8/12/2019 01432444

1/6

2 4 IEEElPES Transmission Distribution Conference Exposition: LatinAmericaI

Adaptive distance protection for SeriesCompensated Transmission LinesRicardo Dum Luis Fabiano Wanner Oliveira Murari Mohan Saha Stig Lidstrom

FURNAS, BRAZIL A B B ,BRAZIL

Abstract : A new high speed distance protect ion sche me forsingle l ine to ground fau l t s and mul t i phase fau l t s fo r EHV(Extra High Voltage) transmission lines is described in thispaper. This scheme also performs well on series compensatedlines. The new complem enta ry fas t t r ipp ing a lgor i thms toge therwi th the ear l i e r exper ienced a lgor i thm and multi-processorbased distance algori thm, have led to the deve lopment to a nadapt ive hybr id scheme. This scheme i s implemented in to awel l -p roven hardw are p la t fo rm wi th modera te requ i rements onthe com munica t ion . The pro tec t ion scheme prov ides h igh speedt r ipp ing less than one cycle) and high speed signall ing.Extensive test ing by using EMTP s imula t ions as well as realt i m e s imula t ion wi th a power sys tem s imula tor have conf i rmedthe va l id i ty o f th i s hybr id concep t as a fas t , secure anddependable p ro tec t ion scheme. The scheme is in operat ion forthe protect ion of Ser ies Com pensa ted Lines n some 25 count r iesa ro t lnd the world.Keywords: distanc e relay, high speed hybrid concept , seriescompensated lines, esthg

I. INTRODUCTIONDistance relays c n benefit from ideas in the newlydeveloped field of adaptive protection, and can offer an evenmore selective and sensitive form of protection, under avariety of system configurations. The benefits of installingseries capacitors in the power system include increasedpower transfer capability, improved power system stability,reduced system losses, improved voltage regulation, and thepossibility to regulate power flow. InstalIation of seriescapacitors, however, introduces challenges to protectionsystems with regards for both the series compensated linesand the adjacent lines [1,2]. Based on the principle ofsuperimposed transient signals outlined in [3], the first

t r ave l l ing wave relay was developed for commercial use.This relay satisfied the ultra high speed requirements for onecycle fault clearances and has been in service for many years.The drawback however is that it is not based on a continuousmeasuring a l g o r i t h m . In order to overcome drawback, theabove mentioned method has been combined with animpedance measuring method. It is, not possible to use onlyone algorithm but rather a hybrid solution with parallel andadaptive algorithms has to be implemented.

0-7803-8775-9/04/ 20.00 02 00 4 IEEE

ABB, SWEDEN

This paper explores such a hybrid solution in order toachieve high speed of operation while maintaining highdependability and security.Verification of the fast measuring algorithms is done byusing EMTPiATP [4], as well as by real time simulation witha power system skulator [ 5 ] and with a real time DigitalTransient Network Analyser [6]. The results are presented inthis paper.

11. DESCRIPTION OF THE NEW PROTECTIONSCHEMEFig. 1shows the schematic diagram of the new protection

that consists of t h e main distance protection, a fast trippingalgorithm for multi phase faults, and a fast hipping algorithmfor single phase to ground faults. Besides that there are alsodifferent blocks representing the AID converter circuit boardand the logical output CPU. The binary outputs Trip andDirection are respectively the phase selective tripping andthe directional information to be used in the communicationscheme.

581

ast npping alwonthmfor muhi phase faults

ZFPP1 Fast tripping algorthm Ifor single phase toground laulb ZFPN

Fig. 1. Block diagramof he new protection scheme.A Main distanceprotection

The main protection function is a fill scheme distanceprotection with three impedance measuring zones, having aquadrilateral characteristic [ 7 ] as shown in Fig. 2. Thesetting for each zone are independent for: reactive andresistive reach, resistive reach for single phase to ground and

8/12/2019 01432444

2/6

multi phase faults, zero sequence compensation factor anddirectionally of all zones.X

Fig. 2. Fault detection elements and adaptive expandingcharacteristic.A filter described in [ 7 j gives initially an underestimationof the current, which increases security o f the scheme. Thecomparison of the currents and voltages gives an impedancecircle (small circle in Fig, 2) and the operating time isshorter. The app arent characteristic will increase (large circle

in Fig. 2) when the filter factors are adjusted towards anarrower bandwidth and as the estimate of the fault currentgrows.B . heory of operution

