10 Unit Protection

8/14/2019 10 Unit Protection

1/18


2/18

N e t w o r k P r o t e c t i o n & A u t o m a t i o n G u i d e 1 5 3

10 .1 INTRODUCTION

The graded overcurrent systems described in Chapter 9,though attractively simple in principle, do not meet allthe protection requirements of a power system.Application difficulties are encountered for two reasons:firstly, satisfactory grading cannot always be arrangedfor a complex network, and secondly, the settings maylead to maximum tripping times at points in the systemthat are too long to prevent excessive disturbances

occurring.These problems led to the concept of 'Unit Protection',whereby sections of the power system are protectedindividually as a complete unit without reference toother sections. One form of Unit Protection is alsoknown as Differential Protection, as the principle is tosense the difference in currents between the incomingand outgoing terminals of the unit being protected.Other forms can be based on directional comparison, ordistance teleprotection schemes, which are covered inChapter 12, or phase comparison protection, which isdiscussed later in this chapter. The configuration of the

power system may lend itself to unit protection; forinstance, a simple earth fault relay applied at the sourceend of a transformer-feeder can be regarded as unitprotection provided that the transformer windingassociated with the feeder is not earthed. In this casethe protection coverage is restricted to the feeder andtransformer winding because the transformer cannottransmit zero sequence current to an out-of-zone fault.

In most cases, however, a unit protection systeminvolves the measurement of fault currents (and possiblyvoltages) at each end of the zone, and the transmissionof information between the equipment at zone

boundaries. It should be noted that a stand-alonedistance relay, although nominally responding only tofaults within their setting zone, does not satisfy theconditions for a unit system because the zone is notclearly defined; it is defined only within the accuracylimits of the measurement. Also, to cater for someconditions, the setting of a stand-alone distance relaymay also extend outside of the protected zone to caterfor some conditions.

Merz and Price [10.1] first established the principle of current differential unit systems; their fundamentaldifferential systems have formed the basis of many

10 Unit Protection of Feeders

Chap10-152-169 21/06/02 8:43 Page 153


3/18

N e t w o r k P r o t e c t i o n & A u t o m a t i o n G u i d e

highly developed protection arrangements for feedersand numerous other items of plant. In one arrangement,an auxiliary pilot circuit interconnects similar currenttransformers at each end of the protected zone, asshown in Figure 10.1. Current transmitted through thezone causes secondary current to circulate round thepilot circuit without producing any current in the relay.For a fault within the protected zone the CT secondarycurrents will not balance, compared with the through-fault condition, and the difference between the currentswill flow in the relay.

An alternative arrangement is shown in Figure 10.2, inwhich the CT secondary windings are opposed forthrough-fault conditions so that no current flows in theseries connected relays. The former system is known asa Circulating Current system, while the latter is knownas a Balanced Voltage system.

Most systems of unit protection function through thedetermination of the relative direction of the faultcurrent. This direction can only be expressed on a

comparative basis, and such a comparative measurementis the common factor of many systems, includingdirectional comparison protection and distanceteleprotection schemes with directional impedancemeasurement.

A major factor in consideration of unit protection is themethod of communication between the relays. This iscovered in detail in Chapter 8 in respect of the latestfibre-optic based digital techniques. For older pilot wiresystems, only brief mention is made. For more detaileddescriptions of pilot wire techniques, see reference[10.2] in Section 10.13.

10 .2 CONVENTION OF DIRECTION

It is useful to establish a convention of direction of current flow; for this purpose, the direction measuredfrom a busbar outwards along a feeder is taken aspositive. Hence the notation of current flow shown inFigure 10.3; the section GH carries a through current

which is counted positive at G but negative at H , whilethe infeeds to the faulted section HJ are both positive.

Neglect of this rule has often led to anomalousarrangements of equipment or difficulty in describingthe action of a complex system. When applied, the rulewill normally lead to the use of identical equipments atthe zone boundaries, and is equally suitable for extensionto multi-ended systems. It also conforms to the standardmethods of network analysis.

10 .3 CONDITIONSFOR DIRECTION COMPARISON

The circulating current and balanced voltage systems of Figures 10.1 and 10.2 perform full vectorial comparisonof the zone boundary currents. Such systems can betreated as analogues of the protected zone of the powersystem, in which CT secondary quantities representprimary currents and the relay operating currentcorresponds to an in-zone fault current.

These systems are simple in concept; they arenevertheless applicable to zones having any number of boundary connections and for any pattern of terminalcurrents.

To define a current requires that both magnitude andphase be stated. Comparison in terms of both of thesequantities is performed in the Merz-Price systems, but itis not always easy to transmit all this informationover some pilot channels. Chapter 8 provides a detaileddescription of modern methods that may be used.

10 .4 CIRCULATING CURRENT SYSTEM

The principle of this system is shown in outline inFigure 10.1. If the current transformers are ideal, thefunctioning of the system is straightforward. The

10

U n

i t P r o t e c t

i o n

F e e

d e r s

1 5 4

EndH EndG

RelayI >

Figure 10.1: Circulating current system

FaultH G J

Source+ _ + +

Source

Figure 10.3: Convention of current direction

Figure 10.2: Balanced voltage system

EndG

Relay RelayH

EndH

I > I >

Chap10-152-169 21/06/02 8:43 Page 154


4/18


transformers will, however, have errors arising from bothWattmetric and magnetising current losses that causedeviation from the ideal, and the interconnectionsbetween them may have unequal impedances. This cangive rise to a spill current through the relay evenwithout a fault being present, thus limiting thesensitivity that can be obtained. Figure 10.4 illustratesthe equivalent circuit of the circulating current scheme.If a high impedance relay is used, then unless the relay islocated at point J in the circuit, a current will flowthrough the relay even with currents I Pg and I Ph beingidentical. If a low impedance relay is used, voltageFF will be very small, but the CT exciting currents will beunequal due to the unequal burdens and relay current I Rwill still be non-zero.

