CIGRE - International Guide On The Protection of Synchronous Generators

479
International Guide
on the
Protection of Synchronous Generators
Working Group
B5.04
October 2011
International Guide
on the
Protection of Synchronous Generators
Working Group
B5.04
Members
Gabriel Benmouyal (Convenor, CA), Fernando Calero (BO), Ronny Goin (NO),
Hans-Yoachim Herrmann (DE), Oscar Martinez Oterino (ES), Lars Messing (SE),
N.N. Misra (IN), Murty V.V.S. Yalla, Ph.D (USA), Tony Yip (UK), Sylwia Wroblewska (PL)
Copyright © 2011
“Ownership of a CIGRE publication, whether in paper form or on electronic support only infers right of
use for personal purposes. Are prohibited, except if explicitly agreed by CIGRE, total or partial
reproduction of the publication for use other than personal and transfer to a third party; hence circulation
on any intranet or other company network is forbidden”.
Disclaimer notice
“CIGRE gives no warranty or assurance about the contents of this publication, nor does it accept any
responsibility, as to the accuracy or exhaustiveness of the information. All implied warranties and
conditions are excluded to the maximum extent permitted by law”.
ISBN: 978- 2- 85873- 170-1

TABLE OF CONTENTS
1. Synchronous generators fundamentals

1.1 Introduction……………………………………………………………………………….... 1
1.2 Types of synchronous generators………………………………………………………...... 1
1.3 Synchronous generator prime movers……………………………………………………... 1
1.4 Generator parameters………………………………………………………………………. 2
1.5 Connections of generators to the power network………………………………………….. 2
1.6 Earthing of synchronous generators………………………………………………………. 3
1.6.1 High resistance earthing…………………………………………………………………....... 4
1.6.2 Low and medium-resistance earthing……………………………………………………….. 4
1.6.3 Reactance earthing…………………………………………………………………………..... 5
1.6.4 Other earthing methods……………………………………………………………………….. 5
1.7 Synchronous generators during a short-circuit……………………………………………. 5
1.8 Symmetrical component representation of generators…………………………………… 7
1.9 Generator steady-state phasors diagram……………………………………………………. 8
1.10 Generator steady-state power-angle characteristic…………………………………………. 11
1.11 Generator controls…………………………………………………………………………. 13
1.11.1 Governor controls……………………………………………………………………………. 13
1.11.2 Automatic voltage regulation (AVR)……………………………………………………… 14
1.12 Generators operation limits………………………………………………………………... 15
1.12.1 Generator capability curve in the P-Q plane……………………………………………… 15
1.12.2 Steady state stability limit…………………………………………………………………… 17
1.12.3 Underexcitation and overexcitation limiters……………………………………………… 18
1.12.4 Generator capability curve in the R-X plane............................................................... 20
1.13 Generator-network interactions……………………………………………………………. 20
1.14 References………………………………………………………………………………….. 21
2. Current state of generator protection
2.1 Introduction………………………………………………………………………………… 22
2.2 Multifunction generator protection systems (MGPS)……………………………………… 22
2.3 Typical applications……………………………………………………………………….. 23
2.4 Redundancy aspects………………………………………………………………………. 29
3. Stator phase fault short circuit protection
3.1 Introduction………………………………………………………………………………… 31
3.2 Generator differential protection…………………………………………………………… 31
3.2.1 Biased differential protection………………………………………………………….......... 32
3.2.2 High impedance differential protection…………………………………………………….. 34
3.2.3 Self-Balancing differential Protection……………………………………………………… 36
3.3 Inter-turn (or turn-to-turn) Fault protection………………………………………………… 37
3.3.1 Overcurrent protection……………………………………………………………………….. 37
3.3.2 Biased differential protection………………………………………………………………… 37
3.3.3 Split phase differential protection…………………………………………………………… 39
i
3.3.4 Zero sequence overvoltage protection………………………………………………………. 41
3.4 Phase fault overcurrent protection…………………………………………………………. 42
3.5 Overall generator transformer differential protection……………………………………… 43
3.6 Other phase backup protection…………………………………………………………….. 43
