17 Generator and Generator Transformer · PDF fileinterconnection between generator and...

Introduction 17.1

Generator earthing 17.2

Stator winding faults 17.3

Stator winding protection 17.4

Differential protection ofdirect-connected generators 17.5

Differential protection of generator –transformer units 17.6

Overcurrent protection 17.7

Stator earth fault protection 17.8

Overvoltage protection 17.9

Undervoltage protection 17.10

Low forward power/reversepower protection 17.11

Unbalanced loading 17.12

Protection against inadvertent energisation 17.13

Under/Overfrequency/Overfluxing protection 17.14

Rotor faults 17.15

Loss of excitation protection 17.16

Pole slipping protection 17.17

Overheating 17.18

Mechanical faults 17.19

Complete generator protection schemes 17.20

Embedded generation 17.21

Examples of generator protection settings 17.22

• 1 7 • G e n e r a t o r a n dG e n e r a t o r T r a n s f o r m e r P r o t e c t i o n

17.1 INTRODUCTION

The core of an electric power system is the generation.With the exception of emerging fuel cell and solar-celltechnology for power systems, the conversion of thefundamental energy into its electrical equivalentnormally requires a 'prime mover' to develop mechanicalpower as an intermediate stage.

The nature of this machine depends upon the source ofenergy and in turn this has some bearing on the designof the generator. Generators based on steam, gas, wateror wind turbines, and reciprocating combustion enginesare all in use. Electrical ratings extend from a fewhundred kVA (or even less) for reciprocating engine andrenewable energy sets, up to steam turbine setsexceeding 1200MVA.

Small and medium sized sets may be directly connectedto a power distribution system. A larger set may beassociated with an individual transformer, throughwhich it is coupled to the EHV primary transmissionsystem.

Switchgear may or may not be provided between thegenerator and transformer. In some cases, operationaland economic advantages can be attained by providinga generator circuit breaker in addition to a high voltagecircuit breaker, but special demands will be placed onthe generator circuit breaker for interruption ofgenerator fault current waveforms that do not have anearly zero crossing.

A unit transformer may be tapped off theinterconnection between generator and transformer forthe supply of power to auxiliary plant, as shown inFigure 17.1. The unit transformer could be of the orderof 10% of the unit rating for a large fossil-fuelled steamset with additional flue-gas desulphurisation plant, butit may only be of the order of 1% of unit rating for ahydro set.

• 17 • G ene ra tor an dG ene ra t or-Tran s f or me r P rote ct ion

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 • 2 8 1 •

Industrial or commercial plants with a requirement forsteam/hot water now often include generating plantutilising or producing steam to improve overalleconomics, as a Combined Heat and Power (CHP)scheme. The plant will typically have a connection to thepublic Utility distribution system, and such generation isreferred to as ‘embedded’ generation. The generatingplant may be capable of export of surplus power, orsimply reduce the import of power from the Utility. Thisis shown in Figure 17.2.

A modern generating unit is a complex systemcomprising the generator stator winding, associatedtransformer and unit transformer (if present), the rotorwith its field winding and excitation system, and theprime mover with its associated auxiliaries. Faults ofmany kinds can occur within this system for whichdiverse forms of electrical and mechanical protection are

required. The amount of protection applied will begoverned by economic considerations, taking intoaccount the value of the machine, and the value of itsoutput to the plant owner.

The following problems require consideration from thepoint of view of applying protection:

a. stator electrical faults

b. overload

c. overvoltage

d. unbalanced loading

e. overfluxing

f. inadvertent energisation

e. rotor electrical faults

f. loss of excitation

g. loss of synchronism

h. failure of prime mover

j. lubrication oil failure

l. overspeeding

m. rotor distortion

n. difference in expansion between rotating andstationary parts

o. excessive vibration

p. core lamination faults

17.2 GENERATOR EARTHING

The neutral point of a generator is usually earthed tofacilitate protection of the stator winding and associatedsystem. Earthing also prevents damaging transientovervoltages in the event of an arcing earth fault orferroresonance.

For HV generators, impedance is usually inserted in thestator earthing connection to limit the magnitude ofearth fault current. There is a wide variation in the earthfault current chosen, common values being:

1. rated current

2. 200A-400A (low impedance earthing)

3. 10A-20A (high impedance earthing)

The main methods of impedance-earthing a generatorare shown in Figure 17.3. Low values of earth faultcurrent may limit the damage caused from a fault, butthey simultaneously make detection of a fault towardsthe stator winding star point more difficult. Except forspecial applications, such as marine, LV generators arenormally solidly earthed to comply with safetyrequirements. Where a step-up transformer is applied,

• 17 •

Gen

erat

or a

nd G

ener

ator

-Tra

nsfo

rmer

Pro

tect

ion

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• 2 8 2 •

Generator Main transformer

HV busbars

Unit transformer

Auxiliarysupplies switchboard

Figure 17.1: Generator-transformer unit

Utility

PCC

Industrial plantmain busbar

Plant feeders - totaldemand: xMW

When plant generator is running:If y>x, Plant may export to Utility across PCCIf x>y, Plant max demand from Utility is reduced

PCC: Point of Common Coupling

GeneratorRating: yMW

Figure 17.2: Embedded generation

the generator and the lower voltage winding of thetransformer can be treated as an isolated system that isnot influenced by the earthing requirements of thepower system.

An earthing transformer or a series impedance can beused as the impedance. If an earthing transformer isused, the continuous rating is usually in the range 5-250kVA. The secondary winding is loaded with a resistorof a value which, when referred through the transformerturns ratio, will pass the chosen short-time earth-faultcurrent. This is typically in the range of 5-20A. Theresistor prevents the production of high transientovervoltages in the event of an arcing earth fault, whichit does by discharging the bound charge in the circuitcapacitance. For this reason, the resistive component offault current should not be less than the residualcapacitance current. This is the basis of the design, andin practice values of between 3-5 Ico

Conventional generator protection systems would beblind to an interturn fault, but the extra cost andcomplication of providing detection of a purely interturnfault is not usually justified. In this case, an interturnfault must develop into an earth fault before it can becleared. An exception may be where a machine has anabnormally complicated or multiple windingarrangement, where the probability of an interturn faultmight be increased.

17.4 STATOR WINDING PROTECTION

To respond quickly to a phase fault with damaging heavycurrent, sensitive, high-speed differential protection isnormally applied to generators rated in excess of 1MVA.For large generating units, fast fault clearance will alsomaintain stability of the main power system. The zoneof differential protection can be extended to include anassociated step-up transformer. For smaller generators,IDMT/instantaneous overcurrent protection is usually theonly phase fault protection applied. Sections 17.5-17.8detail the various methods that are available for statorwinding protection.

17.5 DIFFERENTIAL PROTECTION OF DIRECTCONNECTED GENERATORS

The theory of circulating current differential protection isdiscussed fully in Section 10.4.

High-speed phase fault protection is provided, by use ofthe connections shown in Figure 17.4. This depicts thederivation of differential current through CT secondarycircuit connections. This protection may also offer earthfault protection for some moderate impedance-earthedapplications. Either biased differential or highimpedance differential techniques can be applied. Asubtle difference with modern, biased, numericalgenerator protection relays is that they usually derive thedifferential currents and biasing currents by algorithmic

calculation, after measurement of the individual CTsecondary currents. In such relay designs, there is fullgalvanic separation of the neutral-tail and terminal CTsecondary circuits, as indicated in Figure 17.5(a). This isnot the case for the application of high-impedancedifferential protection. This difference can impose somespecial relay design requirements to achieve stability forbiased differential protection in some applications.

17.5.1 Biased Differential Protection

The relay connections for this form of protection areshown in Figure 17.5(a) and a typical bias characteristicis shown in Figure 17.5(b). The differential currentthreshold setting Is1 can be set as low as 5% of ratedgenerator current, to provide protection for as much ofthe winding as possible. The bias slope break-pointthreshold setting Is2 would typically be set to a valueabove generator rated current, say 120%, to achieveexternal fault stability in the event of transientasymmetric CT saturation. Bias slope K2 setting wouldtypically be set at 150%.

17.5.2 High Impedance Differential Protection

This differs from biased differential protection by themanner in which relay stability is achieved for externalfaults and by the fact that the differential current mustbe attained through the electrical connections of CTsecondary circuits. If the impedance of each relay inFigure 17.4 is high, the event of one CT becomingsaturated by the through fault current (leading to a

• 17 •

Gen

erat

or a

nd G

ener

ator

-Tra

nsfo

rmer

Pro

tect

ion


(a): Relay connections for biased differential protection

Operate

Restrain

(b) Biased differential operating characteristic

I1

IdiffdiffIdiff= I1+I2I

I2I

IS1I

IS2I

K1

K2K

IBIASI = I1+ 2

Figure 17.5: Typical generator biaseddifferential protection

C

A

B

IdI > IdI > IdI >

Stator

Figure 17.4: Stator differential protection

relatively low CT impedance), will allow the current fromthe unsaturated CT to flow mainly through the saturatedCT rather than through the relay. This provides therequired protection stability where a tuned relay elementis employed. In practice, external resistance is added tothe relay circuit to provide the necessary highimpedance. The principle of high-impedance protectionapplication is illustrated in Figure 17.6, together with asummary of the calculations required to determine thevalue of external stabilising resistance.

In some applications, protection may be required to limitvoltages across the CT secondary circuits when thedifferential secondary current for an internal phase faultflows through the high impedance relay circuit(s), butthis is not commonly a requirement for generatordifferential applications unless very high impedancerelays are applied. Where necessary, shunt–connected,non-linear resistors, should be deployed, as shown inFigure 17.7.

To calculate the primary operating current, the followingexpression is used:

Iop = N x (Is1 + nIe)

where:

Iop = primary operating current

N = CT ratio

Is1 = relay setting

n = number of CT’s in parallel with relay element

Ie = CT magnetising current at Vs

Is1 is typically set to 5% of generator rated secondarycurrent.

It can be seen from the above that the calculations forthe application of high impedance differential protectionare more complex than for biased differential protection.However, the protection scheme is actually quite simpleand it offers a high level of stability for through faultsand external switching events.

With the advent of multi-function numerical relays andwith a desire to dispense with external components, highimpedance differential protection is not as popular asbiased differential protection in modern relayingpractice.

17.5.3 CT Requirements

The CT requirements for differential protection will varyaccording to the relay used. Modern numerical relaysmay not require CT’s specifically designed for differentialprotection to IEC 60044-1 class PX (or BS 3938 class X).However, requirements in respect of CT knee-pointvoltage will still have to be checked for the specificrelays used. High impedance differential protection maybe more onerous in this respect than biased differentialprotection.

Many factors affect this, including the other protectionfunctions fed by the CT’s and the knee-pointrequirements of the particular relay concerned. Relaymanufacturers are able to provide detailed guidance onthis matter.

17.6 DIFFERENTIAL PROTECTION OFGENERATOR-TRANSFORMERS

A common connection arrangement for large generatorsis to operate the generator and associated step-uptransformer as a unit without any intervening circuitbreaker. The unit transformer supplying the generatorauxiliaries is tapped off the connection betweengenerator and step-up transformer. Differentialprotection can be arranged as follows.

• 17 •G

ener

ator

and

Gen

erat

or-T

rans

form

er P

rote

ctio

n


RstNLR = Non-linear resistance

(Metrosil)

NLRNLR

V

Figure 17.7: Relay connections for highimpedance differential protection

Healthy CT Saturated CTProtected zone

Id>

Zm

RCT1 RCT2

RL1

Rst

RL3

RL2 RL4

If

Vr

Voltage across relay circuit Vr = If (RCT + 2RL) and Vs = KVr

where 1.0<K≤1.5Stabilising resistor, Rst, limits spill current to <Is (relay setting)

when RR = relay burden

VsIs

Rst = -RR

Figure 17.6: Principle of high impedancedifferential protection

17.6.1 Generator/Step-up TransformerDifferential Protection

The generator stator and step-up transformer can beprotected by a single zone of overall differentialprotection (Figure 17.8). This will be in addition todifferential protection applied to the generator only. Thecurrent transformers should be located in the generatorneutral connections and in the transformer HVconnections. Alternatively, CT’s within the HVswitchyard may be employed if the distance is nottechnically prohibitive. Even where there is a generatorcircuit breaker, overall differential protection can still beprovided if desired.

The current transformers should be rated according toSection 16.8.2. Since a power transformer is includedwithin the zone of protection, biased transformerdifferential protection, with magnetising inrush restraintshould be applied, as discussed in Section 16.8.5.Transient overfluxing of the generator transformer mayarise due to overvoltage following generator loadrejection. In some applications, this may threaten thestability of the differential protection. In such cases,consideration should be given to applying protectionwith transient overfluxing restraint/blocking (e.g. basedon a 5th harmonic differential current threshold).Protection against sustained overfluxing is covered inSection 17.14.

17.6.2 Unit Transformer Differential Protection

The current taken by the unit transformer must beallowed for by arranging the generator differentialprotection as a three-ended scheme. Unit transformercurrent transformers are usually applied to balance thegenerator differential protection and prevent the unittransformer through current being seen as differentialcurrent. An exception might be where the unit

transformer rating is extremely low in relation to thegenerator rating, e.g. for some hydro applications. Thelocation of the third set of current transformers isnormally on the primary side of the unit transformer. Iflocated on secondary side of the unit transformer, theywould have to be of an exceptionally high ratio, orexceptionally high ratio interposing CT’s would have tobe used. Thus, the use of secondary side CT’s is not to berecommended. One advantage is that unit transformerfaults would be within the zone of protection of thegenerator. However, the sensitivity of the generatorprotection to unit transformer phase faults would beconsidered inadequate, due to the relatively low rating ofthe transformer in relation to that of the generator.Thus, the unit transformer should have its owndifferential protection scheme. Protection for the unittransformer is covered in Chapter 16, including methodsfor stabilising the protection against magnetising inrushconditions.