10.4.1 Transient Instability

It is shown in Section 6.4.10 that an asymmetricalcurrent applied to a current transformer will induce aflux that is greater than the peak flux corresponding tothe steady state alternating component of the current. Itmay take the CT into saturation, with the result that thedynamic exciting impedance is reduced and the excitingcurrent greatly increased.

When the balancing current transformers of a unitprotection system differ in excitation characteristics, orhave unequal burdens, the transient flux build-ups willdiffer and an increased 'spill' current will result. There isa consequent risk of relay operation on a healthy circuitunder transient conditions, which is clearly

unacceptable. One solution is to include a stabilisingresistance in series with the relay. Details of how tocalculate the value of the stabilising resistor are usuallyincluded in the instruction manuals of all relays thatrequire one.

When a stabilising resistor is used, the relay current

setting can be reduced to any practical value, the relaynow being a voltage-measuring device. There isobviously a lower limit, below which the relay elementdoes not have the sensitivity to pick up. Relaycalibration can in fact be in terms of voltage. For moredetails, see reference [10.2].

10.4.2 Bias

The 'spill' current in the relay arising from these varioussources of error is dependent on the magnitude of thethrough current, being negligible at low values of

through-fault current but sometimes reaching adisproportionately large value for more severe faults.Setting the operating threshold of the protection abovethe maximum level of spill current produces poorsensitivity. By making the differential settingapproximately proportional to the fault current, the low-level fault sensitivity is greatly improved. Figure 10.5illustrates a typical bias characteristic for a modern relaythat overcomes the problem. At low currents, the bias issmall, thus enabling the relay to be made sensitive. Athigher currents, such as would be obtained from inrush orthrough fault conditions, the bias used is higher, and thusthe spill current required to cause operation is higher. Therelay is therefore more tolerant of spill current at higherfault currents and therefore less likely to maloperate,while still being sensitive at lower current levels.

10

U n

i t P r o t e c t

i o n

F e e

d e r s

Restrain

I 1

I diff

I s1

I s2

I 2

I 3

Percentagebias k2

Percentagebias k1

Operate

I 1I bias+ I 2 I 3+

2=

= I 1+I 2+I 3

Figure 10.5: Typical bias characteristic of relay

Figure 10.4: Equivalent circuit of circulating current scheme

Relay I

RSh RSh

I PhI Pg

i hieg

R

RR

Z hZ eg

ShiSg

RLg RLh

(a)

H

G ''

G

G '

F '

F

J

H '

H

Electro-motive forces with low impedance relayGG GG

(b)

En G En H

ubscripts:

S - CT Secondary

LG

- end H

Chap10-152-169 21/06/02 8:43 Page 155


5/18


10 .5 BALANCED VOLTAGE SYSTEM

This section is included for historical reasons, mainlybecause of the number of such schemes still to be foundin service for new installations it has been almostcompletely superseded by circulating current schemes. Itis the dual of the circulating current protection, and is

summarised in Figure 10.2 as used in theTranslay H04scheme.

With primary through current, the secondary e.m.f.s of the current transformers are opposed, and provide nocurrent in the interconnecting pilot leads or the seriesconnected relays. An in-zone fault leads to a circulatingcurrent condition in the CT secondaries and hence torelay operation.

An immediate consequence of the arrangement is thatthe current transformers are in effect open-circuited, asno secondary current flows for any primary through-current conditions. To avoid excessive saturation of thecore and secondary waveform distortion, the core isprovided with non-magnetic gaps sufficient to absorbthe whole primary m.m.f. at the maximum current level,the flux density remaining within the linear range. Thesecondary winding therefore develops an e.m.f. and canbe regarded as a voltage source. The shunt reactance of the transformer is relatively low, so the device acts as atransformer loaded with a reactive shunt; hence theAmerican name of transactor. The equivalent circuit of the system is as shown in Figure 10.6.

The series connected relays are of relatively highimpedance; because of this the CT secondary windingresistances are not of great significance and the pilotresistance can be moderately large without significantlyaffecting the operation of the system. This is why thescheme was developed for feeder protection.

10.5.1 Stability Limit of the Voltage Balance System

Unlike normal current transformers, transactors are notsubject to errors caused by the progressive build-up of

exciting current, because the whole of the primarycurrent is expended as exciting current. In consequence,the secondary e.m.f. is an accurate measure of theprimary current within the linear range of thetransformer. Provided the transformers are designed tobe linear up to the maximum value of fault current,balance is limited only by the inherent limit of accuracyof the transformers, and as a result of capacitancebetween the pilot cores. A broken line in the equivalentcircuit shown in Figure 10.6 indicates such capacitance.Under through-fault conditions the pilots are energisedto a proportionate voltage, the charging current flowingthrough the relays. The stability ratio that can beachieved with this system is only moderate and a biastechnique is used to overcome the problem.

10 .6 SUMMATION ARRANGEMENTS

Schemes have so far been discussed as though they wereapplied to single-phase systems. A polyphase systemcould be provided with independent protection for eachphase. Modern digital or numerical relayscommunicating via fibre-optic links operate on thisbasis, since the amount of data to be communicated isnot a major constraint. For older relays, use of thistechnique over pilot wires may be possible for relativelyshort distances, such as would be found with industrialand urban power distribution systems. Clearly, eachphase would require a separate set of pilot wires if theprotection was applied on a per phase basis. The cost of providing separate pilot-pairs and also separate relay

elements per phase is generally prohibitive. Summationtechniques can be used to combine the separate phasecurrents into a single relaying quantity for comparisonover a single pair of pilot wires. For details of suchtechniques, see reference [10.2].

