1. ROZA CANAL HK TEST SITE ANDBATHYMETRYATTHETURBINELOCALITY.

FIGURE 2. THE INSTREAM ENERGY SYSTEM’STURBINE,INOPERATIONALCONDITION.

FIGURE 3. A SCHEMATIC VIEW OF THE CROSS‐SECTION,SHOWINGTHEROTORSWEPTAREA.

PRELIMINARY RESULTS  A preliminary campaign to measure waterlevelandvelocitywasconductedinsummer2013toprovide informationon thevelocityandwaterlevel at a flow depth of ~3m. This informationwasused to help specs the sensors, e.g. a torquesensor, and design a cableway system for ADCPand acoustic Doppler velocimeter (ADV)deployments that will be used for the 2014measurements. Velocitymeasurementswereconductedusingan RDI ADCP StreamPro Mode 12 [2], with asampling frequency of 1 Hz. Themeasurementswere conducted at 12 cross‐sections; six cross‐sections are located upstreamof the turbine andsix cross‐sections are located downstream of theturbine. Cross‐section1isthefurthestupstream.The turbine is located between cross‐sections 6and7.Allofthecross‐sectionsareevenlyspaced10 m apart with the turbine being located mid‐waybetweenCS6andCS7suchthatthesecross‐sectionsarelocated5mupanddownstreamfromthe turbine, respectively. The locations of thesecross‐sectionsareillustratedinFigure4. Measurements were conducted for twoconditions: 1) without the turbine present(baseline),and2)withthe turbineoperatingataconstant rotation rate. The ADCP, attached to asmall catamaran, was traversed along a taglineacross the channel in each cross‐section (Figure5).TheADCPwastraversedataspeedof0.1m/s,orless,resultedinhorizontalbinsizesof0.1mor

less.Thesizeoftheverticalbinwassetto0.12m.Two transects were obtained for each cross‐section,whichissufficientforthepurposeofthispreliminary fieldwork. During the 2014fieldwork, four transects or more will bemeasured for each cross‐section, with a goal toobtain at least four transects with differences indischargeofequalorlessthan5%ofthemeanofall the samples. This approach is a standardprocedure that has been adopted by researchersandengineers,e.g.[3,4]. A flow discharge was calculated for eachcross‐sectionbyintegratingthevelocitiesoverthecross‐section area. Due to its measuringmethodology,anADCPcannotmeasurevelocitiesnear hard boundaries (bed and side walls) andnear the water surface, due to sidelobeinterference and blanking distance. Thedischargesintheseunmeasuredregionsareeitherinterpolated to the hard boundaries (to zero) orextrapolated to thewater surface level using themeasured velocities [5]. The dischargecalculations are performed within the datacollection software WinRiver 2 [3]. A validcomparison between conditions 1 and 2 can bemade if the flow discharges between those twoconditions are similar. A representative flowdischarge of the canal was calculated for eachconditionbyaveragingthedischargevaluesofthesix upstream cross‐sections. A dischargedifference of 1.5 % between the two conditionswasobtained. TheADCPvelocitymeasurementsupstreamoftheturbinelocation,withandwithouttheturbineinthewater,aregraphicallyrepresentedinFigure6,whilethesamedatadownstreamoftheturbinedeployment location are shown Figure 7. ThemissingdatainCS5,seenasverticalwhitestreaksor gaps in the contour plots, is likely caused byvegetation that growsat thebottomof the canal,causing poor acoustic signal readings. ThemissingdatainCS7thatincludedtheturbinewasa result of the ADCP transducer losing contactwiththewater.Thestrongwatersurfacewaveinthe near‐wake caused unstable ADCP boatmovement. This condition will be improved byutilizing a larger ADCP boat and a cablewaysystem that can stabilize the boat during highvelocity and roughwater conditions, such as theoneoutlinedin[6]. The turbine seemed to have little effect onupstreamvelocitiesastheupstreamcrosssectionshavesimilarvelocitydistributionsinabsenceandpresence of the turbine. It is interesting to notethat, for all cases, there is a high velocity core atthe right part of the CS. Upstream of themeasurement site, the canal curves to the left(Figure 1), which causes superelevation of the

water surface on the right side at themeasurementlocationandshiftsthehighvelocitycore to the right side downstream of the bend.Secondary flow circulation caused by centrifugalacceleration,whichisoftentermedasthePrandtlsecondaryflowofthefirstkind,isknowntooccurat bends [7]. TheADCPmeasurements from thepreliminary campaigns show an indication ofsecondary flow cells at the cross‐sectionsupstreamoftheturbine.Itisstillunclearifthesecellsarerelatedonlytothesecondaryflowatthebend,oralsoinfluencedbytheturbulencedrivensecondaryflowthat isknowntooccur instraightsectionsofachannel[7,8],suchas thelocationswhere the ADCP measurements were taken.Further analyses, such as the spatio‐temporalaveraging of ADCP transects [9‐11], will beconductedoncemoremeasurementsareobtained.