17.7 OVERCURRENT PROTECTION

Overcurrent protection of generators may take twoforms. Plain overcurrent protection may be used as theprinciple form of protection for small generators, andback-up protection for larger ones where differentialprotection is used as the primary method of generatorstator winding protection. Voltage dependentovercurrent protection may be applied where differentialprotection is not justified on larger generators, or whereproblems are met in applying plain overcurrentprotection.

17.7.1 Plain Overcurrent Protection

It is usual to apply time-delayed plain overcurrentprotection to generators. For generators rated less than1MVA, this will form the principal stator windingprotection for phase faults. For larger generators,overcurrent protection can be applied as remote back-upprotection, to disconnect the unit from any unclearedexternal fault. Where there is only one set of differentialmain protection, for a smaller generator, the overcurrentprotection will also provide local back-up protection forthe protected plant, in the event that the mainprotection fails to operate. The general principles ofsetting overcurrent relays are given in Chapter 9.

In the case of a single generator feeding an isolatedsystem, current transformers at the neutral end of themachine should energise the overcurrent protection, toallow a response to winding fault conditions. Relaycharacteristics should be selected to take into accountthe fault current decrement behaviour of the generator,with allowance for the performance of the excitation

• 17 •

Gen

erat

or a

nd G

ener

ator

-Tra

nsfo

rmer

Pro

tect

ion


GeneratorMain

transformer

HVbusbars

IdI >

Protected zone

Figure 17.8: Overall generator-transformerdifferential protection

system and its field-forcing capability. Without theprovision of fault current compounding from generatorCT’s, an excitation system that is powered from anexcitation transformer at the generator terminals willexhibit a pronounced fault current decrement for aterminal fault. With failure to consider this effect, thepotential exists for the initial high fault current to decayto a value below the overcurrent protection pick-upsetting before a relay element can operate, unless a lowcurrent setting and/or time setting is applied. Theprotection would then fail to trip the generator. Thesettings chosen must be the best compromise betweenassured operation in the foregoing circumstances anddiscrimination with the system protection and passageof normal load current, but this can be impossible withplain overcurrent protection.

In the more usual case of a generator that operates inparallel with others and which forms part of an extensiveinterconnected system, back-up phase fault protectionfor a generator and its transformer will be provided by HVovercurrent protection. This will respond to the higher-level backfeed from the power system to a unit fault.Other generators in parallel would supply this currentand, being stabilised by the system impedance, it will notsuffer a major decrement. This protection is usually arequirement of the power system operator. Settings mustbe chosen to prevent operation for external faults fed bythe generator. It is common for the HV overcurrentprotection relay to provide both time-delayed andinstantaneous high-set elements. The time-delayedelements should be set to ensure that the protected itemsof plant cannot pass levels of through fault current inexcess of their short-time withstand limits. Theinstantaneous elements should be set above themaximum possible fault current that the generator cansupply, but less than the system-supplied fault current inthe event of a generator winding fault. This back-upprotection will minimise plant damage in the event ofmain protection failure for a generating plant fault andinstantaneous tripping for an HV-side fault will aid therecovery of the power system and parallel generation.

17.7.2 Voltage-Dependent Overcurrent Protection

The plain overcurrent protection setting difficultyreferred to in the previous section arises becauseallowance has to be made both for the decrement of thegenerator fault current with time and for the passage offull load current. To overcome the difficulty ofdiscrimination, the generator terminal voltage can bemeasured and used to dynamically modify the basic relaycurrent/time overcurrent characteristic for faults close tothe generating plant. There are two basic alternativesfor the application of voltage-dependent overcurrentprotection, which are discussed in the following sections.

The choice depends upon the power systemcharacteristics and level of protection to be provided.Voltage-dependent overcurrent relays are often foundapplied to generators used on industrial systems as analternative to full differential protection.

17.7.2.1 Voltage controlled overcurrent protection

Voltage controlled overcurrent protection has twotime/current characteristics which are selected accordingto the status of a generator terminal voltage measuringelement. The voltage threshold setting for the switchingelement is chosen according to the following criteria.

1. during overloads, when the system voltage issustained near normal, the overcurrent protectionshould have a current setting above full load currentand an operating time characteristic that will preventthe generating plant from passing current to a remoteexternal fault for a period in excess of the plant short-time withstand limits

2. under close-up fault conditions, the busbar voltagemust fall below the voltage threshold so that thesecond protection characteristic will be selected. Thischaracteristic should be set to allow relay operationwith fault current decrement for a close-up fault atthe generator terminals or at the HV busbars. Theprotection should also time-grade with externalcircuit protection. There may be additional infeeds toan external circuit fault that will assist with grading

Typical characteristics are shown in Figure 17.9.

17.7.2.2 Voltage restrained overcurrent protection

The alternative technique is to continuously vary therelay element pickup setting with generator voltagevariation between upper and lower limits. The voltage issaid to restrain the operation of the current element.

The effect is to provide a dynamic I.D.M.T. protectioncharacteristic, according to the voltage at the machine

• 17 •G

ener

ator

and

Gen

erat

or-T

rans

form

er P

rote

ctio

n


Current pick-up level

I >

KI>

Vs Voltage level

Figure 17.9: Voltage controlled relaycharacteristics

terminals. Alternatively, the relay element may beregarded as an impedance type with a long dependenttime delay. In consequence, for a given fault condition,the relay continues to operate more or lessindependently of current decrement in the machine. Atypical characteristic is shown in Figure 17.10.

17.8 STATOR EARTH FAULT PROTECTION

Earth fault protection must be applied where impedanceearthing is employed that limits the earth fault currentto less than the pick-up threshold of the overcurrentand/or differential protection for a fault located down tothe bottom 5% of the stator winding from the star-point. The type of protection required will depend on themethod of earthing and connection of the generator tothe power system.

17.8.1 Direct-Connected Generators

A single direct-connected generator operating on anisolated system will normally be directly earthed.However, if several direct-connected generators areoperated in parallel, only one generator is normallyearthed at a time. For the unearthed generators, asimple measurement of the neutral current is notpossible, and other methods of protection must be used.The following sections describe the methods available.

17.8.1.1 Neutral overcurrent protection

With this form of protection, a current transformer in theneutral-earth connection energises an overcurrent relayelement. This provides unrestricted earth-faultprotection and so it must be graded with feederprotection. The relay element will therefore have a time-delayed operating characteristic. Grading must becarried out in accordance with the principles detailed inChapter 9. The setting should not be more than 33% ofthe maximum earth fault current of the generator, and alower setting would be preferable, depending on grading

considerations.

17.8.1.2 Sensitive earth fault protection

This method is used in the following situations:

a. direct-connected generators operating in parallel

b. generators with high-impedance neutral earthing,the earth fault current being limited to a few tensof amps

c. installations where the resistance of the groundfault path is very high, due to the nature of theground

In these cases, conventional earth fault protection asdescribed in Section 17.8.1.1 is of little use.

The principles of sensitive earth fault protection aredescribed in Sections 9.17.1, 9.18 and 9.19. The earthfault (residual) current can be obtained from residualconnection of line CT’s, a line-connected CBCT, or a CT inthe generator neutral. The latter is not possible ifdirectional protection is used. The polarising voltage isusually the neutral voltage displacement input to therelay, or the residual of the three phase voltages, so asuitable VT must be used. For Petersen Coil earthing, awattmetric technique (Section 9.19) can also be used.

For direct connected generators operating in parallel,directional sensitive earth fault protection may benecessary. This is to ensure that a faulted generator willbe tripped before there is any possibility of the neutralovercurrent protection tripping a parallel healthygenerator. When being driven by residually-connectedphase CT’s, the protection must be stabilised againstincorrect tripping with transient spill current in the eventof asymmetric CT saturation when phase fault ormagnetising inrush current is being passed. Stabilisingtechniques include the addition of relay circuitimpedance and/or the application of a time delay. Wherethe required setting of the protection is very low incomparison to the rated current of the phase CT’s, itwould be necessary to employ a single CBCT for the earthfault protection to ensure transient stability.

Since any generator in the paralleled group may beearthed, all generators will require to be fitted with bothneutral overcurrent protection and sensitive directionalearth fault protection.

The setting of the sensitive directional earth faultprotection is chosen to co-ordinate with generatordifferential protection and/or neutral voltagedisplacement protection to ensure that 95% of the statorwinding is protected. Figure 17.11 illustrates thecomplete scheme, including optional blocking signalswhere difficulties in co-ordinating the generator anddownstream feeder earth-fault protection occur.

• 17 •

Gen

erat

or a

nd G

ener

ator

-Tra

nsfo

rmer

Pro

tect

ion


Current pick-up level

I>

KI>

VS2VS2V VS1V Voltage level

Figure 17.10: Voltage restrained relaycharacteristics

For cases (b) and (c) above, it is not necessary to use adirectional facility. Care must be taken to use the correctRCA setting – for instance if the earthing impedance ismainly resistive, this should be 0°. On insulated or veryhigh impedance earthed systems, an RCA of -90° wouldbe used, as the earth fault current is predominatelycapacitive.

Directional sensitive earth-fault protection can also beused for detecting winding earth faults. In this case, therelay element is applied to the terminals of the generatorand is set to respond to faults only within the machinewindings. Hence earth faults on the external system donot result in relay operation. However, current flowingfrom the system into a winding earth fault causes relayoperation. It will not operate on the earthed machine, sothat other types of earth fault protection must also beapplied. All generators must be so fitted, since any canbe operated as the earthed machine.

17.8.1.3 Neutral voltage displacement protection

In a balanced network, the addition of the three phase-earth voltages produces a nominally zero residualvoltage, since there would be little zero sequence voltagepresent. Any earth fault will set up a zero sequencesystem voltage, which will give rise to a non-zeroresidual voltage. This can be measured by a suitablerelay element. The voltage signal must be derived froma VT that is suitable – i.e. it must be capable oftransforming zero-sequence voltage, so 3-limb types andthose without a primary earth connection are notsuitable. This unbalance voltage provides a means ofdetecting earth faults. The relay element must beinsensitive to third harmonic voltages that may bepresent in the system voltage waveforms, as these will

sum residually.

As the protection is still unrestricted, the voltage settingof the relay must be greater than the effective setting ofany downstream earth-fault protection. It must also betime-delayed to co-ordinate with such protection.Sometimes, a second high-set element with short timedelay is used to provide fast-acting protection againstmajor winding earth-faults. Figure 17.12 illustrates thepossible connections that may be used.

17.8.2 Indirectly-Connected Generators

As noted in Section 17.2, a directly-earthed generator-transformer unit cannot interchange zero-sequencecurrent with the remainder of the network, and hence anearth fault protection grading problem does not exist.The following sections detail the protection methods forthe various forms of impedance earthing of generators.

17.8.2.1 High resistance earthing – neutral overcurrentprotection

A current transformer mounted on the neutral-earthconductor can drive an instantaneous and/or timedelayed overcurrent relay element, as shown in Figure17.13. It is impossible to provide protection for the wholeof the winding, and Figure 17.13 also details how thepercentage of winding covered can be calculated. For arelay element with an instantaneous setting, protection istypically limited to 90% of the winding. This is to ensurethat the protection will not maloperate with zerosequence current during operation of a primary fuse for aVT earth fault or with any transient surge currents thatcould flow through the interwinding capacitance of thestep-up transformer for an HV system earth fault.

A time-delayed relay is more secure in this respect, and itmay have a setting to cover 95% of the stator winding.Since the generating units under consideration are usuallylarge, instantaneous and time delayed relay elements are

• 17 •G

ener

ator

and

Gen

erat

or-T

rans

form

er P

rote

ctio

n


I > >>I

Feeder

Block*Block*

ReRe

Open

* Optional interlockedearth-fault protectionif grading problems exist

Minimum earth fault level = IFI Re

VV

I

I >

>

>UrsdUrsdU

I

I >

>

UrsdUrsdU

Figure 17.11: Comprehensive earth-faultprotection scheme for direct-connected

generators operating in parallelVaVaVVbVbV

c

VnVnV

3

2

1

Measured from earth impedance2

Measured from broken delta VT 3

Derived from phase neutral voltages1

Figure 17.12: Neutral voltage displacementprotection

often applied, with settings of 10% and 5% of maximumearth fault current respectively; this is the optimumcompromise in performance. The portion of the windingleft unprotected for an earth fault is at the neutral end.Since the voltage to earth at this end of the winding islow, the probability of an earth fault occurring is also low.Hence additional protection is often not applied.

17.8.2.2 Distribution transformer earthingusing a current element

In this arrangement, shown in Figure 17.14(a), thegenerator is earthed via the primary winding of adistribution transformer. The secondary winding is fittedwith a loading resistor to limit the earth fault current.An overcurrent relay element energised from a currenttransformer connected in the resistor circuit is used tomeasure secondary earth fault current. The relay shouldhave an effective setting equivalent to 5% of themaximum earth fault current at rated generator voltage,in order to protect 95% of the stator winding. The relayelement response to third harmonic current should belimited to prevent incorrect operation when a sensitivesetting is applied.

As discussed in Section 17.8.2.1 for neutral overcurrentprotection, the protection should be time delayed whena sensitive setting is applied, in order to preventmaloperation under transient conditions. It also mustgrade with generator VT primary protection (for a VTprimary earth fault). An operation time in the range0.5s-3s is usual. Less sensitive instantaneous protectioncan also be applied to provide fast tripping for a heavierearth fault condition.

• 17 •

Gen

erat

or a

nd G

ener

ator

-Tra

nsfo

rmer

Pro

tect

ion


R

IsIsI

IfIfI

a

V

IfIfI =aV

R

IsIsI Ramin =

V

%covered 1-a in ) x 1100%

generator stator winding using a current element

Figure 17.13: Earth fault protection of high-resistanceearthed generator stator winding using a current element

17.8.2.3 Distribution transformer earthingusing a voltage element

Earth fault protection can also be provided using a voltage-measuring element in the secondary circuit instead. Thesetting considerations would be similar to those for thecurrent operated protection, but transposed to voltage.The circuit diagram is shown in Figure 17.14(b).