10 .7 EXAMPLES OF ELECTROMECHANICALAND STATIC UNIT PROTECTION SYSTEMS

As mentioned above, the basic balanced voltage principleof protection evolved to biased protection systems.Several of these have been designed, some of which

appear to be quite different from others. Thesedissimilarities are, however, superficial. A number of these systems that are still in common use are describedbelow.

10.7.1 Translay Balanced VoltageElectromechanical System

A typical biased, electromechanical balanced voltagesystem, trade name Translay, still giving useful serviceon distribution systems is shown in Figure 10.7.

10

U n

i t P r o t e c t

i o n

F e e

d e r s

1 5 6

Figure 10.6: Equivalent circuit for balanced voltage system

Z eg

RSg RLg RLh RSh

h

I >I >

Relay

PilotParameters

EndG

RelayH

EndH

Chap10-152-169 21/06/02 8:43 Page 156


6/18


The electromechanical design derives its balancing voltagesfrom the transactor incorporated in the measuring relay ateach line end. The latter are based on the induction-typemeter electromagnet as shown in Figure 10.7.

The upper magnet carries a summation winding toreceive the output of the current transformers, and a

secondary winding which delivers the reference e.m.f.The secondary windings of the conjugate relays areinterconnected as a balanced voltage system over thepilot channel, the lower electromagnets of both relaysbeing included in this circuit.

Through current in the power circuit produces a state of balance in the pilot circuit and zero current in the lowerelectromagnet coils. In this condition, no operatingtorque is produced.

An in-zone fault causing an inflow of current from eachend of the line produces circulating current in the pilotcircuit and the energisation of the lower electromagnets.These co-operate with the flux of the upperelectromagnets to produce an operating torque in thediscs of both relays. An infeed from one end only willresult in relay operation at the feeding end, but nooperation at the other, because of the absence of uppermagnet flux.

Bias is produced by a copper shading loop fitted to thepole of the upper magnet, thereby establishing a Ferrarismotor action that gives a reverse or restraining torqueproportional to the square of the upper magnet flux value.

Typical settings achievable with such a relay are:Least sensitive earth fault - 40% of rating

Least sensitive phase-phase fault - 90% of ratingThree-phase fault - 52% of rating

10.7.2 Static Circulating CurrentUnit Protection System Translay S

A typical static modular pilot wire unit protection systemoperating on the circulating current principle is shown inFigure 10.8. This uses summation transformers with aneutral section that is tapped, to provide alternativeearth fault sensitivities. Phase comparators tuned to thepower frequency are used for measurement and arestraint circuit gives a high level of stability for throughfaults and transient charging currents. High-speedoperation is obtained with moderately sized currenttransformers and where space for current transformers islimited and where the lowest possible operating time isnot essential, smaller current transformers may be used.This is made possible by a special adjustment (K t ) bywhich the operating time of the differential protectioncan be selectively increased if necessary, therebyenabling the use of current transformers having acorrespondingly decreased knee-point voltage, whilstensuring that through-fault stability is maintained togreater than 50 times the rated current.

Internal faults give simultaneous tripping of relays atboth ends of the line, providing rapid fault clearanceirrespective of whether the fault current is fed from bothline ends or from only one line end.

10

U n

i t P r o t e c t

i o n

F e e

d e r s

Figure 10.8: Typical static circulating current feeder unit protection circuit diagram

A

C

BT 1 - Summation transformer

T 2 - Auxiliary transformer

RVO - Non linear resistor

Trip Trip

V

T 2 T 2

T 1T 1

P r P

T

T O

T RsRs

RoRo RVORVOV

Pilot wires

T o - Operating winding

T r - Restraining winding

Ro - Linear resistor

P r - Pilots padding resistor

c - Phase comparator

End G End H

ummationwinding

Pilot

ABC

Secondarywinding

Bias loop

Figure 10.7: Typical biased electromechanical differential protection system.

Chap10-152-169 21/06/02 8:43 Page 157


7/18


10 .8 DIGITAL/NUMERICAL CURRENTDIFFERENTIAL PROTECTION SYSTEMS

A digital or numerical unit protection relay may typicallyprovide phase-segregated current differential protection.This means that the comparison of the currents at eachrelay is done on a per phase basis. For digital data

communication between relays, it is usual that a directoptical connection is used (for short distances) or amultiplexed link. Link speeds of up to 64kbit/s (56kbit/sin N. America) are normal. Through current bias istypically applied to provide through fault stability in theevent of CT saturation. A dual slope bias technique(Figure 10.5) is used to enhance stability for throughfaults. A typical trip criterion is as follows:

For |I bias | < I s2|I diff | < k1 |I bias | + I s1

For |I bias | < I s2

|I diff | < k2 |I bias | - (k2 - k1) I s2 + I s1Once the relay at one end of the protected section hasdetermined that a trip condition exists, an intertripsignal is transmitted to the relay at the other end. Relaysthat are supplied with information on line currents at allends of the line may not need to implement intertrippingfacilities. However, it is usual to provide intertripping inany case to ensure the protection operates in the eventof any of the relays detecting a fault.

A facility for vector/ratio compensation of the measuredcurrents, so that transformer feeders can be included in

the unit protection scheme without the use of interposing CTs or defining the transformer as a separatezone increases versatility. Any interposing CTs requiredare implemented in software. Maloperation ontransformer inrush is prevented by second harmonicdetection. Care must be taken if the transformer has awide-ratio on-load tap changer, as this results in thecurrent ratio departing from nominal and may causemaloperation, depending on the sensitivity of the relays.The initial bias slope should be set taking this intoconsideration.

Tuned measurement of power frequency currents

provides a high level of stability with capacitance inrushcurrents during line energisation. The normal steady-state capacitive charging current can be allowed for if avoltage signal can be made available and thesusceptance of the protected zone is known.