Movingtheturbinetothehighvelocityregionmay significantly increase the rate of powergeneration due to the cube relationship betweenvelocityandpower.Movingtheturbine,however,may not be feasible, because the high velocityregion has a shallower depth than the currentturbinelocation,whichmightresultininadequateclearance between the turbine and channelbottom. Nonetheless, it is important to conductsite‐specific full‐cross‐section velocitymeasurement, such as the moving‐boat ADCPmeasurement, to identify local hot spots, whendesigning the deployment strategy for ahydrokineticturbine. Furthermore, the baselinemeasurements forthe upstream and downstream CS show similarvelocitydistributionsbetween the cross‐sections,with the exception of that at cross‐section 11,which has a longer width than the other cross‐sections. Theonlycross‐sectionmeasureddownstreamoftheturbine,whentheturbinewaspresent,wasCS 7. Only ¾ of cross‐section 7 was measuredbefore the ADCP boat capsized due to thepresenceofstrongwavesatthislocation.Despitethis mishap, the measurement results areencouraging. The ADCP was able to capture thevelocity deficit in the near‐wake region. Thequality of the measured data appears to beacceptable as indicated by the values of signalintensity and correlation, which are very similarto thosemeasuredat theothercross‐sections. Itis expected that, with improving ship keeping,future measurements will be attainable in thenearandfarwakeregion. Waterlevelsweremeasuredevery20secondsatcross‐sections1–12foraperiodofthreedaysduring the field measurement campaign. USBRrequireswater levelmeasuredat the edgeof thecanal for comparison with freeboard level,

therefore,thewaterlevelsensorswereplacedattheedgeofthecross‐sections.Figure8showsthetime‐averagedwater levelmeasurements for thebaseline and with turbine cases. Bothmeasurements were time‐averaged over 1.5hours. The flow conditions were relativelystationary during the three‐day fieldmeasurement campaign, indicated by relativelyconstant water levels during this period at themost upstream cross‐section (cross‐section 1).Duringthetestperiod,thewater levelchangesatcross‐section 1 never exceeded 0.05 m for thebaselinescenarioand0.03mforthewithturbinescenario. Figure9compareswaterlevelmeasurementsin the presence and absence of the turbine. Anincreaseinwaterlevelisindicatedwithapositivedifference value and vice versa. These resultsshow that adding the operating turbine in thecanalincreasedthewatersurfaceintheupstreamcross‐sections by a constant difference ofapproximately0.03m,withtheexceptionofCS5(40 m downstream of CS1). Excluding the CS 5measurement,theHGLsforboththebaselineandwith turbine cases upstream of the turbine areclosetolinear.ThemeasurementatCS5seemstocontain a significant error, as indicated by itsmuchlowermeasuredwaterlevelthanthoseatCS4andCS6. Thiserror ispossiblycausedbythevegetation growth at the bottom of the channelthat interferes with the pressure transducermeasurement. Downstream of the turbine, thewater level decreased by 0.05 m at CS 7. Thewater level difference decreases with distancefromtheturbine,anddiminishesatapproximately40mdownstreamof the turbine,or13.3 turbinerotordiameters.Thesemeasurementsweremadeat one canal flow speed with one turbinerotationalspeed. Future fieldmeasurementswill

include additional flow and turbine conditions,whichwillprovideamorecompletepictureontheeffect of deploying a hydrokinetic energy turbineinacanalsystem.

FIGURE 4. THE LOCATIONS OF THE TEST CROSS‐SECTIONS,RELATIVELYTOTHETURBINE.

FIGURE5.ADCPTESTMEASUREMENTS INAUGUST2013.

FIGURE 6. CONTOURS OF VELOCITY MAGNITUDE (LOOKING DOWNSTREAM) AT SEVERAL CROSS‐SECTIONSUPSTREAMOFTHETURBINE LOCATION. CS1 ISTHEMOSTUPSTREAM SECTION. THETURBINE IS LOCATEDBETWEENCS6ANDCS7.