Application of both voltage and current operatedelements to a generator with distribution transformerearthing provides some advantages. The currentoperated function will continue to operate in the eventof a short-circuited loading resistor and the voltageprotection still functions in the event of an open-circuited resistor. However, neither scheme will operatein the event of a flashover on the primary terminals ofthe transformer or of the neutral cable between thegenerator and the transformer during an earth fault. ACT could be added in the neutral connection close to thegenerator, to energise a high-set overcurrent element todetect such a fault, but the fault current would probablybe high enough to operate the phase differentialprotection.

17.8.2.4 Neutral voltage displacement protection

This can be applied in the same manner as for direct-connected generators (Section 17.8.1.3). The only

I >

Loadingresistor

Loadingresistor

U >

(a) Protection using a current element

(b) Protection using a voltage element

Figure 17.14: Generator winding earth-faultprotection - distribution transformer earthing

difference is that the are no grading problems as theprotection is inherently restricted. A sensitive settingcan therefore be used, enabling cover of up to 95% ofthe stator winding to be achieved.

17.8.3 Restricted Earth Fault Protection

This technique can be used on small generators not fittedwith differential protection to provide fast acting earthfault protection within a defined zone that encompassesthe generator. It is cheaper than full differentialprotection but only provides protection against earthfaults. The principle is that used for transformer REFprotection, as detailed in Section 16.7. However, incontrast to transformer REF protection, both biased low-impedance and high-impedance techniques can be used.

17.8.3.1 Low-impedance biased REF protection

This is shown in Figure 17.15. The main advantage isthat the neutral CT can also be used in a modern relay toprovide conventional earth-fault protection and noexternal resistors are used. Relay bias is required, asdescribed in Section 10.4.2, but the formula forcalculating the bias is slightly different and also shownin Figure 17.15.

The initial bias slope is commonly set to 0% to providemaximum sensitivity, and applied up to the rated currentof the generator. It may be increased to counter theeffects of CT mismatch. The bias slope above generatorrated current is typically set to 150% of rated value. Theinitial current setting is typically 5% of the minimumearth fault current for a fault at the machine terminals.

17.8.3.2 High Impedance REF protection

The principle of high impedance differential protection isgiven in Chapter 10 and also described further in Section17.5.2. The same technique can be used for earth-fault

protection of a generator, using three residuallyconnected phase CT’s balanced against a similar singleCT in the neutral connection. Settings of the order of 5%of maximum earth fault current at the generatorterminals are typical. The usual requirements in respectof stabilising resistor and non-linear resistor to guardagainst excessive voltage across the relay must be taken,where necessary.

17.8.4 Earth Fault Protection forthe Entire Stator Winding

All of the methods for earth fault protection detailed sofar leave part of the winding unprotected. In most cases,this is of no consequence as the probability of a faultoccurring in the 5% of the winding nearest the neutralconnection is very low, due to the reduced phase to earthvoltage. However, a fault can occur anywhere along thestator windings in the event of insulation failure due tolocalised heating from a core fault. In cases whereprotection for the entire winding is required, perhaps foralarm only, there are various methods available.

17.8.4.1 Measurement of third harmonic voltage

One method is to measure the internally generated thirdharmonic voltage that appears across the earthingimpedance due to the flow of third harmonic currentsthrough the shunt capacitance of the stator windingsetc. When a fault occurs in the part of the statorwinding nearest the neutral end, the third harmonicvoltage drops to near zero, and hence a relay elementthat responds to third harmonic voltage can be used todetect the condition. As the fault location movesprogressively away from the neutral end, the drop inthird harmonic voltage from healthy conditions becomesless, so that at around 20-30% of the winding distance,it no longer becomes possible to discriminate between ahealthy and a faulty winding. Hence, a conventionalearth-fault scheme should be used in conjunction with athird harmonic scheme, to provide overlapping coverof the entire stator winding. The measurement of thirdharmonic voltage can be taken either from a star-pointVT or the generator line VT. In the latter case, the VTmust be capable of carrying residual flux, and thisprevents the use of 3-limb types. If the third harmonicvoltage is measured at the generator star point, anundervoltage characteristic is used. An overvoltagecharacteristic is used if the measurement is taken fromthe generator line VT. For effective application of thisform of protection, there should be at least 1% thirdharmonic voltage across the generator neutral earthingimpedance under all operating conditions.

A problem encountered is that the level of thirdharmonic voltage generated is related to the output ofthe generator. The voltage is low when generator output

• 17 •G

ener

ator

and

Gen

erat

or-T

rans

form

er P

rote

ctio

n


Neutral CT ratio200/1

Phase BPhase C

Phase APhase CT ratio 1000/1

IBIASI = (highest of IAI B, II N x scaling factor)2

where scaling factor = = = 0.22001000

IDIFF I = IAI IBI ICICI (scaling factor IN )

Figure 17.15: Low impedance biased REFprotection of a generator

is low. In order to avoid maloperation when operating atlow power output, the relay element can be inhibitedusing an overcurrent or power element (kW, kvar or kVA)and internal programmable logic.

17.8.4.2 Use of low-frequency voltage injection

Another method for protecting the entire stator windingof a generator is to deploy signal injection equipment toinject a low frequency voltage between the stator starpoint and earth. An earth fault at any winding locationwill result in the flow of a measurable injection currentto cause protection operation. This form of protectioncan provide earth fault protection when the generator isat standstill, prior to run-up. It is also an appropriatemethod to apply to variable speed synchronousmachines. Such machines may be employed for variablespeed motoring in pumped-storage generation schemesor for starting a large gas turbine prime mover.

17.9 OVERVOLTAGE PROTECTION

Overvoltages on a generator may occur due to transientsurges on the network, or prolonged power frequencyovervoltages may arise from a variety of conditions.Surge arrestors may be required to protect againsttransient overvoltages, but relay protection may be usedto protect against power frequency overvoltages.

A sustained overvoltage condition should not occur for amachine with a healthy voltage regulator, but it may becaused by the following contingencies:

a. defective operation of the automatic voltageregulator when the machine is in isolated operation

b. operation under manual control with the voltageregulator out of service. A sudden variation of theload, in particular the reactive power component,will give rise to a substantial change in voltagebecause of the large voltage regulation inherent ina typical alternator

c. sudden loss of load (due to tripping of outgoingfeeders, leaving the set isolated or feeding a very smallload) may cause a sudden rise in terminal voltage dueto the trapped field flux and/or overspeed

Sudden loss of load should only cause a transientovervoltage while the voltage regulator and governor actto correct the situation. A maladjusted voltage regulatormay trip to manual, maintaining excitation at the valueprior to load loss while the generator supplies little or noload. The terminal voltage will increase substantially,and in severe cases it would be limited only by thesaturation characteristic of the generator. A rise in speedsimply compounds the problem. If load that is sensitiveto overvoltages remains connected, the consequences interms of equipment damage and lost revenue can besevere. Prolonged overvoltages may also occur on

isolated networks, or ones with weak interconnections,due to the fault conditions listed earlier.

For these reasons, it is prudent to provide powerfrequency overvoltage protection, in the form of a time-delayed element, either IDMT or definite time. The timedelay should be long enough to prevent operation duringnormal regulator action, and therefore should takeaccount of the type of AVR fitted and its transientresponse. Sometimes a high-set element is provided aswell, with a very short definite-time delay orinstantaneous setting to provide a rapid trip in extremecircumstances. The usefulness of this is questionable forgenerators fitted with an excitation system other than astatic type, because the excitation will decay inaccordance with the open-circuit time constant of thefield winding. This decay can last several seconds. Therelay element is arranged to trip both the main circuitbreaker (if not already open) and the excitation; trippingthe main circuit breaker alone is not sufficient.

17.10 UNDERVOLTAGE PROTECTION

Undervoltage protection is rarely fitted to generators. Itis sometimes used as an interlock element for anotherprotection function or scheme, such as field failureprotection or inadvertent energisation protection, wherethe abnormality to be detected leads directly orindirectly to an undervoltage condition.

A transmission system undervoltage condition may arisewhen there is insufficient reactive power generation tomaintain the system voltage profile and the conditionmust be addressed to avoid the possible phenomenon ofsystem voltage collapse.

However, it should be addressed by the deployment of’system protection’ schemes. The generation should notbe tripped. The greatest case for undervoltage protectionbeing required would be for a generator supplying anisolated power system or to meet Utility demands forconnection of embedded generation (see Section 17.21).

In the case of generators feeding an isolated system,undervoltage may occur for several reasons, typicallyoverloading or failure of the AVR. In some cases, theperformance of generator auxiliary plant fed via a unittransformer from the generator terminals could beadversely affected by prolonged undervoltage.

Where undervoltage protection is required, it shouldcomprise an undervoltage element and an associatedtime delay. Settings must be chosen to avoidmaloperation during the inevitable voltage dips duringpower system fault clearance or associated with motorstarting. Transient reductions in voltage down to 80% orless may be encountered during motor starting.

• 17 •

Gen

erat

or a

nd G

ener

ator

-Tra

nsfo

rmer

Pro

tect

ion


17.11 LOW FORWARD POWER/REVERSEPOWER PROTECTION

Low forward power or reverse power protection may berequired for some generators to protect the prime mover.Parts of the prime mover may not be designed toexperience reverse torque or they may become damagedthrough continued rotation after the prime mover hassuffered some form of failure.

17.11.1 Low Forward Power Protection

Low forward power protection is often used as aninterlocking function to enable opening of the maincircuit breaker for non-urgent trips – e.g. for a statorearth fault on a high-impedance earthed generator, orwhen a normal shutdown of a set is taking place. This isto minimise the risk of plant overspeeding when theelectrical load is removed from a high-speed cylindricalrotor generator. The rotor of this type of generator ishighly stressed mechanically and cannot tolerate muchoverspeed. While the governor should control overspeedconditions, it is not good practice to open the maincircuit breaker simultaneously with tripping of the primemover for non-urgent trips. For a steam turbine, forexample, there is a risk of overspeeding due to energystorage in the trapped steam, after steam valve tripping,or in the event that the steam valve(s) do not fully closefor some reason. For urgent trip conditions, such asstator differential protection operation, the risk involvedin simultaneous prime mover and generator breakertripping must be accepted.

17.11.2 Reverse Power Protection

Reverse power protection is applied to prevent damageto mechanical plant items in the event of failure of theprime mover. Table 17.1 gives details of the potentialproblems for various prime mover types and the typicalsettings for reverse power protection. For applications

where a protection sensitivity of better than 3% isrequired, a metering class CT should be employed toavoid incorrect protection behaviour due to CT phaseangle errors when the generator supplies a significantlevel of reactive power at close to zero power factor.

The reverse power protection should be provided with adefinite time delay on operation to prevent spuriousoperation with transient power swings that may arisefollowing synchronisation or in the event of a powertransmission system disturbance.

17.12 UNBALANCED LOADING

A three-phase balanced load produces a reaction fieldthat, to a first approximation, is constant and rotatessynchronously with the rotor field system. Anyunbalanced condition can be resolved into positive,negative and zero sequence components. The positivesequence component is similar to the normal balancedload. The zero sequence component produces no mainarmature reaction.

17.12.1 Effect of Negative Sequence Current

The negative sequence component is similar to thepositive sequence system, except that the resultingreaction field rotates in the opposite direction to the d.c.field system. Hence, a flux is produced which cuts therotor at twice the rotational velocity, thereby inducingdouble frequency currents in the field system and in therotor body. The resulting eddy-currents are very largeand cause severe heating of the rotor.

So severe is this effect that a single-phase load equal tothe normal three-phase rated current can quickly heatthe rotor slot wedges to the softening point. They maythen be extruded under centrifugal force until they standabove the rotor surface, when it is possible that they maystrike the stator core.

A generator is assigned a continuous negative sequencerating. For turbo-generators this rating is low; standardvalues of 10% and 15% of the generator continuousrating have been adopted. The lower rating applies whenthe more intensive cooling techniques are applied, forexample hydrogen-cooling with gas ducts in the rotor tofacilitate direct cooling of the winding.

Short time heating is of interest during system faultconditions and it is usual in determining the generatornegative sequence withstand capability to assume thatthe heat dissipation during such periods is negligible.Using this approximation it is possible to express theheating by the law:

I t K22 =

• 17 •G

ener

ator

and

Gen

erat

or-T

rans

form

er P

rote

ctio

n


Table 17.1: Generator reverse power problems

Prime Mover Motoring Power Possible Damage Protection Setting(% of rated)

Fire/explosion dueto unburnt fuel

Mechanical damageto gearbox/shafts

10-15(split shaft)

>50%(single shaft)

0.2-2(blades out of water) blade and runner

>2 cavitation(blades in water)

turbine blade damagegearbox damageon geared sets

50%

of motoring

power

gearbox damage

Diesel Engine

Gas Turbine

Hydro

Steam Turbine

5-25

0.5-6

where:

I2R = negative sequence component(per unit of MCR)

t = time (seconds)

K = constant proportional to the thermal capacityof the generator rotor

For heating over a period of more than a few seconds, itis necessary to allow for the heat dissipated. From acombination of the continuous and short time ratings,the overall heating characteristic can be deduced to be:

where:

I2R = negative phase sequence continuous rating inper unit of MCR

The heating characteristics of various designs ofgenerator are shown in Figure 17.16.

17.12.2 Negative Phase Sequence Protection

This protection is applied to prevent overheating due tonegative sequence currents. Small salient-polegenerators have a proportionately larger negative

MII eR

I t KR

= =− −( )

2

2

1

1 22

sequence capacity and may not require protection.Modern numerical relays derive the negative sequencecurrent level by calculation, with no need for specialcircuits to extract the negative sequence component. Atrue thermal replica approach is often followed, to allowfor:

a. standing levels of negative sequence current belowthe continuous withstand capability. This has theeffect of shortening the time to reach the criticaltemperature after an increase in negative sequencecurrent above the continuous withstand capability

b. cooling effects when negative sequence currentlevels are below the continuous withstandcapability

The advantage of this approach is that cooling effects aremodelled more accurately, but the disadvantage is thatthe tripping characteristic may not follow the withstandcharacteristic specified by the manufacturer accurately.