Where an earthed transformer winding or earthingtransformer is included within the zone of protection,some form of zero sequence current filtering is required.This is because there will be an in-zone source of zerosequence current for an external earth fault. Thedifferential protection will see zero sequence differentialcurrent for an external fault and it could incorrectly

operate as a result. In older protection schemes, theproblem was eliminated by delta connection of the CTsecondary windings. For a digital or numerical relay, aselectable software zero sequence filter is typicallyemployed.

The problem remains of compensating for the time

difference between the current measurements made atthe ends of the feeder, since small differences can upsetthe stability of the scheme, even when using fast directfibre-optic links. The problem is overcome by either timesynchronisation of the measurements taken by therelays, or calculation of the propagation delay of the linkcontinuously.

10.8.1 Time Synchronisation of Relays

Fibre-optic media allow direct transmission of thesignals between relays for distances of up to several km

without the need for repeaters. For longer distancesrepeaters will be required. Where a dedicated fibre pairis not available, multiplexing techniques can be used. Asphase comparison techniques are used on a per phasebasis, time synchronisation of the measurements isvitally important. This requires knowledge of thetransmission delay between the relays. Four techniquesare possible for this:

a. assume a valueb. measurement during commissioning onlyc. continuous online measurementd. GPS time signal

Method (a) is not used, as the error between the assumedand actual value will be too great.

Method (b) provides reliable data if directcommunication between relays is used. As signalpropagation delays may change over a period of years,repeat measurements may be required at intervals andrelays re-programmed accordingly. There is some risk of maloperation due to changes in signal propagation timecausing incorrect time synchronisation betweenmeasurement intervals. The technique is less suitable if rented fibre-optic pilots are used, since the owner mayperform circuit re-routing for operational reasonswithout warning, resulting in the propagation delaybeing outside of limits and leading to schememaloperation. Where re-routing is limited to a fewroutes, it may be possible to measure the delay on allroutes and pre-program the relays accordingly, with therelay digital inputs and ladder logic being used to detectchanges in route and select the appropriate delayaccordingly.

Method (c), continuous sensing of the signal propagationdelay, is a robust technique. One method of achievingthis is shown in Figure 10.9.

10

U n

i t P r o t e c t

i o n

F e e

d e r s

1 5 8

Chap10-152-169 21/06/02 8:43 Page 158


8/18


Relays A and B sample signals at time T A1 ,T A2 andT B1,T B2 respectively. The times will not be coincident,even if they start coincidentally, due to slight differencesin sampling frequencies. At timeT A1 relay A transmitsits data to relay B, containing a time tag and other data.Relay B receives it at time T A1 +T p1 where T p1 is thepropagation time from relay A to relay B. Relay Brecords this time as time T B* . Relay B also sendsmessages of identical format to relay A . It transmits sucha message at time T B3, received by relay A at timeT B3+ T p2 (say time T A* ), where T p2 is the propagationtime from relay B to relay A . The message from relayBto relay A includes the time T B3, the last received timetag from relay A (T A1 ) and the delay time between thearrival time of the message from A (T B* ) and T B3 callthis the delay time T d. The total elapsed time istherefore:

(T A* - T A1 ) = (T d + T p1 + T p2 )

If it is assumed that T p1 = T p2 , then the value of T p1 and

T p2 can be calculated, and hence also T B3. The relayBmeasured data as received at relay A can then beadjusted to enable data comparison to be performed.Relay B performs similar computations in respect of thedata received from relay A (which also contains similartime information). Therefore, continuous measurementof the propagation delay is made, thus reducing thepossibility of maloperation due to this cause to aminimum. Comparison is carried out on a per-phase basis,so signal transmission and the calculations are requiredfor each phase. A variation of this technique is availablethat can cope with unequal propagation delays in the two

communication channels under well-defined conditions.

The technique can also be used with all types of pilots,subject to provision of appropriate interfacing devices.

Method (d) is also a robust technique. It involves bothrelays being capable of receiving a time signal froma GPS satellite. The propagation delay on eachcommunication channel is no longer required to be

known or calculated as both relays are synchronised to acommon time signal. For the protection scheme to meetthe required performance in respect of availability andmaloperation, the GPS signal must be capable of reliablereceipt under all atmospheric conditions. There is extrasatellite signal receiving equipment required at bothends of the line, which implies extra cost.

The minimum setting that can be achieved with suchtechniques while ensuring good stability is 20% of CTprimary current.

10.8.2 Application to Mesh Cornerand 1 1/2 Breaker Switched Substations

These substation arrangements are quite common, andthe arrangement for the latter is shown in Figure 10.10.Problems exist in protecting the feeders due to thelocation of the line CTs, as either Bus 1 or Bus 2 or bothcan supply the feeder. Two alternatives are used toovercome the problem, and they are illustrated in theFigure. The first is to common the line CT inputs (asshown for Feeder A ) and the alternative is to use a secondset of CT inputs to the relay (as shown for FeederB).

10

U n

i t P r o t e c t

i o n

F e e

d e r s

End A

Digital communications link A B

End B

C u r r e n t v e c t o r s

C u r r e n t vec t o r s

Measured sampling time Propagation delay timeT B3* =(T A * -T p2 ) T p1 =T p2 =1/2(T A * -T A1 -T d )

T B*

T A*

T A1 'T A2 ' T B1'T B2'

T p1

T p2

T d

T A1 *

T B*

T B3*

T B3*

T B2

T B3

T B3T d

T d

T B4

T B5

T B1

T A2

T A3

T A4

T A5

T A1

T A1

T A1

T p1

T p2

- sampling instants of relay A- sampling instants of relay B- propagation delay time

from relay A to B- propagation delay time

from relay B to A- time between the arrival

of message T A1 at relay Band despatch of message T B3

- arrival time of message T B3and relay A

- arrival time of message T A1and relay B

- the measured samplingtime of T B3 by relay A

Figure 10.9: Signal propagation delay measurement

Chap10-152-169 21/06/02 8:43 Page 159


9/18


In the case of a through fault as shown, the relayconnected to Feeder A theoretically sees no unbalancecurrent, and hence will be stable. However, with the linedisconnect switch open, no bias is produced in the relay,so CTs need to be well matched and equally loaded if maloperation is to be avoided.

