1ComparisonofApproaches(SESTLC,ROW&HIFREQ)forACInterferenceStudy

1 ComparisonofApproaches(SESTLC,ROW&HIFREQ)forAC InterferenceStudy
YexuLiandSimonFortin

1.1 Introduction
Three independent tools are available in the SES software packages to carry out an AC interference study: SESTLC, ROW (TRALIN/SPLITS) and HIFREQ. To help select the best program for your study, it is important to be aware of the advantages and limitations of the differenttools. The objective of this articleis to have an overall look at those three tools and illustrate their advantages,limitations,andapplicability.

1.2 TheFundamentalsbehindtheThreeTools
SESTLC:SESTLCisbasedonacircuitapproach.Lineparameters(lineconstants)arecomputed usingthemethodoriginallydevelopedinFCDIST.Thebundlereduction,groundwireelimination and sequence components algorithms are derived from the TRALIN module. It is a simplified analysis tool to quickly estimate the line parameters, magnetic and electric fields of arbitrary configurations of parallel transmission and distribution lines in uniform soil. It can also computetheinductive,conductiveandtotalinterferencelevelsonametallicutilitypathsuchas pipelineorrailwaythatrunsparalleltotheelectriclinesinuniformsoil.SESTLCprovidesa veryfastandefficienttoolforestimatingthemagnitudeoftheACinterferenceforsimpleright ofwayconfigurations. ROW(TRALIN/SPLITS):ROWisalsobasedonacircuitapproach.Fromaspecificationofthe location and characteristics of the physical elements of a system (phase and shield/neutral wires, pipelines, soil characteristics, etc), the program derives an equivalent circuit model. The line parameters for the circuit are computed using line constant formulas originally developed in TRALIN. Lumped elements such as ground impedances of substation grids and transmission tower grounds can be added to the circuit. These are normally computed using appropriategroundingtools,suchasMALTorMALZ. This circuit model is then solved using SPLITS to yield the inductive interference component. Next, a MALZ model is built. This model provides an option to include an EMF term in the energization oftheconductors (theEMFisusedto account fortheinducedeffects on avictim line).Asaresult,theMALZmodelgivesthetotalinterferencelevel. ROW is flexible and fast for detailed designs involving complex rightofway network configurations.Themain approximations in the program aremade inthecomputations of the line parameters, which assume that (1) conductors are parallel to each other and (2) conductorsareinfiniteinlength. ROWsolvesthefirstproblembyrepresentingnonparallelconductorsusingaverageseparation distancesforeachsectioncuttocomputetheinductivecoupling.Thisisagoodapproximation aslongasthesectionsaresufficientlyshort. The second approximation will cause noticeable inaccuracy at current discontinuity locations, suchasafaultlocationandlineendsetc. HIFREQ: HIFREQ is based on a fieldtheoretical approach which solves Maxwells equations directly. HIFREQ models the complete conductor network under consideration in three dimensionalspace,andaccommodatesangledconductorswithoutmakinganyapproximations. The inductive, capacitive and conductive interference effects between all the elements in the network are simultaneously taken into account. Therefore, HIFREQ calculates the total
1

interferencelevelaccurately.However,computationtimescanbeconsiderablyhighercompared toSESTLCandROW,forcomplicatednetworks.