4. Stator earth fault protection
4.1 Introduction……………………………………………………………………………….... 45
4.1.1 Stator winding earthing systems……………………………………………………………... 45
4.1.2 Effects of stator phase-to-earth faults……………………………………………………….. 45
4.1.2.1 Overcurrent effect………………………………………………………………………………….. 46
4.1.2.2 Overvoltage effect………………………………………………………………………………….. 46
4.2 Stator earth fault protection schemes……………………………………………………….. 47
4.3 Generators directly connected to the grid…………………………………………………… 47
4.3.1 Solidly earthed generator…………………………………………………………………….. 47
4.3.2 .Resistor earthed generator…………………………………………………………………… 49
4.3.3 .Ungrounded generator………………………………………………………………………… 50
4.4 Generators connected to grid through a step-up transformer………………………………. 53
4.4.1 Ungrounded generators……………………………………………………………………… 53
4.4.2 High impedance stator earth fault protection……………………………………………… 54
4.4.2.1 Stator winding earth fault 59G-(95%) protection………………………………………….…. 54
4.4.2.2 Stator winding ground 59G-(100%) protection…………………………………………….… 56
4.5 References…………………………………………………………………………………. 61
5. Rotor earth fault protection
5.1 Introduction…………………………………………………………………………………. 62
5.2 Voltage divider method……………………………………………………………………… 62
5.3 AC Injection method………………………………………………………………………… 63
5.4 DC Injection method………………………………………………………………………… 64
5.5 DC Switching injection method……………………………………………………………... 65
5.6 Field protection for brushless generators……………………………………………………. 68
5.7 Impact on digital relays in this area…………………………………………………………. 68
5.8 Rotor earth fault tripping practices………………………………………………………….. 68
6. Underexcitation protection
6.1 Introduction…………………………………………………………………………………… 69
6.2 Theory, capability diagram of generators……………………………………………………. 69
6.3 Protection methods…………………………………………………………………………... 77
6.3.1 Measurement of field current…………………………………………………………………. 77
6.3.2 Impedance Measurement……………………………………………………………………… 78
6.3.2.1 Scheme 1- Offset mho circle method (negative offset)……………………………………….. 78
6.3.2.2 Scheme 2 - Impedance circle (positive offset) and directional element……………………. 80
6.3.3 Admittance measurement……………………………………………………………………... 81
6.3.4 Measurement of the reactive power…………………………………………………………. 83
6.3.5 Directional current principle…………………………………………………………………. 84
6.4 Impact of digital relays…………………………………………………………………….. 86
6.5 References………………………………………………………………………………….. 86
ii
7. Abnormal frequency operation
7.1 Introduction…………………………………………………………………………………... 87
7.2 Generator and turbine capability and protection design criteria…………………………….. 88
7.3 Conformance to IEC 60034:1996……………………………………………………………. 88
7.4 Load shedding schemes……………………………………………………………………… 89
7.4.1 Purpose of underfrequency load shedding schemes………………………………………. 89
7.4.2 Coordination of system and turbine-generator underfrequency protection……………. 89
7.5 Steam generation plants…………………………………………………………………… 90
7.5.1 Generator limits………………………………………………………………………………… 90
7.5.2 Turbine limits…………………………………………………………………………………… 90
7.5.3 Turbine underfrequency protection…………………………………………………………… 91
7.5.4 Plant auxiliary considerations………………………………………………………………… 92
7.6 Combustion turbine units……………………………………………………………………. 92
7.7 Combined cycle units……………………………………………………………………….. 93
7.8 Hydro units………………………………………………………………………………… 93
7.9 Nuclear generating units…………………………………………………………………… 94
7.9.1 Pressurized water reactor (PWR) plants……………………………………………………. 94
7.9.2 Boiling water reactors (BWR) plants………………………………………………………. 94
7.10 Requirements on frequency protection…………………………………………………… 95
7.10.1 Frequency accuracy…………………………………………………………………………… 95
7.10.2 Setting range…………………………………………………………………………………… 95
7.10.2 Speed……………………………………………………………………………………………. 95
7.10.3 Voltage influence……………………………………………………………………………… 95
7.10.4 Rate of change of frequency…………………………………………………………………. 95
7.11 Impact of digital protection systems application…………………………………………… 95
7.12 Referemces……………………………………..…………………………………………… 96
8. Breaker failure protection