The typical relay element characteristic takes the form of

…Equation 17.1

where:

Kg = negative sequence withstand coefficient(Figure 17.16)

I2cmr = generator maximum continuous I2 withstand

Iflc = generator rated primary current

Ip = CT primary current

IN = relay rated current

Figure 17.16 also shows the thermal replica timecharacteristic described by Equation 17.1, from which itwill be seen that a significant gain in capability isachieved at low levels of negative sequence current.Such a protection element will also respond to phase-earth and phase-phase faults where sufficient negativesequence current arises. Grading with downstreampower system protection relays is therefore required. Adefinite minimum time setting must be applied to thenegative sequence relay element to ensure correctgrading. A maximum trip time setting may also be usedto ensure correct tripping when the negative sequence

t

K KII

I III

I

K negative

gflc

p

set cmrflc

pn

g

=

= ×

= ×

×

=

=

=

=

=

time to trip

sequence withstand

coefficient (Figure 17.16)

I generator maximum continuous I

withstand

I generator rated primary current

I CT primary current

I relay rated current

2cmr 2

flc

p

n

2

2 2

t KI

IIset

eset=− −

2

22

2

2

1log

• 17 •

Gen

erat

or a

nd G

ener

ator

-Tra

nsfo

rmer

Pro

tect

ion


Tim

e (s

ec)

0.01

0.01

Negative sequence current (p.u.)

0.1

1

10

100

1000

10000

0.1 1 10

Indirectly cooled (air)

Indirectly cooled (H2)

350MW direct cooled

660MW direct cooled

1000MW direct cooled

Using I2I2I2I t modelt

Using true thermalmodel

Figure 17.16: Typical negative phase sequencecurrent withstand of cylindrical

rotor generators

current level is only slightly in excess of the continuouswithstand capability and hence the trip time from thethermal model may depart significantly from the rotorwithstand limits.

17.13 PROTECTION AGAINST INADVERTENTENERGISATION

Accidental energisation of a generator when it is notrunning may cause severe damage to it. With thegenerator at standstill, closing the circuit breaker resultsin the generator acting as an induction motor; the fieldwinding (if closed) and the rotor solid iron/dampercircuits acting as rotor circuits. Very high currents areinduced in these rotor components, and also occur in thestator, with resultant rapid overheating and damage.Protection against this condition is therefore desirable.

A combination of stator undervoltage and overcurrentcan be used to detect this condition. An instantaneousovercurrent element is used, and gated with a three-phase undervoltage element (fed from a VT on thegenerator side of the circuit breaker) to provide theprotection. The overcurrent element can have a lowsetting, as operation is blocked when the generator isoperating normally. The voltage setting should be lowenough to ensure that operation cannot occur fortransient faults. A setting of about 50% of rated voltageis typical. VT failure can cause maloperation of theprotection, so the element should be inhibited underthese conditions.

17.14 UNDER/OVERFREQUENCY/OVERFLUXING PROTECTION

These conditions are grouped together because theseproblems often occur due to a departure fromsynchronous speed.

17.14.1 Overfluxing

Overfluxing occurs when the ratio of voltage tofrequency is too high. The iron saturates owing to thehigh flux density and results in stray flux occurring incomponents not designed to carry it. Overheating canthen occur, resulting in damage. The problem affectsboth direct-and indirectly-connected generators. Eitherexcessive voltage, or low frequency, or a combination ofboth can result in overfluxing, a voltage to frequencyratio in excess of 1.05p.u. normally being indicative ofthis condition. Excessive flux can arise transiently, whichis not a problem for the generator. For example, agenerator can be subjected to a transiently high powerfrequency voltage, at nominal frequency, immediatelyafter full load rejection. Since the condition would notbe sustained, it only presents a problem for the stability

of the transformer differential protection schemesapplied at the power station (see Chapter 16 fortransformer protection). Sustained overfluxing can ariseduring run up, if excitation is applied too early with theAVR in service, or if the generator is run down, with theexcitation still applied. Other overfluxing instances haveoccurred from loss of the AVR voltage feedback signal,due to a reference VT problem. Such sustainedconditions must be detected by a dedicated overfluxingprotection function that will raise an alarm and possiblyforce an immediate reduction in excitation.

Most AVRs’ have an overfluxing protection facilityincluded. This may only be operative when the generatoris on open circuit, and hence fail to detect overfluxingconditions due to abnormally low system frequency.However, this facility is not engineered to protectionrelay standards, and should not be solely relied upon toprovide overfluxing protection. A separate relay elementis therefore desirable and provided in most modernrelays.

It is usual to provide a definite time-delayed alarmsetting and an instantaneous or inverse time-delayedtrip setting, to match the withstand characteristics ofthe protected generator and transformer. It is veryimportant that the VT reference for overfluxingprotection is not the same as that used for the AVR.

17.14.2 Under/Overfrequency

The governor fitted to the prime mover normally providesprotection against overfrequency. Underfrequency mayoccur as a result of overload of generators operating onan isolated system, or a serious fault on the powersystem that results in a deficit of generation comparedto load. This may occur if a grid system suffers a majorfault on transmission lines linking two parts of thesystem, and the system then splits into two. It is likelythat one part will have an excess of generation over load,and the other will have a corresponding deficit.Frequency will fall fairly rapidly in the latter part, and thenormal response is load shedding, either by loadshedding relays or operator action. However, primemovers may have to be protected against excessively lowfrequency by tripping of the generators concerned.

With some prime movers, operation in narrow frequencybands that lie close to normal running speed (eitherabove or below) may only be permitted for short periods,together with a cumulative lifetime duration ofoperation in such frequency bands. This typically occursdue to the presence of rotor torsional frequencies in suchfrequency bands. In such cases, monitoring of the periodof time spent in these frequency bands is required. Aspecial relay is fitted in such cases, arranged to providealarm and trip facilities if either an individual or

• 17 •G

ener

ator

and

Gen

erat

or-T

rans

form

er P

rote

ctio

n


cumulative period exceeds a set time.

17.15 ROTOR FAULTS

The field circuit of a generator, comprising the fieldwinding of the generator and the armature of the exciter,together with any associated field circuit breaker if itexists, is an isolated d.c. circuit which is not normallyearthed. If an earth fault occurs, there will be no steady-state fault current and the need for action will not beevident.

Danger arises if a second earth fault occurs at a separatepoint in the field system, to cause the high field currentto be diverted, in part at least, from the interveningturns. Serious damage to the conductors and possiblythe rotor can occur very rapidly under these conditions.

More damage may be caused mechanically. If a largeportion of the winding is short-circuited, the flux mayadopt a pattern such as that shown in Figure 17.17. Theattracting force at the surface of the rotor is given by:

where:

A = area

B = flux density

It will be seen from Figure 17.17 that the flux isconcentrated on one pole but widely dispersed over theother and intervening surfaces. The attracting force is inconsequence large on one pole but very weak on theopposite one, while flux on the quadrature axis will

F B A=2

8 π

produce a balancing force on this axis. The result is anunbalanced force that in a large machine may be of theorder of 50-100 tons. A violent vibration is set up thatmay damage bearing surfaces or even displace the rotorby an amount sufficient to cause it to foul the stator.

17.15.1 Rotor Earth-Fault Protection

Two methods are available to detect this type of fault.The first method is suitable for generators thatincorporate brushes in the main generator field winding.The second method requires at least a slip-ringconnection to the field circuit:

a. potentiometer method

b. a.c. injection method

17.15.1.1 Potentiometer method

This is a scheme that was fitted to older generators, andit is illustrated in Figure 17.18. An earth fault on thefield winding would produce a voltage across the relay,the maximum voltage occurring for faults at the ends ofthe winding.

A ‘blind spot' would exist at the centre of the fieldwinding. To avoid a fault at this location remainingundetected, the tapping point on the potentiometercould be varied by a pushbutton or switch. The relaysetting is typically about 5% of the exciter voltage.

17.15.1.2 Injection methods

Two methods are in common use. The first is based onlow frequency signal injection, with series filtering, asshown in Figure 17.19(a). It comprises an injectionsource that is connected between earth and one side ofthe field circuit, through capacitive coupling and themeasurement circuit. The field circuit is subjected to analternating potential at substantially the same levelthroughout. An earth fault anywhere in the field systemwill give rise to a current that is detected as anequivalent voltage across the adjustable resistor by therelay. The capacitive coupling blocks the normal d.c. fieldvoltage, preventing the discharge of a large directcurrent through the protection scheme. The combination

• 17 •

Gen

erat

or a

nd G

ener

ator

-Tra

nsfo

rmer

Pro

tect

ion


Short CircuitField Winding

Figure 17.17: Flux distribution on rotorwith partial winding short circuit

Fieldwinding >I Exciter

Figure 17.18: Earth fault protection of fieldcircuit by potentiometer method

of series capacitor and reactor forms a low-pass tunedcircuit, the intention being to filter higher frequencyrotor currents that may occur for a variety of reasons.

Other schemes are based on power frequency signalinjection. An impedance relay element is used, a fieldwinding earth fault reducing the impedance seen by therelay. These suffer the draw back of being susceptible tostatic excitation system harmonic currents when there issignificant field winding and excitation system shuntcapacitance.

Greater immunity for such systems is offered bycapacitively coupling the protection scheme to both endsof the field winding, where brush or slip ring access ispossible (Figure 17.19(b)).

The low–frequency injection scheme is alsoadvantageous in that the current flow through the fieldwinding shunt capacitance will be lower than for apower frequency scheme. Such current would flowthrough the machine bearings to cause erosion of thebearing surface. For power frequency schemes, asolution is to insulate the bearings and provide anearthing brush for the shaft.

17.15.2 Rotor Earth Fault Protectionfor Brushless Generators

A brushless generator has an excitation systemconsisting of:

1. a main exciter with rotating armature andstationary field windings

2. a rotating rectifier assembly, carried on the mainshaft line out

3. a controlled rectifier producing the d.c. fieldvoltage for the main exciter field from an a.c.source (often a small ‘pilot’ exciter)

Hence, no brushes are required in the generator fieldcircuit. All control is carried out in the field circuit of themain exciter. Detection of a rotor circuit earth fault isstill necessary, but this must be based on a dedicatedrotor-mounted system that has a telemetry link toprovide an alarm/data.

17.15.3 Rotor Shorted Turn Protection

As detailed in Section 17.15 a shorted section of fieldwinding will result in an unsymmetrical rotor fluxpattern and in potentially damaging rotor vibration.Detection of such an electrical fault is possible using aprobe consisting of a coil placed in the airgap. The fluxpattern of the positive and negative poles is measuredand any significant difference in flux pattern betweenthe poles is indicative of a shorted turn or turns.Automated waveform comparison techniques can beused to provide a protection scheme, or the waveformcan be inspected visually at regular intervals. Animmediate shutdown is not normally required unless theeffects of the fault are severe. The fault can be keptunder observation until a suitable shutdown for repaircan be arranged. Repair will take some time, since itmeans unthreading the rotor and dismantling thewinding.

Since short-circuited turns on the rotor may causedamaging vibration and the detection of field faults forall degrees of abnormality is difficult, the provision of avibration a detection scheme is desirable – this formspart of the mechanical protection of the generator.

17.15.4 Protection against Diode Failure

A short-circuited diode will produce an a.c. ripple in theexciter field circuit. This can be detected by a relaymonitoring the current in the exciter field circuit,however such systems have proved to be unreliable. Therelay would need to be time delayed to prevent an alarmbeing issued with normal field forcing during a powersystem fault. A delay of 5-10 seconds may be necessary.

• 17 •G

ener

ator

and

Gen

erat

or-T

rans

form

er P

rote

ctio

n


L.F. injectionsupply

Injectionsupply

ExciterGeneratorfieldwinding

ExciterGeneratorfieldwinding

>U

(a) Low frequency a.c. voltage injection - current measurement

<Z<

(b) Power frequency a.c. voltage injection - impedance measurement

∼∼∼

Figure 17.19: Earth fault protectionof field circuit by a.c. injection

Fuses to disconnect the faulty diode after failure may befitted. The fuses are of the indicating type, and aninspection window can be fitted over the diode wheel toenable diode health to be monitored manually.

A diode that fails open-circuit occurs less often. If thereis more than one diode in parallel for each arm of thediode bridge, the only impact is to restrict the maximumcontinuous excitation possible. If only a single diode perbridge arm is fitted, some ripple will be present on themain field supply but the inductance of the circuit willsmooth this to a degree and again the main effect is torestrict the maximum continuous excitation. The set canbe kept running until a convenient shutdown can bearranged.

17.15.5 Field Suppression

The need to rapidly suppress the field of a machine inwhich a fault has developed should be obvious, becauseas long as the excitation is maintained, the machine willfeed its own fault even though isolated from the powersystem. Any delay in the decay of rotor flux will extendthe fault damage. Braking the rotor is no solution,because of its large kinetic energy.

The field winding current cannot be interruptedinstantaneously as it flows in a highly inductive circuit.Consequently, the flux energy must be dissipated toprevent an excessive inductive voltage rise in the fieldcircuit. For machines of moderate size, it is satisfactoryto open the field circuit with an air-break circuit breakerwithout arc blow-out coils. Such a breaker permits onlya moderate arc voltage, which is nevertheless highenough to suppress the field current fairly rapidly. Theinductive energy is dissipated partly in the arc and partlyin eddy-currents in the rotor core and damper windings.

With generators above about 5MVA rating, it is better toprovide a more definite means of absorbing the energywithout incurring damage. Connecting a ‘field dischargeresistor’ in parallel with the rotor winding before openingthe field circuit breaker will achieve this objective. Theresistor, which may have a resistance value ofapproximately five times the rotor winding resistance, isconnected by an auxiliary contact on the field circuitbreaker. The breaker duty is thereby reduced to that ofopening a circuit with a low L/R ratio. After the breakerhas opened, the field current flows through the dischargeresistance and dies down harmlessly. The use of a fairlyhigh value of discharge resistance reduces the field timeconstant to an acceptably low value, though it may stillbe more than one second. Alternatively, generatorsfitted with static excitation systems may temporarilyinvert the applied field voltage to reduce excitationcurrent rapidly to zero before the excitation system istripped.