For Feeder B, the relay also theoretically sees nodifferential current, but it will see a large bias current evenwith the line disconnect switch open. This provides a highdegree of stability, in the event of transient asymmetric CTsaturation. Therefore, this technique is preferred.

Sensing of the state of the line isolator through auxiliarycontacts enables the current values transmitted to andreceived from remote relays to be set to zero when theisolator is open. Hence, stub-bus protection for theenergised part of the bus is then possible, with any faultresulting in tripping of the relevant CB.

10.9 CARRIER UNIT PROTECTION SCHEMES

In earlier sections, the pilot links between relays havebeen treated as an auxiliary wire circuit thatinterconnects relays at the boundaries of the protectedzone. In many circumstances, such as the protection of longer line sections or where the route involvesinstallation difficulties, it is too expensive to provide anauxiliary cable circuit for this purpose, and other meansare sought.

In all cases (apart from private pilots and some shortrented pilots) power system frequencies cannot betransmitted directly on the communication medium.Instead a relaying quantity may be used to vary thehigher frequency associated with each medium (or thelight intensity for fibre-optic systems), and this processis normally referred to as modulation of a carrier wave.Demodulation or detection of the variation at a remotereceiver permits the relaying quantity to be reconstitutedfor use in conjunction with the relaying quantitiesderived locally, and forms the basis for all carrier systemsof unit protection.

Carrier systems are generally insensitive to induced

power system currents since the systems are designed tooperate at much higher frequencies, but each mediummay be subjected to noise at the carrier frequencies thatmay interfere with its correct operation. Variations of signal level, restrictions of the bandwidth available forrelaying and other characteristics unique to eachmedium influence the choice of the most appropriatetype of scheme. Methods and media for communicationare discussed in Chapter 8.

10.10 CURRENT DIFFERENTIAL SCHEME ANALOGUE TECHNIQUES

The carrier channel is used in this type of scheme toconvey both the phase and magnitude of the current atone relaying point to another for comparison with thephase and magnitude of the current at that point.Transmission techniques may use either voice frequencychannels using FM modulation or A/D converters anddigital transmission. Signal propagation delays still needto be taken into consideration by introducing adeliberate delay in the locally derived signal before acomparison with the remote signal is made.

A further problem that may occur concerns the dynamicrange of the scheme. As the fault current may be up to30 times the rated current, a scheme with linearcharacteristics requires a wide dynamic range, whichimplies a wide signal transmission bandwidth. Inpractice, bandwidth is limited, so either a non-linearmodulation characteristic must be used or detection of fault currents close to the setpoint will be difficult.

10.10.1 Phase Comparison Scheme

The carrier channel is used to convey the phase angle of thecurrent at one relaying point to another for comparisonwith the phase angle of the current at that point.

The principles of phase comparison are illustrated inFigure 10.11. The carrier channel transfers a logic or'on/off' signal that switches at the zero crossing pointsof the power frequency waveform. Comparison of a locallogic signal with the corresponding signal from the

remote end provides the basis for the measurement of phase shift between power system currents at the twoends and hence discrimination between internal andthrough faults.

Current flowing above the set threshold results in turn-off of the carrier signal. The protection operates if gapsin the carrier signal are greater than a set duration thephase angle setting of the protection.

Load or through fault currents at the two ends of aprotected feeder are in antiphase (using the normal relayconvention for direction), whilst during an internal faultthe (conventional) currents tend towards the in-phase

10

U n

i t P r o t e c t

i o n

F e e

d e r s

1 6 0

Figure 10.10: Breaker and a half switched substation

Bus 1

A B

Stubbusinputs

F

Bus 2B1I F

I d>I d>

B2

Chap10-152-169 21/06/02 8:46 Page 160


10/18


condition. Hence, if the phase relationship of throughfault currents is taken as a reference condition, internalfaults cause a phase shift of approximately 180 withrespect to the reference condition.

Phase comparison schemes respond to any phase shiftfrom the reference conditions, but tripping is usually

permitted only when the phase shift exceeds an angle of typically 30 to 90 degrees, determined by the time delaysetting of the measurement circuit, and this angle isusually referred to as the Stability Angle. Figure 10.12 isa polar diagram that illustrates the discriminationcharacteristics that result from the measurementtechniques used in phase comparison schemes.

Since the carrier channel is required to transfer only

binary information, the techniques associated withsending teleprotection commands. Blocking orpermissive trip modes of operation are possible, howeverFigure 10.11 illustrates the more usual blocking mode,since the comparator provides an output when neithersquarer is at logic '1'. A permissive trip scheme can berealised if the comparator is arranged to give an outputwhen both squarers are at logic '1'. Performance of thescheme during failure or disturbance of the carrierchannel and its ability to clear single-end-fed faultsdepends on the mode of operation, the type and functionof fault detectors or starting units, and the use of anyadditional signals or codes for channel monitoring andtransfer tripping.

10

U n

i t P r o t e c t

i o n

F e e

d e r s

Signalling equipment andcommunication channel

Receiver

G F

E

SquarerSummationnetwork

A

Transmitter

D '

B

H

DC

Load or through fault Internal faultI G I H I G I H H G

01

01

01

01

01

01

01

01

E =B+D '

Stability setting

D. Squarer output at end H (Received at end G viaideal carrier system as D '

C. Summation voltage at endH

E. Comparator output at end G

F. Discriminator output at end G

B. Squarer output at end G

A. Summation voltage at endG

EndH End

Phasecomparator

Pulse lengthdiscrimination

Figure 10.11: Principles of phase comparison protection.