1.3 ComparisonoftheThreeTools
Inthissection,computationresultsobtainedwiththethreetoolsdescribedabovearepresented andcompared.Thestudyisbasedonareferencecomputermodelfromwhichseveralseriesof simulationsarecreatedbyvaryingoneormoreparametersatatime.Thereferencecomputer modelconsistsofasinglephaseconductorofatransmissionline(T/L)andapipeline(P/L).The heightoftheT/Lis10mabovetheearthsurface.ThecenteroftheP/Lis2mbelowtheearth surface.Itsouterandinnerradiiare0.2mand0.19m,respectively.Therelativeresistivityof theP/L wall is 12 andits relative permeabilityis 250.TheP/L is well coated,with a0.005 m coating thickness and 3,048,781 Wm coating resistivity. The soil resistivity is 100 ohmm. A 1000 amp current is assumed to be flowing in the T/L phase conductor under steady state conditions.Underfault conditions,acurrent of 1000 ampis assumedto beflowingfrom both sidesofthefaultlocation.Thebasecasesectionlengthis100m. Thefollowingcasesareexaminedundersteadystateconditions: Parallel:TheT/LandP/Lareparallel. NonParallel:TheT/LandP/Lareataconstantangletoeachother. Crossing:TheP/LabruptlycrossesovertotheothersideoftheT/L. Phase Transposition: In this variation, a balanced threephase T/L system with a transpositionsomewherealongtheline. FortheNonParallelstudy,theP/Lisplacedatanangleof15degreeswithrespecttotheT/L. TheT/L(andassociatedpipelinesections)issubdividedintosectionlengthsof50mand100m, respectively, in order to estimate the sensitivity of the results to the subdivision of the ROW intoquasiparallelsections. Under fault conditions, sensitivity tests have been made for the following parameters: fault locations,parallellengths,andseparationdistancesbetweenthepipelineandthetransmission line. Note that for the conductive interference, SESTLC represents tower foundations and other sourcesofconductioncurrentaspointsources,whichisagoodapproximationwhenthevictim line(e.g.pipelineorrailway)isrelativelyfarfromtheenergizationsource.ROWusestheMALZ approach, and can model tower or substation grounding systems as is, instead of as a point source.BothSESTLCandROWtakeintoaccountthevoltagedropalongthevictimconductors. On the other hand, SESTLC and ROW do not account for the inductive and capacitive interactions between conductors of the grounding system, while HIFREQ does. The difference caused by this interaction is generally small for most rightofway networks. However, significant differencescan beobtainedaroundasubstation area,in whichcaseuseofHIFREQ couldimprovethecomputationaccuracy[seeRef.3]. The effect of the interaction between conductors is beyond the scope of this article. The comparisonsmadeinthisarticleapplyonlyfortheinductiveinterference. Theinputfiles that areusedinthis article areavailable onyour CDROM,inthefolderUsers Group2005\InputFiles\ComparisonsofTLCROWHIFREQ\.

1.3.1 SteadyStateConditions

1 ComparisonofApproaches(SESTLC,ROW&HIFREQ)forACInterferenceStudy

ComparisonofTLC,ROW&HIFREQ
LoadCondition(1kA),SinglePhase,Parallel(2km) 350 InducedPipeGPR(V) 300 250 200 150 100 50 0 1500 1000 500 0 500 DistancealongPipeline(m) 1000 1500
HIFREQ:2kMTL HIFREQ:40kMTL ROW TLC

Objective: Examine asimpleparallelnetwork case,anddemonstratethe lengtheffect. Network: LoadCurrent:1kA PhaseWires:Single,2km(and40kmforHIFREQ) ShieldWires:No Comments: Pipeline:2km HorizontalSeparationDistance: 0

The figureshows that thereare novisible differencesbetweentheresultsfromSESTLC & ROWand HIFREQapproaches whenthe modeled transmissionlineis longenough(40km)inHIFREQcomparedtothepipelinelength(2km).Inthiscase,theassumptionthatthetransmission lineisinfiniteisessentiallycorrect[seeRef.1].

TL PL

ComparisonofTLC,ROW&HIFREQ
15
0

200 InducedPipeGPR(V)

LoadCondition(1kA),SinglePhase,15Deg. CrossingatCenterofROW

150

100

50

0 0 500 1000 1500

HIFREQ:10m Seg. ROW:50m Seg. ROW:100m Seg. 2000 2500

DistancealongPipeline(m)
Objective: Examine asimplenonparallelnetwork case,anddemonstratethe sensitivity tothe section length. Network: LoadCurrent:1kA PhaseWires:Single,2km forROW(and40kmforHIFREQ) Comments: For a nonparallel rightofway network, decreasing the subdivision length (i.e., section cut length) in ROW can improve the accuracy of the results.Thisisbecausetheaverageseparationdistancesbetweenconductorsarebetterapproximatedwithashortersectionlength.However,an extremelysmallsubdivisionlengthcancausenumericalerrorsinthecircuitapproachcomputations.Asubdivisionlengthof50musuallyisquite adequateforasteadystateconditionstudy.Inthiscase,resultsarenotpresentedforSESTLC,sinceitisdifficulttorepresentthissystemasan equivalentparallellinenetwork [seeRef.2]. ShieldWires:No Pipeline: 2 km, 15 degree to TL, cross at the center of ROW

1 ComparisonofApproaches(SESTLC,ROW&HIFREQ)forACInterferenceStudy

TL

ComparisonofROW&HIFREQ:
LoadCondition(1kA),ThreePhases, Parallel&CrossingatCenterofROW 12 InducedPipeGPR(V) 10 8 6 4 2 0 1500 1000 500 0 500 1000 1500 DistancealongPipeline(m)
HIFREQ ROW