8.1 Introduction………………………………………………………………………………… 97
8.2 Breaker failure logic…………………………………………………………………………. 98
8.3 Open generator flashover protection…………………………………………………………. 100
8.4 References……………………………………………………………………………………. 101
9. Unbalanced load (negative sequence current) protection
9.1 Introduction………………………………………………………………………………….. 102
9.2 Relationship of negative sequence and unbalanced three-phase current……………………. 102
9.3 System condition causing negative sequence current……………………………………….. 103
9.4 Other…………………………………………………………………………………………. 103
9.5 Effect on generator by negative phase sequence current…………………………………….. 103
9.5.1 Negative sequence generator damage ……………………………………………………… 103
9.5.2 Limit and against negative phase sequence current……………………………………… 103
9.5.3 Generator Negative phase sequence capability…………………………………………….. 104
9.6 Negative sequence protective schemes………………………………………………………. 106
9.6.1 Principle of negative sequence protection…………………………………………………... 106
9.6.2 Electromechanical protection……………………………………………………………….. 107
iii
9.6.3 Digital protection…………………………………………………....................................... 107
9.7 Impact of numerical protection in this area………………………………………………….. 108
9.8 Conclusion……………………………………………………………………………………. 108
9.9 References………………………………………………………………………………….… 108
10. System backup protection
10.1 Introduction………………………………………………………………………………… 109
10.2 Generator fault current…………………………………………………………………….. 109
10.3 Generator System Backup Relays……………………………………………………….… 110
10.3.1 Voltage dependant overcurrent Relays…………………………………………………….. 110
10.3.1.1 Voltage controlled overcurrent ………………………………………………………………… 111
10.3.1.2 Voltage restrained overcurrent ……………………………………………………………….. 112
10.3.1.3 General considerations…………………………………………………………………………. 113
10.3.2 Distance relay backup ……………………………………………………………………… 113
10.3.3 Ground fault system backup………………………………………………………………… 115
10.4 References………………………………………………………………………………… 115
11. Overexcitation (Volts per Hertz) and overvoltage protection
11.1 Introduction…………………………………………………………………………………. 116
11.2 Operation Limits as Defined by the Industry Standards……………………………………. 118
11.3 Volts per hertz Protection…………………………………………………………………… 118
11.4 Overvoltage Protection………………………………………………................................... 120
11.5 Coordination with Automatic Voltage Regulator ………………………………………….. 121
11.6 Impact of numerical protection in this area…………………………………………………. 121
11.7 References………………………………………………………………………………… 121
12. Out-of-step protection
12.1 Introduction………………………………………………………………………………… 122
12.2 Theory, static and transient stability………………………………………………………… 122
12.3 Protection Design…………………………………………………………………………… 126
12.3.1 Measuring principle ………………………………………………………………………..... 126
12.3.2 Blinder schemes ………………………………………………………………………………. 128
12.3.3 Lens scheme…………………………………………………………………………………… 130
12.3.4 Two zone schemes with straight lines……………………………………………………… 131
12.4 Impact on digital relays……………………………………………………………….…… 132
12.5 References………………………………………………………………………………… 133
13. Anti-motoring protection
13.1 Introduction………………………………………………………………………………… 134
13.2 Reverse power level for various type of prime mover……………………………………... 135
13.3 Sequential tripping of generators…………………………………………………………… 136
13.4 Impact of digital protection……………………………………………………………….... 137
14. VT applications and loss of voltage signals
14.1 Introduction………………………………………………………………………………… 139
14.2 Detection of VT fuse failure……………………………………………………………….. 139
14.3 Use of symmetrical components …………………………………………………………… 140
iv
14.4 VT ferroresonance…………………………………………………………………………. 141
14.5 Impact of digital protection………………………………………………………………… 142
15. Inadvertent generator energizing
15.1 Introduction…………………………………………………………………………………. 143
15.1.1 Operational Errors……………………………………………………………………………. 143
15.1.2 Breaker Head Flashover……………………………………………………………………… 144
15.2 Generator response to inadvertent energizing………………………………………………. 144