17.16 LOSS OF EXCITATION PROTECTION

Loss of excitation may occur for a variety of reasons. Ifthe generator was initially operating at only 20%-30%of rated power, it may settle to run super-synchronouslyas an induction generator, at a low level of slip. In doingso, it will draw reactive current from the power systemfor rotor excitation. This form of response is particularlytrue of salient pole generators. In these circumstances,the generator may be able to run for several minuteswithout requiring to be tripped. There may be sufficienttime for remedial action to restore the excitation, but thereactive power demand of the machine during the failuremay severely depress the power system voltage to anunacceptable level. For operation at high initial poweroutput, the rotor speed may rise to approximately 105%of rated speed, where there would be low power outputand where a high reactive current of up to 2.0p.u. maybe drawn from the supply. Rapid automaticdisconnection is then required to protect the statorwindings from excessive current and to protect the rotorfrom damage caused by induced slip frequency currents.

17.16.1 Protection against Loss of Excitation

The protection used varies according to the size ofgenerator being protected.

17.16.1.1 Small generators

On the smaller machines, protection againstasynchronous running has tended to be optional, but itmay now be available by default, where the functionalityis available within a modern numerical generatorprotection package. If fitted, it is arranged either toprovide an alarm or to trip the generator. If thegenerator field current can be measured, a relay elementcan be arranged to operate when this drops below apreset value. However, depending on the generatordesign and size relative to the system, it may well be thatthe machine would be required to operate synchronouslywith little or no excitation under certain systemconditions.

The field undercurrent relay must have a setting belowthe minimum exciting current, which may be 8% of thatcorresponding to the MCR of the machine. Time delayrelays are used to stabilise the protection againstmaloperation in response to transient conditions and toensure that field current fluctuations due to pole slippingdo not cause the protection to reset.

If the generator field current is not measurable, then thetechnique detailed in the following section is utilised.

17.16.1.2 Large generators (>5MVA)

For generators above about 5MVA rating, protectionagainst loss of excitation and pole slipping conditions isnormally applied.

• 17 •

Gen

erat

or a

nd G

ener

ator

-Tra

nsfo

rmer

Pro

tect

ion


Consider a generator connected to network, as shown inFigure 17.20. On loss of excitation, the terminal voltagewill begin to decrease and the stator current will increase,resulting in a decrease of impedance viewed from thegenerator terminals and also a change in power factor.

A relay to detect loss of synchronism can be located atpoint A. It can be shown that the impedance presentedto the relay under loss of synchronism conditions (phaseswinging or pole slipping) is given by:

…Equation 17.2

where:

θ = angle by which EG leads Es

If the generator and system voltages are equal, the aboveexpression becomes:

ZX X Z j

XRG T S

G=+ +( ) −( )

−1 2

2

cotθ

n EE voltageG

S= =

generatedsystem

ZX X Z n n j

n

X

RG T S

G

=+ +( ) − −( )

−( ) +

−

cos sin

cos sin

θ θ

θ θ2 2

The general case can be represented by a system ofcircles with centres on the line CD; see Figure 17.21.Also shown is a typical machine terminal impedancelocus during loss of excitation conditions.

The special cases of EG=ES and EG=0 result in astraight-line locus that is the right-angled bisector ofCD, and in a circular locus that is shrunk to point C,respectively.

When excitation is removed from a generator operatingsynchronously the flux dies away slowly, during whichperiod the ratio of EG/ES is decreasing, and the rotor angleof the machine is increasing. The operating conditionplotted on an impedance diagram therefore travels alonga locus that crosses the power swing circles. At the sametime, it progresses in the direction of increasing rotorangle. After passing the anti-phase position, the locusbends round as the internal e.m.f. collapses, condensing onan impedance value equal to the machine reactance. Thelocus is illustrated in Figure 17.21.

The relay location is displaced from point C by thegenerator reactance XG. One problem in determining theposition of these loci relative to the relay location is thatthe value of machine impedance varies with the rate ofslip. At zero slip XG is equal to Xd, the synchronousreactance, and at 100% slip XG is equal to X’’d, the sub-transient reactance. The impedance in a typical case hasbeen shown to be equal to X’d, the transient reactance,at 50% slip, and to 2X’d with a slip of 0.33%. The sliplikely to be experienced with asynchronous running is

• 17 •G

ener

ator

and

Gen

erat

or-T

rans

form

er P

rote

ctio

n


+jX

-jX

+R-R

0.5

0.6

0.7

C

5.0

2.5

2.0

1.8

Load point

Loss of fieldlocus

=1ESE

EG

=1.5ESE

EG

D

Figure 17.21: Swing curves andloss of synchronism locus

EG

XGG

ZS

ZRZ

XTXTX

XGXG

-jX

A

C

θ

ZS

XGXG+ T+T ZS

+R-R

ES

XTXTX

+jX

A

D

EG

ES1

Figure 17.20: Basic interconnected system

low, perhaps 1%, so that for the purpose of assessing thepower swing locus it is sufficient to take the valueXG=2X’d.

This consideration has assumed a single value for XG.However, the reactance Xq on the quadrature axis differsfrom the direct-axis value, the ratio of Xd/Xg beingknown as the saliency factor. This factor varies with theslip speed. The effect of this factor during asynchronousoperation is to cause XG to vary at slip speed. Inconsequence, the loss of excitation impedance locusdoes not settle at a single point, but it continues todescribe a small orbit about a mean point.

A protection scheme for loss of excitation must operatedecisively for this condition, but its characteristic mustnot inhibit stable operation of the generator. One limitof operation corresponds to the maximum practicablerotor angle, taken to be at 120°. The locus of operationcan be represented as a circle on the impedance plane,as shown in Figure 17.22, stable operation conditionslying outside the circle.

On the same diagram the full load impedance locus forone per unit power can be drawn. Part of this circlerepresents a condition that is not feasible, but the pointof intersection with the maximum rotor angle curve canbe taken as a limiting operating condition for settingimpedance-based loss of excitation protection.

17.16.2 Impedance-Based Protection Characteristics

Figure 17.21 alludes to the possibility that a protection

scheme for loss of excitation could be based onimpedance measurement. The impedance characteristicmust be appropriately set or shaped to ensure decisiveoperation for loss of excitation whilst permitting stablegenerator operation within allowable limits. One or twooffset mho under impedance elements (see Chapter 11for the principles of operation) are ideally suited forproviding loss of excitation protection as long as agenerator operating at low power output (20-30%Pn)does not settle down to operate as an inductiongenerator. The characteristics of a typical two-stage lossof excitation protection scheme are illustrated in Figure17.23. The first stage, consisting of settings Xa1 and Xb1can be applied to provide detection of loss of excitationeven where a generator initially operating at low poweroutput (20-30%Pn) might settle down to operate as aninduction generator.

Pick-up and drop-off time delays td1 and tdo1 areassociated with this impedance element. Timer td1 isused to prevent operation during stable power swingsthat may cause the impedance locus of the generator totransiently enter the locus of operation set by Xb1.However, the value must short enough to preventdamage as a result of loss of excitation occurring. Ifpole-slipping protection is not required (see Section17.17.2), timer tdo1 can be set to give instantaneousreset. The second field failure element, comprisingsettings Xa2, Xb2, and associated timers td2 and tdo2 canbe used to give instantaneous tripping following loss ofexcitation under full load conditions.

17.16.3 Protection Settings

The typical setting values for the two elements varyaccording to the excitation system and operating regimeof the generator concerned, since these affect thegenerator impedance seen by the relay under normal andabnormal conditions. For a generator that is never

• 17 •

Gen

erat

or a

nd G

ener

ator

-Tra

nsfo

rmer

Pro

tect

ion


Normal machine operating impedance

R

Xb1Xb1XXb2X

-Xa1X-Xa2X

X

Alarmangle

Figure 17.23: Loss of excitation protectioncharacteristics

+R-R

+jX

ZS

XTXTX

XdX

2X'd

'd

-jX

Limitinggenerationpoint

Relay

Diameter = d/d/d 2/2/

Locus of constant MVA

Locus of constant load angle

Figure 17.22: Locus of limiting operatingconditions of synchronous machine

operated at leading power factor, or at load angles inexcess of 90° the typical settings are:

impedance element diameter Xb1 = Xd

impedance element offset Xa1 = -0.5X’d

time delay on pick-up, td1 = 0.5s – 10s

time delay on drop-off, tdo1 = 0s

If a fast excitation system is employed, allowing loadangles of up to 120° to be used, the impedance diametermust be reduced to take account of the reducedgenerator impedance seen under such conditions. Theoffset also needs revising. In these circumstances,typical settings would be:

impedance element diameter Xb1 = 0.5Xd

impedance element offset Xa1 = -0.75X’d

time delay on pick-up, td1 = 0.5s – 10s

time delay on drop-off, tdo1 = 0s

The typical impedance settings for the second element, ifused, are:

impedance element diameter

Xb2 =

Xa2 = -0.5X’d

The time delay settings td2 and tdo2 are set to zero to giveinstantaneous operation and reset.

17.17 POLE SLIPPING PROTECTION

A generator may pole-slip, or fall out of synchronismwith the power system for a number of reasons. Theprincipal causes are prolonged clearance of a heavy faulton the power system, when the generator is operating ata high load angle close to the stability limit, or partial orcomplete loss of excitation. Weak transmission linksbetween the generator and the bulk of the power systemaggravate the situation. It can also occur withembedded generators running in parallel with a strongUtility network if the time for a fault clearance on theUtility network slow, perhaps because only IDMT relaysare provided. Pole slipping is characterised by large andrapid oscillations in active and reactive power. Rapiddisconnection of the generator from the network isrequired to ensure that damage to the generator isavoided and that loads supplied by the network are notaffected for very long.

Protection can be provided using several methods. Thechoice of method will depend on the probability of poleslipping occurring and on the consequences should itoccur.

kVMVA

2

17.17.1 Protection using Reverse Power Element

During pole-slipping, there will be periods where thedirection of active power flow will be in the reversedirection, so a reverse power relay element can be usedto detect this, if not used for other purposes. However,since the reverse power conditions are cyclical, theelement will reset during the forward power part of thecycle unless either a very short pick-up time delay and/ora suitable drop-off time delay is used to eliminateresetting.

The main advantage of this method is that a reversepower element is often already present, so no additionalrelay elements are required. The main disadvantages arethe time taken for tripping and the inability to controlthe system angle at which the generator breaker tripcommand would be issued, if it is a requirement to limitthe breaker current interruption duty. There is also thedifficulty of determining suitable settings.Determination of settings in the field, from a deliberatepole-slipping test is not possible and analytical studiesmay not discover all conditions under which pole-slipping will occur.

17.17.2 Protection using an Under ImpedanceElement

With reference to Figure 17.21, a loss of excitation underimpedance characteristic may also be capable ofdetecting loss of synchronism, in applications where theelectrical centre of the power system and the generatorlies ‘behind’ the relaying point. This would typically bethe case for a relatively small generator that isconnected to a power transmission system (XG >> (XT +XS)). With reference to Figure 17.23; if pole-slippingprotection response is required, the drop-off timer tdo1 ofthe larger diameter impedance measuring elementshould be set to prevent its reset of in each slip cycle,until the td1 trip time delay has expired.

As with reverse power protection, this would be anelementary form of pole-slipping protection. It may notbe suitable for large machines where rapid tripping isrequired during the first slip cycle and where somecontrol is required for the system angle at which thegenerator circuit breaker trip command is given. Whereprotection against pole-slipping must be guaranteed, amore sophisticated method of protection should be used.A typical reset timer delay for pole-slipping protectionmight be 0.6s. For generator transformer units, theadditional impedance in front of the relaying point maytake the system impedance outside the under impedancerelay characteristic required for loss of excitationprotection. Therefore, the acceptability of this pole-slipping protection scheme will be dependent on theapplication.

• 17 •G

ener

ator

and

Gen

erat

or-T

rans

form

er P

rote

ctio

n


17.17.3 Dedicated Pole-Slipping Protection

Large generator-transformer units directly connected togrid systems often require a dedicated pole-slippingprotection scheme to ensure rapid tripping and withsystem angle control. Historically, dedicated protectionschemes have usually been based on an ohm-typeimpedance measurement characteristic.

17.17.3.1 Pole slipping protection by impedancemeasurement

Although a mho type element for detecting the changein impedance during pole-slipping can be used in someapplications, but with performance limits, a straight lineohm characteristic is more suitable. The protectionprinciple is that of detecting the passage of thegenerator impedance through a zone defined by twosuch impedance characteristics, as shown in Figure17.24. The characteristic is divided into three zones, A,B, and C. Normal operation of the generator lies in zoneA. When a pole-slip occurs, the impedance traverseszones B and C, and tripping occurs when the impedancecharacteristic enters zone C.

Tripping only occurs if all zones are traversedsequentially. Power system faults should result in thezones not being fully traversed so that tripping will notbe initiated. The security of this type of protectionscheme is normally enhanced by the addition of a plainunder impedance control element (circle about the originof the impedance diagram) that is set to prevent trippingfor impedance trajectories for remote power systemfaults. Setting of the ohm elements is such that they lieparallel to the total system impedance vector, andenclose it, as shown in Figure 17.24.

17.17.3.2 Use of lenticular characteristic

A more sophisticated approach is to measure theimpedance of the generator and use a lenticularimpedance characteristic to determine if a pole-slippingcondition exists. The lenticular characteristic is shown inFigure 17.25. The characteristic is divided into two halvesby a straight line, called the blinder.

The inclination, θ, of the lens and blinder is determined bythe angle of the total system impedance. The impedanceof the system and generator-transformer determines theforward reach of the lens, ZA, and the transient reactanceof the generator determines the reverse reach ZB.