FIGURE 7. CONTOURS OF VELOCITY MAGNITUDE (LOOKING DOWNSTREAM) AT SEVERAL CROSS‐SECTIONSDOWNSTREAMOFTHETURBINELOCATION.THETURBINEISLOCATEDINBETWEENCS6ANDCS7. CS7WITHTURBINE CONTOUR ONLY SHOWS MEASUREMENTS TO UP TO ¾ OF THE CS LENGTH DUE TO THE BOATCAPSIZING.MEASUREMENTSWITHTHETURBINEOPERATINGWERENOTSUCCESSFULATCROSS‐SECTIONS8‐12DUETOTHEBOATCAPSIZING.

FIGURE 8. TIME‐AVERAGED WATER LEVELMEASUREMENTS FOR THE BASELINE AND WITHTURBINE CASES. USBR DATUM IS USED AS AREFERENCE.

FIGURE 9. WATER LEVEL DIFFERENCE BETWEENTHEWITHTURBINEANDBASELINECASES.

CONCLUSIONS  AcomprehensivefieldmeasurementtestplantoinvestigatetheeffectofHKturbinedeploymentin a canal test site was proposed based on apreliminary observation of velocity and waterlevelmeasurementsaroundaverticalaxisturbinedeployed at the site. The test plan includesdetailed velocity measurements to characterizeinflow and wake flow, and water levelmeasurements to quantify the effect of turbineoperation on water level, energy grade line andhydraulic grade line. The test plan will beimplemented at an HK energy test site in RozaCanal, Yakima, WA, USA, in mid‐2014. It isexpectedthat themethodsused inthisstudywillhelp industry and regulators evaluate the impactofHK turbinedeploymentandapplyappropriatemitigationmeasuresforthenegativeimpactsthatmayoccur.ACKNOWLEDGEMENTS  This research was supported by the U.S.Department of Energy’s (DOE) Office of Energy

Efficiency and Renewable Energy, Wind andWater Power Technologies Program. SandiaNationalLaboratories,amulti‐programlaboratorymanagedandoperatedbySandiaCorporation,isawholly owned subsidiary of Lockheed MartinCorporation, for the U.S. Department of Energy'sNational Nuclear Security Administration undercontractDE‐AC04‐94AL85000.REFERENCES [1] Omega Engineering Inc., 2014, “Positioningstrain gages tomonitor bending, axial, shear andtorsionalloads.”http://www.omega.com/faq/pressure/pdf/positioning.pdf[2] Teledyne RD Instruments, 2013, “StreamProADCPShallowStreamflowMeasurementSystem,”Datasheet.(http://www.rdinstruments.com/datasheets/streampro_ds_lr.pdf).[3] Teledyne RD Instruments, 2014, “WinRiver 2User’sGuide,”RDInstruments,Poway,CA,USA.[4] Mueller, D.S. and Wagner, C.R., 2009,“Measuring discharge with acoustic Dopplercurrent profilers from a moving boat.” USGeologicalSurveyTechniquesandMethods3A‐22,72p.USGS:http://pubs.water.usgs.gov/tm3a22[5] Teledyne RD Instruments, 1996, “AcousticDopplercurrentprofiler,principlesofoperation,apracticalprimer.”TechnicalManual.[6] Gunawan, B., Sterling, M. and Knight, D.W.,2010,“UsinganAcousticDopplerCurrentProfilerinasmallriver,”WaterandEnvironmentJournal.Vol.24no.2pp147‐158.[7] Perkins, H.J., 1970, “The formation ofstreamwisevorticityinturbulentflow,”Journalof FluidMechanics, Cambridge University Press,44(04):721‐740.[8] Einstein, H.A. and Li, H., 1958, “Secondarycurrentsinstraightchannels.”TransactionsoftheAmerican Geophysical Union, Vol. 39, No.6, pp1085‐1088.[9] Szupiany, R.N., Amsler, M.L., Best, J.L. &Parsons, D.R., 2007, “Comparison of fixed‐ andmoving‐vesselflowmeasurementswithanaDPina large river,” Journal of Hydraulic Engineering,ASCE,pp.1299‐1309[10]LeCoz,J.,Gilles,P.,Magali,J.&Andre,P.,2007, “Mean vertical velocity profiles fromADCP river discharge measurement datasets,”Proceedings 32nd Congress of InternationalAssociation of Hydraulic Engineering andResearch,Venice,Italy.[11]Gunawan,B., Sun,X., Sterling,M., Shiono,K.,Tsubaki, R., Rameshwaran, P., Knight, D.W.,Chandler, J., Tang, X. and Fujita, I., 2012, “TheapplicationofLS‐PIVtoasmallirregularriverfor

inbankandoverbank flows,“ FlowMeasurementandInstrumentation,Volume24,pp1–12.