15.2.1 Generator response to three-phase energizing……………………………………………. 144
15.2.2 Generator response due to single phase energizing…………………………………….. 145
15.3 Generator damage due to inadvertent energizing…………………………………………... 145
15.4 Response of conventional generator protection to inadvertent energizing ………………… 145
15.4.1 Generator breaker failure protection……………………………………………………… 146
15.5 Dedicated protection schemes to detect inadvertent energizing …………………………… 147
15.5.1 Frequency supervised overcurrent relays…………………………………………………. 147
15.5.2 Voltage supervised overcurrent relays……………………………………………………. 148
15.5.3 Directional overcurrent relays……………………………………………………………… 150
15.5.4 Impedance relays……………………………………………………………………………… 150
15.5.5 Auxiliary contact enabled overcurrent relays……………………………………............. 151
15.6 Dedicated protection schemes to detect generator breaker head flashover…………………. 151
15.6.1 Modified breaker failure scheme …………………………………………………………... 152
15.7 Conclusions…………………………………………………………………………………. 152
15.8 References………………………………………………………………………………… 153
16. Generator tripping practice
16.1 Introduction……………………………………………………………………………….. 154
16.1.1 Security requirement………………………………………………………………………… 154
16.1.2 Selectivity requirement………………………………………………………………………. 154
16.2 Type of tripping schemes………………………………………………………………… 154
16.2.1 Simultaneous tripping………………………………………………………………………… 155
16.2.2 Generator tripping…………………………………………………………………………….. 155
16.2.3 Unit separation………………………………………………………………………………… 155
16.2.4 Sequential tripping……………………………………………………………………………. 155
16.3 Suggested trip logic for specific relay functions………………………………………….. 156
16.4 Control circuit redundancy………………………………………………………………... 158
16.5 Impact of numerical relaying in this area……………………………………………….… 158
17. Power plant transmission grid interaction
17.1 General………………………………………………………………………………………. 159
17.2 Phenomena in the grid influencing power plants and requirements for protection system… 159
17.2.1 Short circuits in the grid……………………………………………………………………. 159
17.2.2 Unbalance load……………………………………………………………………………….. 162
17.2.3 Undamped power oscillations (small signal stability)……………………………………. 163
17.2.4 Voltage instability…………………………………………………………………………….. 164
17.2.5 Sub-synchronous oscillations……………………………………………………………….. 166
17.3 Phenomena in power plants influencing the grid…………………………………………. 166
v
17.3.1 Excitation system or automatic voltage regulator malfunction…………………………. 166
17.4 Short circuit power………………………………………………………………………….. 168
17.5 References ………………………………………………………………………………….. 168
18. Oscillographic monitoring and sequence of events
18.1 Introduction…………………………………………………………………………………. 169
18.2 Oscillographic recorder configuration……………………………………………………… 169
18.2.1 Analog waveforms…………………………………………………………………………… 169
18.2.1.1 Current waveforms………………………………………………………………………………. 169
18.2.1.2 Voltage waveforms………………………………………………………………………………. 170
18.2.2 Binary input signals………………………………………………………………………… 171
18.2.3 Triggering inputs………………………………………………………………………………. 171
18.2.4 Output signals…………………………………………………………………………………. 171
18.2.5 Auxiliary supply……………………………………………………………………………… 171
18.2.6 Synchronization……………………………………………………………………………… 172
18.2.7 Recording……………………………………………………………………………………… 172
18.3 Record management……………………………………………………………………….. 172
18.3.1 Triggering……………………………………………………………………………………… 172
18.3.2 Communication and monitoring…………………………………………………………… 173
18.3.3 Record information…………………………………………………………………………… 173
18.4 Independent recording equipment vs. oscillograph function in digital relay………………. 173
18.4.1 Oscillograph function in digital relays…………………………………………………… 173
18.4.1.1 Advantages………………………………………………………………………………………. 173
18.4.1.2 Disadvantages……………………………………………………………………………………. 173
18.4.2 Independent recorder equipment…………………………………………………………… 174
18.4.2.1 Advantages……………………………………………………………………………………….. 174
18.4.2.2 Disadvantages…………………………………………………………………………………… 174
Annex A Example of generator protection settings calculation………………………….. 1-A