PL

Objective: Examine asimple 3phasesystem withaPLcrossing. Network: LoadCurrent:1kA PhaseWires:Threephases,40km Comments: For balanced threephase currents (steadystate condition), with no shield wires, the discontinuityin pipelineGPRthat iscausedbyapipeline crossingatransmissionline isnearlyidenticalwhencomputedwith HIFREQand ROW.ThisisbecausethesourceofEMF(thetransmissionline) isuninterruptedinthiscase.Hence,thetransmissionlineeffectively looksinfiniteinHIFREQ(asinROW). Inthiscase,resultsarenotpresentedforSESTLC,sinceitisdifficulttorepresentthissystemasanequivalentparallellinenetwork. ShieldWires:No Pipeline:2km,Paralleltotransmissionline,CrossatthecenterofROW

ComparisonofROW&HIFREQ: LoadCondition(1kA),ThreePhase, Parallel&PhaseTranspositionatCenterofROW 16 InducedPipeGPR(V) 14 12 10 8 6 4 2 0 1500 ROW HIFREQ 1000 500 0 500 DistancealongPipeline(m) 1000 1500

Objective: Examine aphasetransposition case. Network: LoadCurrent:1kA PhaseWires:Threephases,40km,PhasetranspositionatthecenterofROW ShieldWires:No Pipeline:2km,Paralleltotransmissionline Comments: For balanced threephase currents (steadystate condition), no shield wires and with a transmission line discontinuity caused by a phase transposition,theresultsobtained usingHIFREQand ROWarevery similar. Inthiscase,resultsarenotpresentedforSESTLC,sinceitisdifficulttorepresentthissystemasanequivalentparallellinenetwork.

1 ComparisonofApproaches(SESTLC,ROW&HIFREQ)forACInterferenceStudy

1.3.2 UnderFaultConditions
ComparisonofTLC,ROW&HIFREQ:
FaultLocation: Ce nte rofROW HorizontalSe parationDistance :0 ROW/TLCMaxError:27.3% 200 ParallelLength: 2k m FaultCurrent: 2kA

InducedPipeGPR(V)

160 120 80 40 0 1500 HIFREQ ROW TLC 1000 500 0 500 1000 1500

DistancealongPipeline(m)

Objective: Examine asimple symmetricfaultcase. Network: FaultCurrent&Location:2kAatcenterofROW PhaseWires:Single,40km ShieldWires:No Comments: Themaximumdifferencesbetweenthefieldapproach(HIFREQ)andthecircuitapproach(SESTLC&ROW)resultsoccuratthefaultlocation, whilethesedifferencesarelessimportantatthepipelineends.Thesedifferencesareduetothemagneticfielddiscontinuityatthefaultlocation andatthepipelineends.Thisdiscontinuityislargeratthefaultthelocationthanattheendsoftheline,sincethecurrentjumpsfrom1Ampto1 Ampatthefaultlocation,whileitjumpsfrom1Amptozeroattheendsoftheline.Thismagneticfielddiscontinuityisnottakenintoaccountin thecircuitapproach,whichassumesaninfinitelinelengthwhencomputingthelineparameters. Pipeline:2km,Paralleltotransmissionline SeparationDistance:0 MaximumROW/SESTLCError:27.3%fora2kmexposurelength.

ComparisonofROW&HIFREQ:AsaFunctionofFaultLocations
Paralle lLength: 2km FaultCurrent:2k A 30 HorizontalSe parationDis tance: 0

AtFaultLocation
25

AtPipeEnds(MaximumGPR)

ROWError(%)

20 15 10 5 0 5 0 500 1000 1500 2000 2500

FaultLocationfromCenterofROW(m)

Objective: Examine theeffectsofthefaultlocation onthe computationaccuracy. Network: FaultCurrent&Location:2kAthelocation variesbetweenthecenteroftheROWand2.2kmawayfromit PhaseWires:Single,40km Separation Distance:0 ShieldWires:No MaximumROW/SESTLCError:27.3%. Pipeline:2km,Paralleltotransmissionline Comments: Thelargestdifferencebetweenthefieldapproach(HIFREQ)andcircuitapproach(ROW,SESTLC)isobservedforafaultlocatedmidwayalong theROW.

1 ComparisonofApproaches(SESTLC,ROW&HIFREQ)forACInterferenceStudy

ComparisonofROW&HIFREQ: AsaFunctionofParallelLength
FaultLocation: Center FaultCurrent: 2kA HorizontalSeparationDistance: 0

45 40 ROWError(%) DeltaGPR(V) 35 30 25 20 15 10 5 0 0 5 10 15 ParallelLength(km) 20 25 ROWErrorin%atFaultLocation DeltaGPRatFaultLocation

Objective: Examine theeffectsofthe lengthofparallelismonthecomputationaccuracy. Network: FaultCurrent&Location:2kACenterofROW ShieldWires:No PhaseWires:Single,40km SeparationDistance:0 Pipeline:Paralleltotransmissionline,ParallelLengthvaryingbetween2and20km Comments: The difference between the induced pipeline potential calculated using the circuit approach (ROW) and the results using the field approach (HIFREQ)decreaseswithincreasingexposurelength.Thedifferenceisnegligiblewhentheexposurelengthisconsiderable(lessthan5%when theexposurelengthisabout10km).Whentheparallelismissignificant,eventhoughtheendeffectatthefaultlocationstill exists,thecircuit approachgivesaccurateresultsbecausethetotalinducedpotentialispredominantlyduetotheinductivecouplingalongthewholelength,thusthe presenceofan abruptelectromagneticfielddiscontinuityatthefaultlocationdoesnotaffecttheresultssignificantly.