The width of the lens is set by the angle α and the linePP’, perpendicular to the axis of the lens, is used todetermine if the centre of the impedance swing during atransient is located in the generator or power system.

Operation in the case of a generator is as follows. Thecharacteristic is divided into 4 zones and 2 regions, asshown in Figure 17.26.

Normal operation is with the measured impedance inzone R1. If a pole slip develops, the impedance locus willtraverse though zones R2, R3, and R4. When enteringzone R4, a trip signal is issued, provided the impedancelies below reactance line PP’ and hence the locus ofswing lies within or close to the generator – i.e. thegenerator is pole slipping with respect to the rest of thesystem.

• 17 •

Gen

erat

or a

nd G

ener

ator

-Tra

nsfo

rmer

Pro

tect

ion


R

ZBZ

ZAZ

Blinder

X

Lens

P

P'

θα

Figure 17.25: Pole-slipping protection usinglenticular characteristic and blinder

+R

ZS

TT

XGXGX

-R

+jX

-jXOhm relay 2

Relaying pointing

BC A

Ohm relay 1

Slip locusEG=ES

Figure 17.24: Pole slipping detectionby ohm relays

If the impedance locus lies above line PP’, the swing liesfar out in the power system – i.e. one part of the powersystem, including the protected generator, is swingingagainst the rest of the network. Tripping may still occur,but only if swinging is prolonged – meaning that thepower system is in danger of complete break-up. Furtherconfidence checks are introduced by requiring that theimpedance locus spends a minimum time within eachzone for the pole-slipping condition to be valid. The tripsignal may also be delayed for a number of slip cycleseven if a generator pole-slip occurs – this is to bothprovide confirmation of a pole-slipping condition andallow time for other relays to operate if the cause of thepole slip lies somewhere in the power system. Shouldthe impedance locus traverse the zones in any othersequence, tripping is blocked.

17.18 STATOR OVERHEATING

Overheating of the stator may result from:

i. overload

ii. failure of the cooling system

iii. overfluxing

iv. core faults

Accidental overloading might occur through thecombination of full active load current component,governed by the prime mover output and an abnormallyhigh reactive current component, governed by the levelof rotor excitation and/or step-up transformer tap. Witha modern protection relay, it is relatively simple toprovide a current-operated thermal replica protectionelement to estimate the thermal state of the stator

windings and to issue an alarm or trip to preventdamage.

Although current-operated thermal replica protectioncannot take into account the effects of ambienttemperature or uneven heat distribution, it is oftenapplied as a back-up to direct stator temperaturemeasuring devices to prevent overheating due to highstator current. With some relays, the thermal replicatemperature estimate can be made more accuratethrough the integration of direct measuring resistancetemperature devices.

Irrespective of whether current-operated thermal replicaprotection is applied or not, it is a requirement tomonitor the stator temperature of a large generator inorder to detect overheating from whatever cause.

Temperature sensitive elements, usually of the resistancetype, are embedded in the stator winding at hot-spotlocations envisaged by the manufacturer, the numberused being sufficient to cover all variations. Theelements are connected to a temperature sensing relayelement arranged to provide alarm and trip outputs. Thesettings will depend on the type of stator windinginsulation and on its permitted temperature rise.

17.19 MECHANICAL FAULTS

Various faults may occur on the mechanical side of agenerating set. The following sections detail the moreimportant ones from an electrical point of view.

17.19.1 Failure of the Prime Mover

When a generator operating in parallel with others losesits power input, it remains in synchronism with thesystem and continues to run as a synchronous motor,drawing sufficient power to drive the prime mover. Thiscondition may not appear to be dangerous and in somecircumstances will not be so. However, there is a dangerof further damage being caused. Table 17.1 lists sometypical problems that may occur.

Protection is provided by a low forward power/reversepower relay, as detailed in Section 17.11

17.19.2 Overspeed

The speed of a turbo-generator set rises when the steaminput is in excess of that required to drive the load atnominal frequency. The speed governor can normallycontrol the speed, and, in any case, a set running inparallel with others in an interconnected system cannotaccelerate much independently even if synchronism islost. However, if load is suddenly lost when the HVcircuit breaker is tripped, the set will begin to accelerate

• 17 •G

ener

ator

and

Gen

erat

or-T

rans

form

er P

rote

ctio

n


Blinder

Z

1

ZSO

ZBZ

X

T2T

XTXTXM

Stable Power Swing

a

A

X

Left-lens Right-lensB

Power Swing In System

S

Pole SlippingCharacteristic

R4 R3 R2 R1

P

P'

Figure 17.26: Definition of zones for lenticularcharacteristic

rapidly. The speed governor is designed to prevent adangerous speed rise even with a 100% load rejection,but nevertheless an additional centrifugal overspeed tripdevice is provided to initiate an emergency mechanicalshutdown if the overspeed exceeds 10%.

To minimise overspeed on load rejection and hence themechanical stresses on the rotor, the following sequenceis used whenever electrical tripping is not urgentlyrequired:

i. trip prime mover or gradually reduce power input tozero

ii. allow generated power to decay towards zero

iii. trip generator circuit breaker only when generatedpower is close to zero or when the power flowstarts to reverse, to drive the idle turbine

17.19.3 Loss of Vacuum

A failure of the condenser vacuum in a steam turbinedriven generator results in heating of the tubes. Thisthen produces strain in the tubes, and a rise intemperature of the low-pressure end of the turbine.Vacuum pressure devices initiate progressive unloadingof the set and, if eventually necessary, tripping of theturbine valves followed by the high voltage circuitbreaker. The set must not be allowed to motor in the

event of loss of vacuum, as this would cause rapidoverheating of the low-pressure turbine blades.

17.20 COMPLETE GENERATOR PROTECTIONSCHEMES

From the preceding sections, it is obvious that theprotection scheme for a generator has to take account ofmany possible faults and plant design variations.Determination of the types of protection used for aparticular generator will depend on the nature of theplant and upon economic considerations, which in turnis affected by set size. Fortunately, modern, multi-function, numerical relays are sufficiently versatile toinclude all of the commonly required protectionfunctions in a single package, thus simplifying thedecisions to be made. The following sections provideillustrations of typical protection schemes for generatorsconnected to a grid network, but not all possibilities areillustrated, due to the wide variation in generator sizesand types.

17.20.1 Direct-Connected Generator

A typical protection scheme for a direct-connectedgenerator is shown in Figure 17.27. It comprises thefollowing protection functions:

• 17 •

Gen

erat

or a

nd G

ener

ator

-Tra

nsfo

rmer

Pro

tect

ion


Excitationcircuitbreaker

Governortrip

Generatorcircuitbreaker

Stator winding temperature

Loss of excitation

Stator E/F (or neutral voltage

Reverse/low forward power

Back-up overcurrent (or voltage

Stator differential (biased/high

Low powerinterlock

Emergency push button

Lubricating oil failure

Electrical trip of governor

Under/overvoltage

Unbalanced loading

impedance)

displacement)

dependent O/C)

Underfrequency

N.B. Alarms and time delays omitted for simplicity

Mechanical faults (non-urgent)

Mechanical faults (urgent)

Pole slipping

OverfluxingInadvertent energisation

Figure 17.27: Typical protection arrangement for a direct-connected generator

1. stator differential protection

2. overcurrent protection – conventional or voltagedependent

3. stator earth fault protection

4. overvoltage protection

5. undervoltage protection

6. overload/low forward power/ reverse powerprotection (according to prime mover type)

7. unbalanced loading

8. overheating

9. pole slipping

10. loss of excitation

11. underfrequency

12. inadvertent energisation

13. overfluxing

14. mechanical faults

Figure 17.27 illustrates which trips require an

instantaneous electrical trip and which can be timedelayed until electrical power has been reduced to a lowvalue. The faults that require tripping of the prime moveras well as the generator circuit breaker are also shown.

17.20.2 Generator-Transformer Units

These units are generally of higher output than direct-connected generators, and hence more comprehensiveprotection is warranted. In addition, the generatortransformer also requires protection, for which theprotection detailed in Chapter 16 is appropriate

Overall biased generator/generator transformerdifferential protection is commonly applied in addition,or instead of, differential protection for the transformeralone. A single protection relay may incorporate all ofthe required functions, or the protection of thetransformer (including overall generator/generatortransformer differential protection) may utilise aseparate relay.

Figure 17.28 shows a typical overall scheme.

• 17 •G

ener

ator

and

Gen

erat

or-T

rans

form

er P

rote

ctio

n


Generatorcircuitbreaker

Excitationcircuitbreaker

Stator winding temperature

Loss of excitation

Stator E/F (or neutral voltage

Reverse/low forward power

Stator differential (biased/high

Low powerinterlock

Emergency push button

Lubricating oil failure

Governortrip

Electrical trip of governor

Under/overvoltage

Unbalanced loading

Pole slipping

Transformer winding temperature

Overall differential (transformerdifferential)

Buchholz

HV overcurrent

HV restricted E/F

impedance)

displacement)Back-up overcurrent (or voltage

dependent O/C)

Underfrequency

N.B. Alarms and time delays omitted for simplicity

Mechanical faults (non-urgent)

Mechanical faults (urgent)

Overfluxing

Inadvertent energisation

Figure 17.28: Typical tripping arrangements for generator-transformer unit

17.21 EMBEDDED GENERATION

In recent years, through de-regulation of the electricitysupply industry and the ensuing commercialcompetition, many electricity users connected to MVpower distribution systems have installed generating setsto operate in parallel with the public supply. Theintention is either to utilise surplus energy from othersources, or to use waste heat or steam from the primemover for other purposes. Parallel connection ofgenerators to distribution systems did occur before de-regulation, but only where there was a net power importfrom the Utility. Power export to Utility distributionsystems was a relatively new aspect. Since generation ofthis type can now be located within a Utility distributionsystem, as opposed to being centrally dispatchedgeneration connected to a transmission system, the term‘Embedded Generation’ is often applied. Figure 17.2illustrates such an arrangement. Depending on size, theembedded generator(s) may be synchronous orasynchronous types, and they may be connected at anyvoltage appropriate to the size of plant being considered.

The impact of connecting generation to a Utilitydistribution system that was originally engineered onlyfor downward power distribution must be considered,particularly in the area of protection requirements. Inthis respect, it is not important whether the embeddedgenerator is normally capable of export to the Utilitydistribution system or not, since there may exist faultconditions when this occurs irrespective of the designintent.

If plant operation when disconnected from the Utilitysupply is required, underfrequency protection (Section17.4.2) will become an important feature of the in-plantpower system. During isolated operation, it may berelatively easy to overload the available generation, suchthat some form of load management system may berequired. Similarly, when running in parallel with theUtility, consideration needs to be given to the mode ofgenerator operation if reactive power import is to becontrolled. The impact on the control scheme of asudden break in the Utility connection to the plant mainbusbar also requires analysis. Where the in-plantgeneration is run using constant power factor orconstant reactive power control, automatic reversion tovoltage control when the Utility connection is lost isessential to prevent plant loads being subjected to avoltage outside acceptable limits.

Limits may be placed by the Utility on the amount ofpower/reactive power import/export. These may demandthe use of an in-plant Power Management System tocontrol the embedded generation and plant loadsaccordingly. Some Utilities may insist on automatictripping of the interconnecting circuit breakers if there isa significant departure outside permissible levels of

frequency and voltage, or for other reasons.

From a Utility standpoint, the connection of embeddedgeneration may cause problems with voltage control andincreased fault levels. The settings for protection relaysin the vicinity of the plant may require adjustment withthe emergence of embedded generation. It must also beensured that the safety, security and quality of supply ofthe Utility distribution system is not compromised. Theembedded generation must not be permitted to supplyany Utility customers in isolation, since the Utility supplyis normally the means of regulating the system voltageand frequency within the permitted limits. It alsonormally provides the only system earth connection(s), toensure the correct performance of system protection inresponse to earth faults. If the Utility power infeed fails,it is also important to disconnect the embeddedgeneration before there is any risk of the Utility powersupply returning on to unsynchronised machines. Inpractice this generally requires the following protectionfunctions to be applied at the ‘Point of CommonCoupling’ (PCC) to trip the coupling circuit breaker:

a. overvoltage

b. undervoltage

c. overfrequency

d. underfrequency

e. loss of Utility supply

In addition, particular circumstances may requireadditional protection functions:

f. neutral voltage displacement

g. reverse power

h. directional overcurrent

In practice, it can be difficult to meet the protectionsettings or performance demanded by the Utility withouta high risk of nuisance tripping caused by lack of co-ordination with normal power system faults anddisturbances that do not necessitate tripping of theembedded generation. This is especially true whenapplying protection specifically to detect loss of theUtility supply (also called ‘loss of mains’) to cater foroperating conditions where there would be noimmediate excursion in voltage or frequency to causeoperation of conventional protection functions.

17.21.1 Protection Against Loss of Utility Supply

If the normal power infeed to a distribution system, or tothe part of it containing embedded generation is lost, theeffects may be as follows:

a. embedded generation may be overloaded, leadingto generator undervoltage/underfrequency

• 17 •

Gen

erat

or a

nd G

ener

ator

-Tra

nsfo

rmer

Pro

tect

ion


b. embedded generation may be underloaded, leadingto overvoltage/overfrequency

c. little change to the absolute levels of voltage orfrequency if there is little resulting change to theload flow through the PCC

The first two effects are covered by conventional voltageand frequency protection. However, if condition (c)occurs, conventional protection may not detect the lossof Utility supply condition or it may be too slow to do sowithin the shortest possible auto-reclose dead-timesthat may be applied in association with Utility overheadline protection. Detection of condition (c) must beachieved if the requirements of the Utility are to be met.Many possible methods have been suggested, but theone most often used is the Rate of Change of Frequency(ROCOF) relay. Its application is based on the fact thatthe rate of change of small changes in absolutefrequency, in response to inevitable small load changes,will be faster with the generation isolated than when thegeneration is in parallel with the public, interconnectedpower system. However, problems with nuisancetripping in response to national power system events,where the system is subject to significant frequencyexcursions following the loss of a large generator or amajor power interconnector, have occurred.This is particularly true for geographically islanded powersystems, such as those of the British Isles. An alternativeto ROCOF protection is a technique sometimes referredto as ‘voltage vector shift’ protection. In this techniquethe rate of phase change between the directly measuredgenerator bus voltage is compared with a memorised a.c.bus voltage reference.