1. Technical data of unit equipment……………………………………………………………… 1-A
1.1 Generator…………………………………………………………………………………………… 1-A
1.2 Unit transformer…………………………………………………………………………………… 1-A
1.3 Generator circuit breaker………………………………………………………………………… 1-A
1.4 Auxiliary transformer……………………………………………………………………………… 1-A
1.5 Excitation transformer……………………………………………………………………………. 2-A
1.6 Phase short circuit power on the 110 kV bus bars……………………………………………. 2-A
2. The short-circuit electrical quantities…………………………………………………………. 2-A
3. Protection settings………………………………………………………………………..…… 8-A
vi
EXECUTIVE SUMMARY
WG B5.04 International Guide on the Protection of Synchronous Generators
Introduction
The guide comprises 18 chapters plus one Annex corresponding to the next topics:

6. Underexcitation (loss-of-field) Protection
8. Breaker failure Protection
14. VT applications and loss of voltage signals
Annex A Example of generator protection settings calculation
Guide Content
This chapter is an introduction to the basic physics principles of synchronous

generators. It presents typical values of generator impedances and time constants.
Earthing techniques for generators are introduced. Symmetrical components
representation of generators for short-circuit studies is addressed together with the
vector representation of generators in steady-state. A brief overview is made of
generator controls and operation limits. Function characteristics representation in the R-
X, P-Q and G-B planes is finally introduced.
The purpose of this section is to present the state-of-the-art in generator protection

applications. The current trend in the usage of digital multifunction generator relay is first
discussed. Four protection schemes for generator are then introduced. The first scheme
covers generators with power smaller than 5 MW (see figure 1), the second scheme
covers unit with power less than 100 MW, the third scheme is applicable to unit with
power between 100 and 200 MW and finally the fourth scheme is suitable for units with
power greater than 200 MW. The last paragraph of this section discusses the
application of redundant protection to increase dependability.
vii
Q1 Q2 T
Q1
I1 87G Generator
I3 I3 Differential
I2 67N Directional Earth
U Fault Protection
46.1 Negative Phase Alarm
I3
46.2 Sequence
I1 51V Overcurrent with
U Voltage restraint
49S Stator Overload Alarm
I1
I3 40 Loss of Field
U
U 32R Reverse
I3
I4 U Power
50ET Excitation
I4 Transformer
Ins.Overcurrent
Excitation Note 3 51ET Excitation
Transformer I4 Transformer
I2
(ET) Overcurrent
Uf 64R.1 Rotor Earth Alarm
If 64R.2 Fault
Q2
59 Over Note 1
U Voltage
+
27 Under Note 2
U
- T Voltage
81O Over Note 1
Uf U Frequency
If
81U Under Note 2
I1 U Frequency
Note 1: Only for hydro generators

Note 2: Only if required by independent generator producer
Note 3: Core balanced CT
Q1 – Generator Circuit Breaker Q2 – Field Circuit Breaker T – Prime Mover
Fig. 1 Protection system for generator (<5 MW) connected to a distribution system
Fig.2 Solution with full-redundancy
Stator phase faults occur when faults exist on one or more than one stator winding
without involvement of the ground. Stator phase faults are conventionally covered by
current differential relays. This chapter starts by covering the various stator winding
differential protection schemes: biased differential, high impedance or self-balancing
differential protection schemes. Inter-turn protection using overcurrent, biased or split-
phase differential or zero-sequence overvoltage protection scheme are then addressed.
The section finally discusses the application of phase overcurrent protection for small
generators and overall generator-step up transformer differential protection.