The high speed h c t i o n is achieved by measuring thethree phase to ground loops as well as the three phase tophase loops. The trip, in case of three phase faults is issuedby the phase to phase measuring loops (operation of one ofthe loops is sufficient for a three phase trip).The measurement is of the full scheme type, the threephase to ground loops and the three phase to phase loops arecalculated in parallel. In proposed protection scheme, a newset of samples is issued every ms. All calculations arerepeatedly performed on each new set of samples and aresult is available every ms. The trip as well as the carriersend function require that the operation criterion has been

fidfilled during a number of calcdation. The results areaccumulated in a trip counter.C. Basic charucteristic

The characteristic can be described by the Fig. 2.Due tothe transient character of the measuring principle staticmeasurement can not veri@ the Characteristic. Dynamically,it can be verified that wt n the accuracy no operation willoccur outside the characteristic.

~

582

The different measuring criterion can be identified in thecharacteristic and all of them have to be fulfilled foroperation.The characteristic is principally identical for all type offaults. The reactive and resistive reach settings are differentfor the phase to ground and the phase to phase measuringloops.

D. ast tripping algorithm or multiphasefauh ZFPP)The main interest is to gain speed at multi phase faults,close to the relaying point. For such faults, the fault currentsare quite high. This provides good measuring conditions andenables faster measurements. This has been achieved by acomparison of the estimated filtered current amplitudes withthe corresp onding voltage amplitudes, such that

I k * U ,where Xkel is a set reactance, I, U are the estimated currentand voltage amplitude respectively and k is a constantsuitable for fast tripping.

The presence of zero sequence current preventsunselective three phase tripping at ground faults, wheresingle phase tripping is to be provided by some other means(as for example by using the ZFPN scheme).The fast tripping algorithm func tion utilizes fault quantitiesw i t h o n e ha lf cyc le in order to avoid overreach duringsubsequent breaker opening transients and other phenomena.This function i s supervised with an instantaneousmeasurement of the phase-to-phase current before an output

is issued.E. Fast tripping algorithm for single phase t ground faults

This algorithm presents a new approach to the distanceprotection with adaptive features. It consists of three mainparts. One part is for the determination of the h e c t i o n of afault utilizing superimposed currents and polarized voltages.The directional finction provides a very high speed phaseselective signal for drectiona l comparison,The second part is the determination of the faulty phases.

T h e method of phase selection uses a novel techniqueutilizing the pre-fault quantities of the voltages and currents.The pre-fault quantities are ob tained with help of informationfrom the healthy phases. The phase selection is the mostimportant function to prevent unwanted operation in theunfaulted phases.The third part is the determination of the fault loopimpedance. This is done by using the algorithm ( Z l )measures the impedance of the faulted loop by utilizing theline model, neutral model and the fault model. Correlation

8/12/2019 01432444

3/6

3

and adaptive filter techniques are used in order to improvemeasuring condition.The fast tripping algorithm for the single phase to groundfault m odule, ZFPN, will be described in more detail below.In the equations that follow, sample values will have lowercase letters, while amplitudes will have upper case letters.The amplitudes are calculated from voltage and currentsamples and their deriva tives. Superscript is used todefine the time derivatives.The ZFPN scheme is a further development of theprinciples introduced in Ultra High Speed protectionsdescribed in references [3, 81. The philosophy of the hybridsolution is to enable the fast tripping algorithm to provide ahigh security and let the main distance protection algorithmimprove the dependability. The ZFPN is optimized to give ahigh speed operation in case of single phase to ground faultsand is only activated by the following ground faultconditions :- for currents inall phases 1 > 0.40 rated enl andf, 0.20 m a x i o f d if ferences of phase currents

influence of load current. The change in current is calculatedby substracting h faulted phase current from the respectivepre-fault value.An impedance is calculated for each phase. As forexample, for the phase L1, the impedance is calculated byutilizing:

where U L ~ 3 is the different of voltages between healthyphases L2 and L3, and di~l s the change of current in phaseL1.For the detection of fault in forward direction theargument for the calculated impedaflce shall be within -15to 115 degrees.Fast zone 1 Z I ) - A zone fault will be indicated if themeasured impedance is within the rectangular characteristic

defmed by the zone 1 settings and if the following conditionis satisfies:where f is the amplitude of the sum of the phasecurrents p * ( ~ l , , ~ x > R 0se