Chap10-152-169 21/06/02 8:46 Page 161


11/18


Signal transmission is usually performed by voicefrequency channels using frequency shift keying (FSK) orPLC techniques.

Voice frequency channels involving FSK use two discretefrequencies either side of the middle of the voice band.This arrangement is less sensitive to variations in delay orfrequency response than if the full bandwidth was used.

Blocking or permissive trip modes of operation may beimplemented. In addition to the two frequencies usedfor conveying the squarer information, a third tone isoften used, either for channel monitoring or transfertripping dependent on the scheme.

For a sensitive phase comparison scheme, accuratecompensation for channel delay is required. However,since both the local and remote signals are logic pulses,simple time delay circuits can be used, in contrast to theanalogue delay circuitry usually required for currentdifferential schemes.

The principles of the Power Line Carrier channeltechnique are illustrated in Figure 10.13. The schemeoperates in the blocking mode. The 'squarer' logic is useddirectly to turn a transmitter 'on' or 'off' at one end, andthe resultant burst (or block) of carrier is coupled to andpropagates along the power line which is being protectedto a receiver at the other end. Carrier signals above athreshold are detected by the receiver, and hence producea logic signal corresponding to the block of carrier. Incontrast to Figure 10.11, the signalling system is a 2-wirerather than 4-wire arrangement, in which the localtransmission is fed directly to the local receiver alongwith any received signal. The transmitter frequencies at

both ends are nominally equal, so the receiver respondsequally to blocks of carrier from either end. Through-fault current results in transmission of blocks of carrierfrom both ends, each lasting for half a cycle, but with aphase displacement of half a cycle, so that the compositesignal is continuously above the threshold level and thedetector output logic is continuously '1'. Any phase shiftrelative to the through fault condition produces a gap inthe composite carrier signal and hence a corresponding'0' logic level from the detector. The duration of the logic'0' provides the basis for discrimination between internaland external faults, tripping being permitted only when atime delay setting is exceeded. This delay is usuallyexpressed in terms of the corresponding phase shift indegrees at system frequency s in Figure 10.12.

The advantages generally associated with the use of thepower line as the communication medium apply namely,that a power line provides a robust, reliable, and low-lossinterconnection between the relaying points. In additiondedicated 'on/off' signalling is particularly suited for usein phase comparison blocking mode schemes, as signalattenuation is not a problem. This is in contrast topermissive or direct tripping schemes, where high poweroutput or boosting is required to overcome the extraattenuation due to the fault.

The noise immunity is also very good, making the schemevery reliable. Signal propagation delay is easily allowedfor in the stability angle setting, making the scheme verysensitive as well.

10.11 PHASE COMPARISION PROTECTIONSCHEME CONSIDERATIONS

One type of unit protection that uses carrier techniquesfor communication between relays is phase comparisonprotection. Communication between relays commonlyuses PLCC or frequency modulated carrier modemtechniques. There are a number of considerations thatapply only to phase comparison protection systems,which are discussed in this section.

10.11.1 Lines with Shunt Capacitance

A problem can occur with the shunt capacitance currentthat flows from an energising source. Since this current isin addition to the load current that flows out of the line,and typically leads it by more than 90, significantdifferential phase shifts between the currents at the endsof the line can occur, particularly when load current is low.

The system differential phase shift may encroach into thetripping region of the simple discriminator characteristic,regardless of how large the stability angle setting maybe. Figure 10.14 illustrates the effect and indicatestechniques that are commonly used to ensure stability.

10

U n

i t P r o t e c t

i o n

F e e

d e r s

1 6 2

Discriminator stability angle setting.

System differential phase shift referred to through fault referenceconditionOR Through faultreference condition H

I G =-I H I G I H

G

RStability

=0

=270

=180 -Tripping

=90

(I G ' I H conventional relay currents at ends of protected feeder)

O

Figure 10.12: Polar diagram for phase comparison scheme

Chap10-152-169 21/06/02 8:46 Page 162


12/18


Operation of the discriminator can be permitted onlywhen current is above some threshold, so thatmeasurement of the large differential phase shifts whichoccur near the origin of the polar diagram is avoided. Bychoice of a suitable threshold and stability angle, a'keyhole' characteristic can be provided such that thecapacitive current characteristic falls within theresultant stability region. Fast resetting of the faultdetector is required to ensure stability following theclearance of a through fault when the currents tend tofall towards the origin of the polar diagram.

The mark-space ratio of the squarer (or modulating)waveform can be made dependent on the currentamplitude. Any decrease in the mark-space ratio willpermit a corresponding differential phase shift to occurbetween the currents before any output is given from thecomparator for measurement in the discriminator. Asquarer circuit with an offset or bias can provide a

10

U n

i t P r o t e c t

i o n

F e e

d e r s

D. Discriminator output

C. Carrier detector output

B. Composite carrier signal at endG

1

1

0

0

10

0

1

Receiver

Transmitter

Trip

A. Squarer output at end G

Pulse lengthdiscriminator

Squarer

D

A

C

En G

Summation

network

B

Couplingfilter

Line trapLine trap

Stability setting

EndH

Identicalrelay

to end G

Blocks of carrier transmitted from end G

Squarer output at end H

Blocks of carrier transmitted from end H

Load or through fault Internal fault

0

1

0

1

0

1

1

0

Trip

Figure 10.13: Principles of power line carrier phase comparison

Figure 10.14: Capacitive current in phase comparisonschemes and techniques used to avoid instability