Stator earth faults are the most common faults on a generator. They usually occur due
to a breakdown of the insulation of the stator winding to earth through the stator core.
The chapter discusses the impact of a stator fault in terms of the resulting potentially
viii
destructive effects due to high currents and high voltages. Various techniques for the
detection of earth faults are introduced depending upon the type of connection to the
grid: direct or through a step-up transformer and the earthing type: ungrounded
generators, resistive or solidly grounded or connected to the earth through a distribution
transformer.
The chapter introduces the various techniques used for the detection of rotor-earth
faults: voltage-divider method, AC injection method, DC injection method and finally the
more modern DC switching injection method.
6. Underexcitation (loss-of-field) protection
A generator will go underexcited when the field supply is lost. Generator instability is the
frequent consequence of underexcitation state. This section starts by introducing the
two basic modes of operation of synchronous generators: over and under-excited. It
then discusses the limits of generator stability in the under-excited mode as an
introduction to the necessity of having a loss-of-field protection. The existing techniques
to detect loss-of-field are then introduced: direct measurement of the field current, dual
(negative) offset mho circles scheme, combination of impedance circle with positive
offset and admittance measurement with straight line characteristic in the admittance
plane.
In the introduction, it is reminded that both generators and turbines have limited
operation capability outside the rated frequency. A word of caution is introduced to the
effect that generators and turbines frequency capability characteristics very often are not
always public information and that any protection scheme should be discussed with the
manufacturer in order to get his approval. When designing a frequency protection
scheme, verification should be made if the IEC 60034:1996 standard is applicable. This
will have further constraints on the operation of the generator. Any generator/turbine
frequency protection scheme should be coordinated with an existing load-shedding
scheme. The chapter discusses frequency capability operation and protection schemes
for steam generation plants, combustion turbine units, combined cycle units, hydro units
and finally nuclear generation units.
8. Breaker failure protection
Breaker failure protection is necessary when a breaker fails to operate following fault
detection. This chapter starts by discussing the necessity for generator breaker failure
protection and then introduces the base logic for breaker failure protection. Modifications
to the base scheme to cover breaker flashover with generator breaker open are then
discussed.
Unbalance load will create ohmic losses in a generator that could end up with a
catastrophic level of damages on the generator. This section addresses the impact of
unbalanced load as measured by negative sequence current on a synchronous
generator. Two tables present the permissible negative sequence capability of salient
and cylindrical rotors and two other tables present the limiting K value for the same rotor
types. Inverse-type negative sequence current characteristics are then introduced.
ix
The purpose of generator backup protection is to clear faults external to the generator
zone when the primary protection has failed to do it. Faults external to the generator
zone cannot be detected using overcurrent relays because the fault current very often is
going to be below the load current. One type of conventional backup protection for
phase faults consists in using voltage dependent or voltage restrained overcurrent
elements or overcurrent element with undervoltage holding function. The other type of
conventional backup protection for phase faults is the usage of distance elements. For
ground faults, an inverse overcurrent element measuring the current in the generator
neutral is used for bus connected generator. For unit connected generators, an inverse
overcurrent element measures the current flowing in the transformer neutral.
Overexcitation occurs when the ratio of the voltage over the frequency exceeds the
design limits. The consequence of overexcitation or overfluxing is excessive heating that
could lead to destructive damages to the machine. Overexcitation protection is
accomplished using volts per hertz relay in the form of definite-time elements, inverse-
time elements or a combination of the two. Overvoltage protection is needed when it
could occur without exceeding the volts per hertz limits. Attention should be provided to
the coordination between the Automatic Voltage Regulator and any overvoltage
protection.
This section starts by defining the rotor-angle stability problem with a two-machine
system. The classical problem of the trajectory of the positive sequence impedance in
the complex plane during a power-swing is then addressed. Power swings detection
principles are then introduced: single or double-blinder schemes, scheme based on
lenticular characteristic and finally two-zone scheme with straight lines.
This chapter starts by determining the motoring power likely to exist with the different
prime movers in order to help set the real-power relay. Anti-motoring protection is
conventionally applied by using a real power relay associated with a definite-time delay.
Implementation of sequential tripping principle is then introduced with a particular logic
scheme.
14. VT Applications and loss of voltage signals
This section starts by listing the generator protective functions likely to be affected by a
loss of voltage signals. The concept of the conventional voltage balance relay with the
use of two sets of VT’s is then discussed. A digital scheme based on symmetrical
components to detect loss of voltage signal is presented. The chapter last paragraph
discusses VT ferroresonance issues.
When a generator is energized while off-line on turning gear, or coasting to a stop, it