Sources of embedded generation are not normallyearthed, which presents a potential safety hazard. In theevent of an Utility system earth fault, the Utilityprotection should operate to remove the Utility powerinfeed. In theory, this should also result in removal of theembedded generation, through the action of thestipulated voltage/frequency protection and dependable‘loss of mains’ protection. However, in view of safetyconsiderations (e.g. fallen overhead line conductors inpublic areas), an additional form of earth fault protectionmay also be demanded to prevent the backfeed of anearth fault by embedded generation. The only way ofdetecting an earth fault under these conditions is to useneutral voltage displacement protection. The additionalrequirement is only likely to arise for embeddedgeneration rated above 150kVA, since the risk of thesmall embedded generators not being cleared by othermeans is negligible.

17.21.2 ROCOF Relay Description

A ROCOF relay detects the rate of change of frequency inexcess of a defined setpoint. The signal is obtained froma voltage transformer connected close to the Point ofCommon Coupling (PCC). The principal method used isto measure the time period between successive zero-crossings to determine the average frequency for eachhalf-cycle and hence the rate of change of frequency.The result is usually averaged over a number of cycles.

17.21.3 Voltage Vector Shift Relay Description

A voltage vector shift relay detects the drift in voltagephase angle beyond a defined setpoint as long as it takesplace within a set period. Again, the voltage signal isobtained from a voltage transformer connected close tothe Point of Common Coupling (PCC). The principalmethod used is to measure the time period betweensuccessive zero-crossings to determine the duration ofeach half-cycle, and then to compare the durations withthe memorised average duration of earlier half-cycles inorder to determine the phase angle drift.

17.21.4 Setting Guidelines

Should loss of the Utility supply occur, it is extremelyunlikely that there will be an exact match between theoutput of the embedded generator(s) and the connectedload. A small frequency change or voltage phase anglechange will therefore occur, to which can be added anychanges due to the small natural variations in loading ofan isolated generator with time. Once the rate of changeof frequency exceeds the setting of the ROCOF relay fora set time, or once the voltage phase angle drift exceedsthe set angle, tripping occurs to open the connectionbetween the in-plant and Utility networks.

While it is possible to estimate the rate of change offrequency from knowledge of the generator set inertiaand MVA rating, this is not an accurate method forsetting a ROCOF relay because the rotational inertia ofthe complete network being fed by the embeddedgeneration is required. For example, there may be otherembedded generators to consider. As a result, it isinvariably the case that the relay settings are determinedat site during commissioning. This is to ensure that theUtility requirements are met while reducing thepossibility of a spurious trip under the various operatingscenarios envisaged. However, it is very difficult todetermine whether a given rate of change of frequencywill be due to a ‘loss of mains’ incident or aload/frequency change on the public power network, andhence spurious trips are impossible to eliminate. Thusthe provision of Loss of Utility Supply protection to meetpower distribution Utility interface protection

• 17 •G

ener

ator

and

Gen

erat

or-T

rans

form

er P

rote

ctio

n


requirements, may actually conflict with the interests ofthe national power system operator. With the growingcontribution of non-dispatched embedded generation tothe aggregate national power demand, the loss of theembedded generation following a transmission systemincident that may already challenge the security of thesystem can only aggravate the problem. There have beenclaims that voltage vector shift protection might offerbetter security, but it will have operation times that varywith the rate of change of frequency. As a result,depending on the settings used, operation times mightnot comply with Utility requirements under allcircumstances. Reference 17.1 provides further detailsof the operation of ROCOF relays and the problems thatmay be encountered.

Nevertheless, because such protection is a commonrequirement of some Utilities, the ‘loss of mains’protection may have to be provided and the possibility ofspurious trips will have to be accepted in those cases.Site measurements over a period of time of the typicalrates of frequency change occurring may assist innegotiations of the settings with the Utility, and with thefine-tuning of the protection that may already becommissioned.

17.22 EXAMPLES OF GENERATOR PROTECTIONSETTINGS

This section gives examples of the calculations requiredfor generator protection. The first is for a typical smallgenerator installed on an industrial system that runs inparallel with the Utility supply. The second is for a largergenerator-transformer unit connected to a grid system.

17.22.1 Protection Settings of aSmall Industrial Generator

Salient details of the generator, network and protectionrequired are given in Table 17.2. The examplecalculations are based on a MiCOM P343 relay in respectof setting ranges, etc.

17.22.1.1 Differential protection

Biased differential protection involves the determinationof values for four setting values: Is1, Is2, K1 and K2 inFigure 17.5. Is1 can be set at 5% of the generator rating,in accordance with the recommendations for the relay,and similarly the values of Is2 (120%) and K2 (150%) ofgenerator rating. It remains for the value of K1 to bedetermined. The recommended value is generally 0%,but this only applies where CT’s that conform to IEC60044-1 class PX (or the superseded BS 3938 Class X)are used – i.e. CT’s specifically designed for use indifferential protection schemes. In this application, theCT’s are conventional class 5P CT’s that meet the relayrequirements in respect of knee-point voltage, etc.Where neutral tail and terminal CT’s can saturate atdifferent times due to transiently offset magnetisinginrush or motor starting current waveforms with an r.m.s.level close to rated current and where there is a high L/Rtime constant for the offset, the use of a 0% bias slopemay give rise to maloperation. Such waveforms can beencountered when plant of similar rating to thegenerator is being energised or started. Differencesbetween CT designs or differing remanent flux levels canlead to asymmetric saturation and the production of adifferential spill current. Therefore, it is appropriate toselect a non-zero setting for K1, and a value of 5% isusual in these circumstances.

17.22.1.2 Voltage controlled overcurrent protection

This protection is applied as remote backup to thedownstream overcurrent protection in the event ofprotection or breaker failure conditions. This ensuresthat the generator will not continue to supply the faultunder these conditions.

At normal voltage, the current setting must be greaterthan the maximum generator load current of 328A. Amargin must be allowed for resetting of the relay at thiscurrent (reset ratio = 95%) and for the measurementtolerances of the relay (5% of Is under referenceconditions), therefore the current setting is calculated as:

The nearest settable value is 365A, or 0.73In.

The minimum phase-phase voltage for a close-up single-phase to earth fault is 57%, so the voltage setting Vsmust be less than this. A value of 30% is typically used,giving Vs = 33V. The current setting multiplying factor

I

A

vcset > ×

>

3280 95

1 05

362 5

..

.

• 17 •

Gen

erat

or a

nd G

ener

ator

-Tra

nsfo

rmer

Pro

tect

ion


Table 17.2: Data for small generator protection example

Generator Data

kVA kW PF Rated Rated Rated Rated Prime Mover voltage current frequency speed type

6250 5000 0.8 11000 328 50 1500 Steam Turbine

Generator Parameters

Generator type Xd p.u. X’d p.u. CT Ratio VT Ratio

Salient Pole 2.349 0.297 500/1 11000/110

Network Data

Earthing Maximum earth Minimum phase Maximum downstreamresistor fault current fault current phase fault current

31.7Ω 200A 145A 850A

Existing Protection

CT Ratio Overcurrent Settings Earth Fault Settings

Characteristic Setting TMS Characteristic Setting TMS

200/1 SI 144A 0.176 SI 48A 0.15

K must be chosen such that KIS is less than 50% of thegenerator steady-state current contribution to anuncleared remote fault. This information is not available(missing data being common in protection studies).However, the maximum sustained close-up phase faultcurrent (neglecting AVR action) is 145A, so that a settingchosen to be significantly below this value will suffice. Avalue of 87.5A (60% of the close-up sustained phasefault current) is therefore chosen, and hence K = 0.6.This is considered to be appropriate based on knowledgeof the system circuit impedances. The TMS setting ischosen to co-ordinate with the downstream feederprotection such that:

1. for a close-up feeder three-phase fault, that resultsin almost total voltage collapse as seen by the relay

2. for a fault at the next downstream relay location, ifthe relay voltage is less than the switching voltage

It should also be chosen so that the generator cannot besubjected to fault or overload current in excess of thestator short-time current limits. A curve should beprovided by the manufacturer, but IEC 60034-1 demandsthat an AC generator should be able to pass 1.5 timesrated current for at least 30 seconds. The operating timeof the downstream protection for a three-phase faultcurrent of 850A is 0.682s, so the voltage controlled relayelement should have a minimum operating time of 1.09s(0.4s grading margin used as the relay technology usedfor the downstream relay is not stated – see Table 9.2).With a current setting of 87.5A, the operating time ofthe voltage controlled relay element at a TMS of 1.0 is:

Therefore a TMS of:

is required. Use 0.375, nearest available setting.

17.22.1.3 Stator earth fault protection

The maximum earth fault current, from Table 17.2, is200A. Protection for 95% of the winding can beprovided if the relay is set to detect a primary earth faultcurrent of 16.4A, and this equates to a CT secondarycurrent of 0.033A. The nearest relay setting is 0.04A,providing protection for 90% of the winding.

The protection must grade with the downstream earthfault protection, the settings of which are also given inTable 17.2. At an earth fault current of 200A, thedownstream protection has an operation time of 0.73s.The generator earth fault protection must therefore have

1 093 01

0 362..

.=

0 14

85087 5

1

3 010 02

.

.

. s.

−=

an operation time of not less than 1.13s. At a TMS of 1.0,the generator protection relay operating time will be:

=2.97s, so the required TMS is .

Use a setting of 0.4, nearest available setting.

17.22.1.4 Neutral voltage displacement protectionThis protection is provided as back-up earth-faultprotection for the generator and downstream system(direct-connected generator). It must therefore have asetting that grades with the downstream protection. Theprotection is driven from the generator star-connectedVT, while the downstream protection is current operated.

It is therefore necessary to translate the current settingof the downstream setting of the current-operatedearth-fault protection into the equivalent voltage for theNVD protection. The equivalent voltage is found fromthe formula:

where:

Veff = effective voltage setting

Ipe = downstream earth-fault current setting

Ze = earthing resistance

Hence a setting of 48V is acceptable. Time grading isrequired, with a minimum operating time of the NVDprotection of 1.13s at an earth fault current of 200A.Using the expression for the operation time of the NVDelement:

t = K/(M-1) sec

where:

and

V = voltage seen by relay

Vsnvd = relay setting voltage

the value of K can be calculated as 3.34. The nearestsettable value is 3.5, giving an operation time of 1.18s.

M VVsnvd

=

VI Z

VT ratio

V

effpe e=

×( )×

= × ×

=

3

48 31 7 3100

45 6

.

.

1 132 97

0 38..

.=

0 14

20020 1

0 02

. s.( ) −

• 17 •G

ener

ator

and

Gen

erat

or-T

rans

form

er P

rote

ctio

n


17.22.1.5 Loss of excitation protection

Loss of excitation is detected by a mho impedance relayelement, as detailed in Section 17.16.2. The standardsettings for the P340 series relay are:

Xa = 0.5X’d x (CT ratio/VT ratio)

(in secondary quantities)

= -0.5 x 0.297 x 19.36 x 500/100

= -14.5ΩXb = Xd x (CT ratio/VT ratio)

= 2.349Ω x 19.36 x (500/100)

= 227ΩThe nearest settings provided by the relay are Xa = -14.5Ω Xb = 227Ω. The time delay td1 should be set toavoid relay element operation on power swings and atypical setting of 3s is used. This value may need to bemodified in the light of operating experience. To preventcyclical pick-up of the relay element without tripping,such as might occur during pole-slipping conditions, adrop-off time delay tdo1 is provided and set to 0.5s.

17.22.1.6 Negative phase sequence current protection

This protection is required to guard against excessiveheating from negative phase sequence currents, whateverthe cause. The generator is of salient pole design, so fromIEC 60034-1, the continuous withstand is 8% of ratingand the I2

2t value is 20s. Using Equation 17.1, therequired relay settings can found as I2>> = 0.05 and K =8.6s. The nearest available values are I2>> = 0.05 andK = 8.6s. The relay also has a cooling time constantKreset that is normally set equal to the value of K. To co-ordinate with clearance of heavy asymmetric systemfaults, that might otherwise cause unnecessary operationof this protection, a minimum operation time tmin shouldbe applied. It is recommended to set this to a value of 1.Similarly, a maximum time can be applied to ensure thatthe thermal rating of the generator is not exceeded (asthis is uncertain, data not available) and to take accountof the fact that the P343 characteristic is not identicalwith that specified in IEC 60034. The recommendedsetting for tmax is 600s.

17.22.1.7 Overvoltage protection

This is required to guard against various failure modes,e.g. AVR failure, resulting in excessive stator voltage. Atwo-stage protection is available, the first being a low-set time-delayed stage that should be set to grade withtransient overvoltages that can be tolerated followingload rejection. The second is a high-set stage used forinstantaneous tripping in the event of an intolerableovervoltage condition arising.

Generators can normally withstand 105% of ratedvoltage continuously, so the low-set stage should be sethigher than this value. A setting of 117.7V in secondary

quantities (corresponding to 107% of rated statorvoltage) is typically used, with a definite time delay of10s to allow for transients due to load switch-off/rejection, overvoltages on recovery from faults ormotor starting, etc.

The second element provides protection in the event of alarge overvoltage, by tripping excitation and thegenerator circuit breaker (if closed). This must be setbelow the maximum stator voltage possible, taking intoaccount saturation. As the open circuit characteristic ofthe generator is not available, typical values must beused. Saturation will normally limit the maximumovervoltage on this type of generator to 130%, so asetting of 120% (132V secondary) is typically used.Instantaneous operation is required. Generatormanufacturers are normally able to providerecommendations for the relay settings. For embeddedgenerators, the requirements of the local Utility may alsohave to be taken into account. For both elements, avariety of voltage measurement modes are available totake account of possible VT connections (single or three-phase, etc.), and conditions to be protected against. Inthis example, a three-phase VT connection is used, andovervoltages on any phase are to be detected, so aselection of ‘Any’ is used for this setting.