Limits of differential phase shift due to capacitive currentI C Encroachment into tripping region for discriminatorwith stability angle setting s

Through FaultReference

A

Squarer ThresholdStarter Threshold

`Keyhole' characteristicMinimum starter threshold =

Characteristic of system with amplitude dependentcompensation s = angular compensation for current of magnitudeOA

O

sin scapacitive current

c

where s = tan -1I C I L

OA2sin -1 for squarer threshold I C

I C

I L = load current

I C s I L

Chap10-152-169 21/06/02 8:46 Page 163


13/18


14/18


Most phase comparison schemes use summationtechniques to produce a single modulating quantity,responsive to faults on any of the three phases. Phasesequence components are often used and a typicalmodulating quantity is:

I M = MI 2 + NI 1 Equation 10.5

whereI 1 = Positive phase sequence componentI 2 = Negative phase sequence componentM , N = constants

With the exception of three phase faults all internalfaults give rise to negative phase sequence (NPS)currents, I 2 , which are approximately in phase at theends of the line and therefore could form an idealmodulating quantity. In order to provide a modulatingsignal during three phase faults, which give rise topositive phase sequence (PPS) currents, I 1 , only, a

practical modulating quantity must include someresponse to I 1 in addition to I 2 .

Typical values of the ratioM: N exceed 5:1, so that themodulating quantity is weighted heavily in favour of NPS, and any PPS associated with load current tends tobe swamped out on all but the highest resistance faults.

For a high resistance phase-earth fault, the systemremains well balanced so that load current I L is entirelypositive sequence. The fault contribution I F providesequal parts of positive, negative and zero sequencecomponents I F /3 . Assuming the fault is on 'A' phase andthe load is resistive, all sequence components are inphase at the infeed end G :

At the outfeed end load current is negative,

Now, forI mH > 0, H = 0, and |G - H | = 0and for

I mH < 0, H = 180, and |G - H | = 180Hence for correct operation I mH 0Let I mH = 0

Then

Equation 10.6

I I

M N

I FH L

E =+

=31

= + +I NI MI NI mH L FH FH 3 3

= + +

and

I NI MI NI

mG LFG FG

G

3 3

0

The fault current in Equation 10.6 is the effective earthfault sensitivity I E of the scheme. For the typical values of

M = 6 and N = -1

Comparing this with Equation 10.4, a scheme usingsummation is potentially 1.667 times more sensitivethan one using phase current for modulation.

Even though the use of a negative value of M gives alower value of I E than if it were positive, it is usuallypreferred since the limiting condition of I m = 0 thenapplies at the load infeed end. Load and faultcomponents are additive at the outfeed end so that acorrect modulating quantity occurs there, even with thelowest fault levels. For operation of the scheme it issufficient therefore that the fault current contributionfrom the load infeed end exceeds the effective setting.

For faults on B or C phases, the NPS components aredisplaced by 120 or 240 with respect to the PPScomponents. No simple cancellation can occur, butinstead a phase displacement is introduced. For trippingto occur, Equation 10.2 must be satisfied, and to achievehigh dependability under these marginal conditions, asmall effective stability angle is essential. Figure 10.15illustrates operation near to the limits of earth faultsensitivity.

Very sensitive schemes may be implemented by usinghigh values of M_ N but the scheme then becomes moresensitive to differential errors in NPS currents such asthe unbalanced components of capacitive current or spillfrom partially saturated CT's.Techniques such as capacitive current compensation andreduction of M_ N at high fault levels may be required toensure stability of the scheme.

10.11.5 Fault Detection and Starting

For a scheme using a carrier system that continuouslytransmits the modulating quantity, protecting an idealline (capacitive current=0 ) in an interconnectedtransmission system, measurement of current magnitudemight be unnecessary. In practice, fault detector orstarting elements are invariably provided and the schemethen becomes a permissive tripping scheme in whichboth the fault detector and the discriminator mustoperate to provide a trip output, and the fault detectormay limit the sensitivity of the scheme. Requirementsfor the fault detectors vary according to the type of carrier channel used, mode of operation used in the

=I I E L35

M N

=6

10

U n

i t P r o t e c t

i o n

F e e

d e r s

Chap10-152-169 21/06/02 8:46 Page 165


15/18


phase angle measurement, that is, blocking orpermissive, and the features used to provide tolerance tocapacitive current.

10.11.6 Normally Quiescent Power Line Carrier(Blocking Mode)

To ensure stability of through faults, it is essential thatcarrier transmission starts before any measurement of the width of the gap is permitted. To allow for

equipment tolerances and the difference in magnitude of the two currents due to capacitive current, two startingelements are used, usually referred to as 'Low Set' and'High Set' respectively. Low Set controls the start-up of transmission whilst High Set, having a setting typically1.5 to 2 times that of the Low Set element, permits thephase angle measurement to proceed.

The use of impulse starters that respond to the change incurrent level enables sensitivities of less than ratedcurrent to be achieved. Resetting of the starters occursnaturally after a swell time or at the clearance of thefault. Dwell times and resetting characteristics must

ensure that during through faults, a High Set is neveroperated when a Low Set has reset and potential raceconditions are often avoided by the transmitting of anunmodulated (and therefore blocking) carrier for a shorttime following the reset of low set; this feature is oftenreferred to as 'Marginal Guard.'

10.11.7 Scheme without CapacitiveCurrent Compensation

The 'keyhole' discrimination characteristic of depends onthe inclusion of a fault detector to ensure that nomeasurements of phase angle can occur at low currentlevels, when the capacitive current might cause largephase shifts. Resetting must be very fast to ensurestability following the shedding of through load.