behaves as an induction motor and can be damaged within a few seconds. Turbine
damage can also occur. Energizing of the generator can occur due to human error or
breaker flashover. The chapter addresses the various protective means to counter this
x
situation: frequency supervised overcurrent elements, voltage supervised overcurrent
elements, directional overcurrent elements or impedance relays. A paragraph addresses
the issue of generator breaker head flashover that needs to initiate the breaker failure
function.

Depending upon the fault type, various tripping schemes are available: simultaneous
tripping, generator tripping, unit separation or sequential tripping. This section provides
a table recommending a tripping scheme depending upon the fault type.

In this chapter, phenomenons in the transmission grid like short-circuits, unbalanced
load, voltage instability etc… that affect the operation of the generator and alternatively,
phenomenons in the generator like an excitation system malfunction that affect the
transmission grid are discussed.

This section discusses the relative merits of imbedding oscillographic capability and
sequence of events inside digital generator relays as opposed to implementing these
functions in external equipment. A number of issues like what analog waveforms and
binary signals to record during a fault, what triggering inputs to use to initiate the record
are also addressed.
Annex A Example of generator protection settings calculation
This section provides an example of protection elements settings calculation for a 270
MVA generator.
Conclusion
In this guide, an effort has been made to report on the nature and the implementation of all
the possible protective functions that could be encountered in a synchronous generator.
Trying to reflect the different protection principles in various countries and various continents
has also been a constant concern.
xi
1. SYNCHRONOUS GENERATORS FUNDAMENTALS
1.1 INTRODUCTION
Synchronous generators are the complex assembly of windings, copper and steel that
convert the mechanical energy supplied by a prime mover into electrical power supplied to
users by means of an electrical power network. Although faults in the generator are rare,
they could be destructive if not rapidly and properly taken care of by protective system.
1.2 TYPES OF SYNCHRONOUS GENERATORS
There exist two types of synchronous machines: round-rotor machines and salient-pole
machines (Figure 1.1). Generators driven by steam or gas turbines (turbo generators) have
cylindrical (round) rotors with slots in which distributed field windings are placed. Solid steel
forgings constitute the material of most of the cylindrical rotors though thick steel disks can
be the basic material. The number of poles could be two, four or six. When water wheels
(hydraulic turbines) are the prime mover, rotors are made of laminated salient-poles with
concentrated field windings and the number of poles is usually large.
The generator type is related to its rated speed. The relation between a generator rated
frequency f, its speed n in revolutions per minutes and its pole number P is:
P n
f = (1.1)
2 60
Whereas steam turbines have a higher speed, they require less number of poles to achieve
the rated frequency. On the contrary, hydro generators have a much slower speed and
therefore require a substantial number of poles in order to provide the rated frequency.
Figure 1.1 Round-rotor (a) and salient-rotor (b) synchronous generator
1.3 SYNCHRONOUS GENERATOR PRIME MOVERS
The prime sources for generating electricity are the kinetic energy of water or the thermal
energy derived from fossil fuels (coal, oil or gas) and nuclear fission. The prime mover is the
mechanical system that supplies the energy to the generator so that rotation at constant
speed is achieved.
When kinetic energy from water is being used, the prime mover will be a hydraulic turbine.
Hydraulic turbine could be of three types: impulse, reaction or propeller turbine. Hydraulic
generators will rotate at lower speeds between 100 and 300 RPM.
1
When fossil fuels or nuclear fission is used, a steam turbine will convert the high pressure
and high temperature in the steam into rotating energy to drive the generator. Generators
will then rotate at speed typically between 1500/1800 and 3000/3600 RPM for 50/60 Hz.
1.4 GENERATOR PARAMETERS
It is beyond the scope of this guide to address any synchronous generator electrical
modeling but there exist a number of parameters that characterize any generator.
First of these characteristics are those related to the generator rating as shown in Table 1.1.
Table 1.2 shows different impedances used to describe the characteristics of a synchronous
generator. Table 1.3 describes the different time constants of a synchronous generator.
1.5 CONNECTIONS OF GENERATORS TO THE POWER NETWORK
There are two ways to connect a generator to a power network: direct or by way of a step-up
transformer (unit connected).
Direct connection implies that the load voltage level is the same as the generator voltage
with no transformer between the two. This type of arrangement is used with small generators
in industrial applications.
In unit connection, the generator is connected to the power network by way of a step-up
transformer. This arrangement is applicable to practically all large generators where the
step-up transformer has a wye-delta connection. The generator is connected to the delta
windings so that earth fault currents are substantially reduced by way of a high impedance
earthing system.
Table 1.1 Generator Ratings
Description Unit
Rated power (S) kVA (kilovolt-amperes)
Rated Voltage (U) kV (kilovolts)
(Phase-to-phase voltage)
Cosϕ or Power Factor (pf) Unitless
Rpm Speed in revolutions per minute
Rated current (I) A
(amperes at rated power and
voltage)
Inertia constant WR2 per unit (pu)
2
Table 1.2 Generator Impedances*
Description Symbol Turbo-generator Hydro-generator