ComparisonofROW&HIFREQ:AsaFunction ofRatioofSeparationDistancetoParallelLength
FaultLocation: Center 250 200 ROWError(%) 150 100 50 0 50 0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.2 0.22 2kmParallelLength,AtFaultLocation 10kmParallelLength,AtFaultLocation 2kmParallelLength,AtPipeEnd 10kmParallelLength,AtPipeEnd FaultCurrent: 2kA

RatioofHorizontalSeparationDistancetoParallelLength

Objective: Examine theeffectsofthe separationdistanceonthecomputationaccuracy. Network: FaultCurrent&Location:2kACenterofROW PhaseWires:Single,40km ShieldWires:No Pipeline:Paralleltotransmissionline,2km SeparationDistance:AsaFunctionof theRatioofSeparationDistancetoParallelLength Comments: The difference between the induced pipeline potential calculated using the circuit approach (SESTLC, ROW) and the results using the field approach(HIFREQ)increaseswithincreasingratiobetweenthehorizontalseparationdistanceandpipelineparallellength.Thisisduetothefact thatasthedistancebetweenthepipelineandthetransmissionlineincreases,themagneticdiscontinuitiesatthefaultlocationandattheendsofthe lineaffectsalargerandlargerportionofthepipeline

1 ComparisonofApproaches(SESTLC,ROW&HIFREQ)forACInterferenceStudy

1.4 Conclusion
This article compared the performance of three tools (SESTLC, ROW, and HIFREQ) when computing AC interference effects caused by transmission lines. The main advantages and limitationsofthethreetoolsarelistedbelow. SESTLC provides a quick estimate of ac interference level for simple rightofway configurationsundersteadystateandfaultconditions.However,formoredetailedstudiesor for a final design involving complex configurations, the ROW or HIFREQ modules should alwaysbeused. ROW is fast and flexible. It can provide accurate results under steadystate conditions for most rightofway network configurations: parallel, nonparallel, crossing, phase transpositions,etc.Thedifference between ROW andHIFREQunder steadystate conditions is usually small, especially when the transmission line is sufficiently long. Decreasing the sectioncutlengthcanimprovethecomputationaccuracyfornonparallelcases. Under fault conditions, the error, which is most noticeable at the fault location, can be importantforsomenetworkconfigurations.ThepredictionsfromROW,however,areusually conservative. Furthermore,ROWcanmodeldifferentsoilmodelsalongtherightofway.Italsoallowsthe userto automatically createfaults alongany transmissionline, at any given intervals (with theMonitorFaultmodule).Finally,itgeneratessummaryfilescontainingcertaininformation about the victim phase conductor (i.e., pipeline/railway maximum GPR, maxim railtorail voltages, GPR at fault locations etc.), a reference phase conductor (i.e.: tower injected currents)andphaseconductors(i.e.,faultcurrentsforanyfaultlocation). HIFREQ offersthemost accuratemodel,andaccounts correctlyforthe finite length ofthe transmission line conductors. It is also considerably simpler to specify the input data for HIFREQ.However,the program can betimeconsuming, especially for complicatedrightof way configurations. In addition, HIFREQ can only model one soil model along the entire rightofwayunderstudy.

1.5 References
[1]Y.Li,F.P.Dawalibi,andJ.Ma,EffectsofConductorLengthandAngleontheAccuracyofInductive InterferenceComputations,TransmissionandDistributionConferenceandExposition,IEEE/PES2001, Atlanta,UAS,October28November2,2001. [2]Y. Li, F. P. Dawalibi, and J. Ma, Effects of Conductor Angle between Transmission Lines and Neighboring Utilities on the Accuracy of Inductive Interference Computations, PowerCon2002, Kunming,China,November2002. [3]R. Southey, F. P. Dawalibi, Y. Li andW. Yuan, Increasing the CostEffectiveness of AC Interference MitigationDesignswithIntegratedElectromagneticFieldModeling,Corrosion/2005NACEInternational Conference,Huston,USA,March2005.

Page11