- the phase selective function can define only one faultedphase. The measurement is performed during a timewindow of 15 msAs a next step, pre-fault load current and voltage values ina faulted phase are calculated from currents and voltages inunfaulted phases. These values are calculated by usingderivatives of current and voltage samples. The pre-faultvaIues are used to calculate superimposed voltage and current(A-quantities) in the faulted phase. The phase selection,directional, and distancemeasurement parts will be describedbelow.Phase selection (p) - The phase selection algorithm iscommon for both single phase to ground and multi phasefaults. The phase selection is performed by the comparison ofchanges in the phase to phase currents between phases L1

and L2, L2 and L3, and L3 and L1. The changes in the phaseto phase currents are obtained by substracting the actual faultcurrents with respective pre-fault quantities. The quantities(changes of currents) should be above certain operationlevels in order to indicate the faulted phase. Due to the fact

This fiction is similar to the ZFPP scheme describedbefore.Logic functions - The output from the various ZFPN

algorithms are combined in the form of logic output. Theinternal signals for each phase are: phase selection P),direction D),ast zone 1 tripping Z I ) and overreachingzone 2 ZZ). Using as an and operator, combinations usedto form output signal from the logic as follows:Send =D P 2 and Trip = Send ZI

The 22 function is used to limit the reach of Send. Thedirectional phase selection function can be used to speed upthe main 2nd zone tripping in the remote end. It could also beused in directional comparison if the ZFPN directionalfunctions in both ends are used and compared.111. IMPLEMENTATION

A . Hardwure Platformthat, the in the phase currents can not be Implementation of the distance protection with fastcontiiuously, the phase selection for all types of faults mustbe blocked after 15 ms and facilitating the performance ofhigh speed m easurement.

hipping algorithms has beenas shown in Fig. 3, that is a part of n automatedsubstation concept The transformer module, the AID

in a modular line

conversion module, the main processing module, the powerDirection D) - The directional measurement is performed supply the binary I o module and thecommunicationmodule are shown in Fig, 3 .ith full cross polarization. The polarizing voltage is takenentireIy from the healthy phases. The change in the phasecurrent (aiphase)s used for-the directionality to eliminate the

583

8/12/2019 01432444

4/6

The main processing module can have up to 12 DSPs andone 32 bit CPU (Central Processing Unit). The CPU is usedfor logic and comm unication. Fig. 3also illustrates the s ignaland informationflow as a pipe-line within a line terminal. 5currents and 5 voltages are connected to the A/D converterthat has a sampling rate of 2 kHz or 40 sampleslcycle in a 50Hz system. In addition to the phase quantities, the 31 of theparallel line and the bus voltage are included to allow mutualcompen sation and synchrocheck. The 2 kHz signal is filtereddown to lcHz with sliding average filtering. This oversampling technique gives both fast operation and goodtransient behaviour. Each millisecond, numerical data for 10analogue (5 voltage, current) are sent to 12 DSPs, whichare operating in parallel. This means, a separate continuousmeasuring micro-processor is used for each main function,which allows for both rapid and complex algorithms, as wellas tailor made filtering for each function. The total capacityO fthe DSPs and CPU is approximately 100 MIPS. The datafrom the A /D converter is available for all 12 DSPs. EachDSP performs both the specific filtering and algorithm,required for each function. By this architecture each DSPwith its software can be seen a s one protection function. Aprogram change in one DSP will not affect any of the others.The two modules ZFPP and ZFPN are implemented on oneDSP each.

Fig. 3. The protectionpla formB. Sofhyare Pla mn

The software platform is structured in different levels ,Software incorporating basic elements as start-uproutines, self supervision , operative system, etc.Functional platform, containing software library fordifferent Eunctions, programmable logicltimers, MMI,etc.

such as:

Product specific configuration.The application library contains all the basic softwaremodules, which have been developed. It also containspossibilities to connect future modules. The applicationlibrary for protection, control and monitoring functions areshown in Fig. 4. Each software module corresponds to a

~

584

4

function block. The function blocks for a particdar terrnindare chosen based on the application requirements. Thisconcept offers the possibility to custom-design the protectionand control terminals based on the actual field requirements.