17.22.1.8 Underfrequency protection

This is required to protect the generator from sustainedoverload conditions during periods of operation isolatedfrom the Utility supply. The generating set manufacturerwill normally provide the details of machine short-timecapabilities. The example relay provides four stages ofunderfrequency protection. In this case, the first stage isused for alarm purposes and a second stage would beapplied to trip the set.

The alarm stage might typically be set to 49Hz, with atime delay of 20s, to avoid an alarm being raised undertransient conditions, e.g. during plant motor starting.The trip stage might be set to 48Hz, with a time delay of0.5s, to avoid tripping for transient, but recoverable, dipsin frequency below this value.

17.22.1.9 Reverse power protection

The relay setting is 5% of rated power.

This value can be set in the relay. A time delay isrequired to guard against power swings while generatingat low power levels, so use a time delay of 5s. No resettime delay is required.

settingCT ratio VT ratio

W

= × ××

= × ××

=

0 05 5 10

0 05 5 10500 100

5

6

6

.

.

• 17 •

Gen

erat

or a

nd G

ener

ator

-Tra

nsfo

rmer

Pro

tect

ion


17.22.2 Large Generator Transformer Unit Protection

The data for this unit are given in Table 17.4. It is fittedwith two main protection systems to ensure security oftripping in the event of a fault. To economise on space,the setting calculations for only one system, that using aMiCOM P343 relay are given. Settings are given inprimary quantities throughout.

17.22.2.1 Biased differential protection

The settings follow the guidelines previously stated. As100% stator winding earth-fault protection is provided,high sensitivity is not required and hence Is1 can be setto 10% of generator rated current. This equates to 602A,and the nearest settable value on the relay is 640A (=0.08 of rated CT current). The settings for K1, Is2 and K2follow the guidelines in the relay manual.

17.22.2.2 Voltage restrained overcurrent protection

The setting current Iset has to be greater than the full-load current of the generator (6019A). A suitable marginmust be allowed for operation at reduced voltage, so usea multiplying factor of 1.2. The nearest settable value is7200A. The factor K is calculated so that the operatingcurrent is less than the current for a remote end threephase fault. The steady-state current and voltage at thegenerator for a remote-end three-phase fault are givenby the expressions:

where:

If = minimum generator primary current for amulti-phase feeder-end fault

VN = no-load phase-neutral generator voltage

Xd = generator d-axis synchronous reactance

Xt = generator transformer reactance

rf = feeder resistance

Xf = feeder reactance

n = number of parallel generators

IV

nR X X nX

where

I imum generator primary

current for a multi phase

feeder end fault

V no load phase neutra

enerator voltage

X generator d axis synchronous

react ce

X generator transformer reac ce

r feeder resis ce

X feeder reac

fltN

f d t f

f

N

d

t

f

f

=+ + +

=

−

−

= − −

= −

=

=

=

( ) ( )

:

min

l

g

an

tan

tan

tan

2 2

cece

n number of paralle enerators= l g

• 17 •G

ener

ator

and

Gen

erat

or-T

rans

form

er P

rote

ctio

n


Table 17.3: Small generator protection example – relay settings

Is1 5%

Is2 120%

K1 5%

K2 150%

Ise 0.04

TMS 0.4

Vsnvd 48V

K 3.5

Xa -14.5ΩXb 227Ωtd1 3s

tDO1 0.5s

Ivcset 0.73

Vs 33

K 0.6

TMS 0.375

I2>> 0.05

K 8.6s

Kreset 8.6s

tmin 1.5s

tmax 600s

V> meas mode three-phase

V> operate mode any

V>1 setting 107%

V>1 function DT

V>1 time delay 10s

V>2 setting 120%

V>2 function DT

V>2 time delay 0sec

F<1 setting 49Hz

F<1 time delay 20s

F<2 setting 48Hz

F<2 time delay 0.5s

P1 function reverse power

P1 setting 5W

P1 time delay 5s

P1 DO time 0s

Protection Quantity Value

Differential protection

Stator earth fault

Neutral voltage displacement

Loss of excitation

Voltage controlled overcurrent

Negative phase sequence

Overvoltage

Underfrequency

Reverse Power

Parameter Value Unit

Generator MVA rating 187.65 MVA

Generator MW rating 160 MW

Generator voltage 18 kV

Synchronous reactance 1.93 pu

Direct-axis transient reactance 0.189 pu

Minimum operating voltage 0.8 pu

Generator negative sequence capability 0.08 pu

Generator negative sequence factor, Kg 10

Generator third harmonic voltage under load 0.02 pu

Generator motoring power 0.02 pu

alarm 1.1 pu

Generator overvoltage time delay 5 s

trip 1.3 pu

Generator undervoltage not required

Max pole slipping frequency 10 Hz

Generator transformer rating 360 MVA

Generator transformer leakage reactance 0.244 pu

Generator transformer overflux alarm 1.1 pu

Generator transformer overflux alarm 1.2 pu

Network resistance (referred to 18kV) 0.56 mΩNetwork reactance (referred to 18kV) 0.0199 ΩSystem impedance angle (estimated) 80 deg

Minimum load resistance 0.8 ΩGenerator CT ratio 8000/1

Generator VT ratio 18000/120

Number of generators in parallel 2

Table 17.4: System data for large generator protection example

hence,

Iflt = 2893A

= 0.361N

and

A suitable value of K is therefore .

A suitable value of V2set is 120% of Vflt, giving a valueof 1565V. The nearest settable value is 3000V, minimumallowable relay setting. The value of V1set is required tobe above the minimum voltage seen by the generator fora close-up phase-earth fault. A value of 80% of ratedvoltage is used for V1set, 14400V.

17.22.2.3 Inadvertent energisation protection

This protection is a combination of overcurrent withundervoltage, the voltage signal being obtained from aVT on the generator side of the system. The currentsetting used is that of rated generator current of 6019A,in accordance with IEEE C37.102 as the generator is forinstallation in the USA. Use 6000A nearest settablevalue. The voltage setting cannot be more than 85% ofthe generator rated voltage to ensure operation does notoccur under normal operation. For this application, avalue of 50% of rated voltage is chosen.

17.22.2.4 Negative phase sequence protection

The generator has a maximum steady-state capability of8% of rating, and a value of Kg of 10. Settings of I2cmr= 0.06 (=480A) and Kg = 10 are therefore used.Minimum and maximum time delays of 1s and 1300s areused to co-ordinate with external protection and ensuretripping at low levels of negative sequence current areused.

17.22.2.5 Overfluxing protection

The generator-transformer manufacturer supplied thefollowing characteristics:

Alarm:

Trip: time characteristic

Hence the alarm setting is .

A time delay of 5s is used to avoid alarms due totransient conditions.

The trip setting is .18000 1 260 360× =. V Hz

18000 1 0560 315× =. V Hz

Vf inverse

time characteristic

>1 2. ,

Vf >1 1.

0 3611 2 0 3.. .=

VV nR X nX

nR X X nX

V

U

fltN f t f

f d t f

N

=+ +

+ + +

=

=

3

1304

0 07

2 2

2 2

(( ) ( ) )

( ) ( )

.

A TMS value of 10 is selected, to match the withstandcurve supplied by the manufacturer.

17.22.2.6 100% Stator earth fault protection

This is provided by a combination of neutral voltagedisplacement and third harmonic undervoltageprotection. For the neutral voltage displacementprotection to cover 90% of the stator winding, theminimum voltage allowing for generator operation at aminimum of 92% of rated voltage is:

Use a value of 935.3V, nearest settable value thatensures 90% of the winding is covered. A 0.5s definitetime delay is used to prevent spurious trips. The thirdharmonic voltage under normal conditions is 2% of ratedvoltage, giving a value of:

The setting of the third harmonic undervoltageprotection must be below this value, a factor of 80%being acceptable. Use a value of 166.3V. A time delayof 0.5s is used. Inhibition of the element at lowgenerator output requires determination duringcommissioning.

17.22.2.7 Loss of excitation protection

The client requires a two-stage loss of excitationprotection function. The first is alarm only, while thesecond provides tripping under high load conditions. Toachieve this, the first impedance element of the P343loss of excitation protection can be set in accordancewith the guidelines of Section 17.16.3 for a generatoroperating at rotor angles up to 120o, as follows:

Xb1 = 0.5Xd = 1.666Ω

Xa1 = 0.75X’d = 0.245Ω

Use nearest settable values of 1.669Ω and 0.253Ω. Atime delay of 5s is used to prevent alarms undertransient conditions. For the trip stage, settings for highload as given in Section 17.16.3 are used:

The nearest settable value for Xb2 is 1.725Ω. A timedelay of 0.5s is used.

X kVMVA

X X

b

a d

2

2 2

2

18187 65

1 727

0 75 0 1406

= = =

=− ′ =−.

.

. .

Ω

Ω

18 0 023

207 8

kV

V

×

=

.

.

0 92 18 0 13

956 1

. .

.

× ×

=

kV

V

• 17 •

Gen

erat

or a

nd G

ener

ator

-Tra

nsfo

rmer

Pro

tect

ion


17.22.2.8 Reverse power protection

The manufacturer-supplied value for motoring power is2% of rated power. The recommended setting istherefore 1.6MW. An instrumentation class CT is used inconjunction with the relay for this protection, to ensureaccuracy of measurement. A time delay of 0.5s is used.The settings should be checked at the commissioningstage.

17.22.2.9 Over/under-frequency protection

For under-frequency protection, the client has specifiedthe following characteristics:

Alarm: 59.3Hz, 0.5s time delay

1st stage trip: 58.7Hz, 100s time delay

2nd stage trip: 58.2Hz, 1s time delay

Similarly, the overfrequency is required to be set as follows:

Alarm: 62Hz, 30s time delay

Trip: 63.5Hz, 10s time delay

These characteristics can be set in the relay directly.

17.22.2.10 Overvoltage protection

The generator manufacturers’ recommendation is:

Alarm: 110% voltage for 5s

Trip: 130% voltage, instantaneous

This translates into the following relay settings:

Alarm: 19800V, 5s time delay

Trip: 23400V, 0.1s time delay

17.22.2.11 Pole slipping protection

This is provided by the method described in Section17.7.3.2. Detection at a maximum slip frequency of 10Hzis required. The setting data, according to the relaymanual, is as follows.

Forward reach, ZA = Zn + Zt

= 0.02 + 0.22

= 0.24Ω

Reverse reach, ZB = ZGen

= 2 x X’d

= 0.652Ω

Reactance line, ZC = 0.9 x Z

= 0.9 x 0.22

= 0.198Ω

where:

Z1 = generator transformer leakage impedance

Zn = network impedance

• 17 •G

ener

ator

and

Gen

erat

or-T

rans

form

er P

rote

ctio

n


Table 17.5: Relay settings for large generator protection example

Is1 8%

Is2 100%

K1 0%

K2 150%

Vn3H< 166.3V

Vn3H delay 0.5s

Vsnvd 935.3V

Time Delay 0.5s

Xa1 -0.245ΩXb1 1.666Ωtd1 5s

Xa2 -0.1406ΩXb2 1.725Ωtd2 0.5s

tDO1 0s

Iset 7200A

K 3

V1set 14400V

V2set 3000V

I2>> 0.06

Kg 10

Kreset 10

tmin 1s

tmax 1300s

V> meas mode three-phase

V> operate mode any

V>1 setting 19800V

V>1 function DT

V>1 time delay 5s

V>2 setting 23400V

V>2 function DT

V>2 time delay 0.1s


P1 setting 1.6MW

P1 time delay 0.5s

P1 DO time 0s

Dead Mach I> 6000A

Dead Mach V< 9000V

Za 0.243ΩZb 0.656ΩZc 0.206Ωα 90°

θ 80°

T1 15ms

T2 15ms

F>1 setting 62Hz

F>1 time delay 30s

F>2 setting 63.5Hz

F>2 time delay 10s


P1 setting 1.6MW

P1 time delay 0.5s

P1 DO time 0s

F<1 setting 59.3Hz

F<1 time delay 0.5s

F<2 setting 58.7Hz

F<2 time delay 100s

F<3 setting 58.2Hz

F<3 time delay 1s

Protection Quantity Value

Differential protection

100% Stator earth fault

Neutral voltage displacement

Loss of excitation

Voltage controlled overcurrent

Negative phase sequence

Overvoltage

Underfrequency

Reverse Power

Inadvertent energisation

Pole Slipping Protection

Reverse Power

Overfrequency

Underfrequency

The nearest settable values are 0.243Ω, 0.656Ω, and0.206Ω respectively.

The lens angle setting, α, is found from the equation:

and, substituting values,

αmin = 62.5°

Use the minimum settable value of 90°. The blinder angle,θ, is estimated to be 80°, and requires checking duringcommissioning. Timers T1 and T2 are set to 15ms asexperience has shown that these settings are satisfactoryto detect pole slipping frequencies up to 10Hz.

This completes the settings required for the generator,and the relay settings are given in Table 17.5. Of course,additional protection is required for the generatortransformer, according to the principles described inChapter 16.

17.23 REFERENCES

17.1 Survey of Rate Of Change of Frequency Relaysand Voltage Phase Shift Relays for Loss of MainsProtection. ERA Report 95-0712R, 1995. ERATechnology Ltd.

α minmintan

.= − −+( )

−180 2 1 541o l

A B

RZ Z

• 17 •

Gen

erat

or a

nd G

ener

ator

-Tra

nsfo

rmer

Pro

tect

ion


17 Generator and Generator Transformer · PDF fileinterconnection between generator and...

Documents

Transcript of 17 Generator and Generator Transformer · PDF fileinterconnection between generator and...