10.11.8 Scheme with CapacitiveCurrent Compensation (Blocking Mode)

When the magnitude of the modulating quantity is lessthan the threshold of the squarer, transmission if itoccurred, would be a continuous blocking signal. Thismight occur at an end with a weak source, remote froma fault close to a strong source. A fault detector isrequired to permit transmission only when the currentexceeds the modulator threshold by some multiple(typically about 2 times) so that the effective stabilityangle is not excessive. For PLCC schemes, the low setelement referred to in Section 10.11.6 is usually used forthis purpose. If the fault current is insufficient tooperate the fault detector, circuit breaker tripping willnormally occur sequentially.

10.11.9 Fault Detector Operating Quantities

Most faults cause an increase in the corresponding phasecurrent(s) so measurement of current increase could formthe basis for fault detection. However, when a line isheavily loaded and has a low fault level at the outfeedend, some faults can be accompanied by a fall in current,which would lead to failure of such fault detection,resulting in sequential tripping (for blocking modeschemes) or no tripping (for permissive schemes).Although fault detectors can be designed to respond toany disturbance (increase or decrease of current), it ismore usual to use phase sequence components. Allunbalanced faults produce a rise in the NPS componentsfrom the zero level associated with balanced load current,whilst balanced faults produce an increase in the PPScomponents from the load level (except at ends with verylow fault level) so that the use of NPS and PPS faultdetectors make the scheme sensitive to all faults. Forschemes using summation of NPS and PPS componentsfor the modulating quantity, the use of NPS and PPS fault

10

U n

i t P r o t e c t

i o n

F e e

d e r s

1 6 6

Assumptions for examples:Infeed of load I L at end G Outfeed of load I L at end G

System voltage reference

(a) A phase to earth faultI F = 0.9 I E

(c) B phase to earth faultI F = I E (d) C phase to earth faultI F = I E

(b) A phase to earth faultI F = 1.1 I E

N M =-6 therefore I m = 6I 2 - I 2 and from Equation 10.6

NI LH

NI LG NI LG

NI LG NI LG

NI LH NI LH

I mH H =0

I mG I mG

G =180 G =0

| G - H | =180

| G - H | =70

| G - H |=0

G G H

H

NI LH

3

NI E 0.9

3 NI E 0.9

3MI E 0.9

3 NI E 1.1

3 NI E 1.1

3 NI E

3 NI E

3 NI E

3 NI E

3MI E

3MI E

3MI E

3MI E

I mH I mH

I mG I mG

3MI E

0.93

MI E 1.1

3MI E 1.1

effective earth fault sensitivity I E =-35

I Lalso I F1 =

I F 3

120 120

Figure 10.15: Effect of load current on differential phase shift |g - H | for resistive earth faults at the effective earth fault sensitivity I E

Chap10-152-169 21/06/02 8:46 Page 166


16/18


17/18


10.12.2 Unit Protection of a Transformer Feeder

Figure 10.18 shows unit protection applied to atransformer feeder. The feeder is assumed to be a 100mlength of cable, such as might be found in someindustrial plants or where a short distance separates the

33kV and 11kV substations. While 11kV cablecapacitance will exist, it can be regarded as negligible forthe purposes of this example.

The delta/star transformer connection requires phaseshift correction of CT secondary currents across thetransformer, and in this case software equivalents of interposing CTs are used.

Since the LV side quantities lag the HV side quantities by30 , it is necessary to correct this phase shift by using

software CT settings that produce a 30 phase shift.There are two obvious possibilities:

a. HV side:Yd1LV side: Yy0

b. HV side:Yy0LV side: Yd11

Only the second combination is satisfactory, since onlythis one provides the necessary zero-sequence currenttrap to avoid maloperation of the protection scheme forearth faults on the LV side of the transformer outside of the protected zone.

Ratio correction must also be applied, in order to ensurethat the relays see currents from the primary andsecondary sides of the transformer feeder that are wellbalanced under full load conditions. This is not alwaysinherently the case, due to selection of the main CTratios. For the example of Figure 10.18,

Transformer turns ratio at nominal tap

Required turns ratio according to the CT ratios used

= = 400 11250

10 32 .

= =1133

0 3333 .

10

U n

i t P r o t e c t

i o n

F e e

d e r s

1 6 8

I d i f f

I ias

00

1

2

3

4

5

6

7

8

1 2 4 5 6

Figure 10.17: Relay characteristic; plain feeder example

Figure 10.18: Unit protection of a transformer feeder

20 MVA33 11kV

Dyn13kV 11kV

I > I >

Digital communicationchannel

1250 1400 1

50A 1050A

0.875A 0.84A

Ratio correction: 1.14oftware CT:Yy0

Ratio correction: 1.1software CT:Yd11

0 -30

a le100m

Chap10-152-169 21/06/02 8:46 Page 168


18/18


Spill current that will arise due to the incompatibility of the CT ratios used with the power transformer turns ratiomay cause relay maloperation. This has to be eliminatedby using the facility in the relay for CT ratio correctionfactors. For this particular relay, the correction factorsare chosen such that the full load current seen by therelay software is equal to 1A .

The appropriate correction factors are:

HV: 400 / 350 = 1.14

LV: 1250 / 1050 = 1.19

where:

transformer rated primary current = 350A

transformer rated secondary current = 1050A

With the line charging current being negligible, thefollowing relay settings are then suitable, and allow fortransformer efficiency and mismatch due to tap-

changing:I S1 = 20% (minimum possible)

I S1 = 20%

k1 = 30%

k2 = 150%

10.13 REFERENCES

10.1 Merz-Price Protective Gear. K. Faye-Hansen andG. Harlow. IEE Proceedings, 1911.

10.2 Protective Relays Application Guide 3rd Edition. ALSTOM Transmission and DistributionProtection and Control, 1987.

10

U n

i t P r o t e c t

i o n

F e e

d e r s

Chap10-152-169 21/06/02 8:46 Page 169

Documents

10 Unit Protection