Fig, 4.Protection , Conirol and Monitoring Functions-Application LibraryC. Functional Con guration

The various functions are arranged as individual blocks,that can be combined either as predetermined schemes orcustom designed utilising connectables as shown in Fig. 5A n output signal from one function can be used as an inputsignal to another function. The function blocks include allprotection functions, tripping and autoreclosing logic, allcontrol functions for apparatus control and interlocking,binary inputs and outputs as well as a logical fun ction librarywith AND, OR and Time Delayed elements. As an example,each distance zone can be programmed individualIy and alsoaccessible individually in the logic. External (or internal)signals can be used to block or enable the auto reclosure.

Fig. S.Programmable logicFor the control finctions different software modules areavailable. For apparatus control the select before executeprinciple is utilised .The protection and control functions canbe integrated in the terminal in a very cost efficient way. Toallow the user to take advantage of the configuration

8/12/2019 01432444

5/6

flexibility, a Computer Aided Tool (CAT) for PC isavailable. The tool CAP 531 is based on the IEC standard1131 -3 and allows the user to configure the terminal usinggraphic symbols, which makes the handling of theconfiguration tool very simple.IV. APPLICATION WITH SERIES COMPENSATEDLINESSeries compensation is a method which is increasinglyused to allow transmission of heavy load over long distances,with maintained stability. Short clearing time is most

important to maintain stability. This is especially true forclose in faults. In systems with series compensation , lowfrequency transients, large phase shiits, current and voltagereversals may furthermore influence the measured quantities,Finally, both protective gaps and MOV will change thevoltages and currents used by the protection system.The fast tripping algorithms proposed in this paper have

been accomplished with added features for the protection ofseries compensated transmission lines. Some of them can bementioned as follows:- In order to maintain correct directional discrimination incase of voltage reversal a polarizat ion hc t i o n is addedto the main distance protection. The polarization voltageis based on healthy phase and memorized voltage.The directional determination associated with ZFPNscheme utilize superimposed currents and polarizingvoltages during a short interval after fault occurrence. Itmakes it possible to extend the reach of the fast trippingzone regardless of the capacitor presence in the faultloop, since the apparent reactance changes slowly.The comparison of the current level with voltage levelgives impedance circles ( Fig. 2) for close in faults withhigher fault current. During this condition the effectivecompensation is reduced due to the influence of the overvoItage protection. This enables the protection to covera larger position of the protected line.

V. TESTINGDevelopment and evaluation of the new protectionschemes has been done within an interactive softwareenvironment. The software testing of the protection schemesis accomplished by using EMTP-MATLAB software tools.EMTP/ATP files and recorded data from a real time power

system simulator (with a connected disturbance recorder)have been used as input data. To thoroughly evaluateoperation of the proposed protection schemes, severalnetwork configurations have been simulated 191.Additional tests have been done using a real time powersystem simulator [ 5 ] and using a real time Digital TransientNetwork Analyser (DTNA)[6]. The DTNA test was camedout on a 500 kV, 288 krn parallel transmission line as shown

in Fig. 6 . The transmission line has four (4) series capacitors(each 28 YO ompensation) which are protected by MOV. Aphase to phase fau lt was initiated at point F3 (aprox. 40 ofthe total length seen at relay R1, Fig. 6 ) . The protections areplaced at both end points (RI & R2).

Fig. 6 A parallel series compensated transmissionfine for DTNA test

Fig. . Phase l o phase fault at F3 as n Fig.The response of the protection as well as voltages andcurrents at relay point RI are shown in Fig. 7. The totaloperate time including output relays is aprox. 14.5 ms. TheDTNA tests have confirmed the validity of the high speedprotection algorithm presented in this paper.

585

8/12/2019 01432444

6/6

VI. CONCLUSIONSThis paper d escribes a new, high speed protection for EHVtransmission lines. The presented scheme also performs wellon series compensated lines. The protection consists of a

main distance protection and fast tripping algorithm for multiphase faults and single phase to ground faults. The fasttripping algorithm for multi phase faults (ZFPP) and forsingle line to ground faults (ZFPN) are introduced toovercome the effects if the series capacitor installation ispresent in the fault loop. Comprehensive test results onvarious test systems and a utility system [9 ] show the benefitsof the proposed schemes. The fast tripping algorithm (ZFPN)that minimizes the fault time in case of single phase toground faults is described in detail in this paper. Thehighlights of the scheme are:Combination of measuring algorithms provides asolution for a high speed protection for EHV as well asfor series compensated transmission lines. The schemeprovides a high-speed operate time (typically less th nhalf a cycle). Total protection operate time is less th none cycle.The protection can be set to cover up to 70 of thetotal uncompensated positive sequence reactance, incase of low resistance faults. Large fault resistance maycause a reduction in the protection reach.The scheme is in operation in some 25 countries aroundthe world for the protection of Series CompensatedTransmission Lines. The operational experiences arevery good.

VII. REFERENCESF. Anderson and W. Elmore, Overview of Series-CompensatedLine Protection Phlosophles, Westem Relay ProtectiveConference Washington State University, Spokane, WashingtonOctober 1990.I. Cheetham, A Newbould, and G . Sh am e, Series-CompensatedLine Protection: System Modelling and Test Results, 15 AnnualWestem Relay Protective Conference, Washington State University,Spokane,Washington, October 1988.M Chamia and S Liberman, Ultra high speed relay forEHVLJHV tr nsms ion lines- Development, design andapplication, IEEE Transactions on PAS, Vol. PAS-97, No.6,Nov./Dec. 1978.Alternative Transient Prograrr ATP), Rule Book, LU.Leuven,EMTP Centre, Leuven, Belgium 1987.G. NimnersjB, B. Hillsbrim, 0 W e r n e r - E r i c h , and G.D.Rockefeller A digitally controlled real time, analogue power-syst m imulator for closed loop protective retaymg testing, IEEETrans. on PWRD, Vol. 3, No. 1, January 1988.G. Nimmersjb, M.M. Saha, and B. Hillstr6m I Protective relaytesting using a modern digital real t ime simulator , companionpaper to be presented at IEEE PES Winter meting 2000, Singapore,January 23-27 ,2000.A. Engqvist and L.Eriksson, Numerical Distance Protection forSubtransmission ines, Cigre, 34-04, AugJSept. 1988.

6

[8] G. Nitnmersja and M.M. Saha, A new approach to high speedrelaymg based on transient phenomena, IEED PSP-89, Edinburgh,VK, April 1989.C. ohlen, J. Esztergalyos, G. immersjo, and M.M. Saha, EMTPused in testing of a protection scheme for series compensatednetwork, Cigri, Stockholm, June 11-17, 1995.

[9]

VIII. ACKNOWLEDGMENTSThe authors extend their gratehl acknowledgement toGunnar Nimmersj i, and k i f Eriksson of ABB AutomationTechnologies, Sweden, for their substantial contributionsduring basic development of this concept.

IX. BIOGRAPHIESMurari Mohar Saha ( M76, SM87) was born in 1947 in Bangladesh.He received B3c.E.E. from Bangladesh University of Engineering andTechnology (ELJET), Dhaka in 1968 and completed M.Sc.E.E. in 1970.From 1969 to 1971 he was a lecturer at the department of ElectricalEngineering at BUET, Dhaka. In 1972, he completed M.S.E.E. and in 1975

he was awarded with Ph.D. from the Techrucal University of Warsaw,Poland. He joined ASEA, Sweden, in 1975 as a Development Engineer andcurrently is a Senior Research and Development Engineer at ABBAutonation Technologies , Vide&, Sweden. He is a Senior Member ofIEEE and a Fellow of I E W ) . He is a registered European Engineer EURING) and a Chartered Engineer CEng). His areas of interest are measuringtransformers, power system analysis and simulation, and digital protectiverelays.

Correspo nding Author:Dr. Murari Mohan Saha, Senior Development EngineerABB Automation Technologies.SE-721 59VasterAs,SWEDENE-mail: murari.saha@se.abb.com

586

mailto:murari.saha@se.abb.commailto:murari.saha@se.abb.com