Typical values Typical values
(pu) (pu)
Direct-axis reactance xd 1.0-2.1** 0.65-1.50
Quadrature-axis reactance xq 1.0-2.1** 0.39-1.00
**
Direct-axis transient reactance xd’ 0.11-0.35 0.20-.50
**
Quadrature-axis transient xq’ 0.11-0.35 0.40-1.00
reactance
Direct-axis subtransient reactance xd’’ 0.1-0.24** 0.12-0.36
Quadrature-axis subtransient xq’’ 0.1-0.24** 0.22-0.46
reactance
Potier reactance xp 0.22-0.29 0.17-0.40
Negative sequence reactance X2 0.24-0.30 0.11-0.36
Zero sequence reactance X0 0.09-0.145 0.02-0.21
AC armature resistance ra 0.00137-0.008 0.003-0.015
Negative sequence resistance r2 0.0259-0.0045 0.012-0.20
*)
for generators with rated power P in the range = (75-600) MW
**)
saturated state
Table 1.3 Generator Electrical time constants*
Description Symbol Turbo-generator Hydro-generator

Typical values (s) Typical values(s)
Direct-axis transient open –circuit Td0’ 4.9-9.2 1.5-12
time constant
Direct-axis transient short-circuit Td’ 0.85-1.36 0.5-3.3
time constant
Direct-axis subtransient short- Td’’ 0.10-0.17 0.01-0.05
circuit time constant
Armature short-circuit time Ta 0.04-0.36 0.03-0.25
constant
*)
for generators with rated power P = (75-600) MW
1.6 EARTHING OF SYNCHRONOUS GENERATORS
The primary objective of the earthing technique for a generator is to limit the generator
current during a single-phase-to-earth short-circuit so that the thermal limit of the generator
will not be exceeded. The other objective is to limit the transient overvoltages during earth
faults. A generator is normally designed to be able to sustain a three-phase short-circuit for a
limited amount of time. Since a single-phase-to-earth short-circuit current will be higher in
3
magnitude than the three-phase short-circuit, it is necessary to use the proper grounding so
as to limit the single-phase short-circuit current to an acceptable magnitude.
There are basically 4 methods for earthing a generator:
• high-resistance earthing
• low and medium resistance earthing
• reactance earthing
• isolated
1.6.1 High-resistance earthing
High-resistance earthing is the most common, and it is normally used to limit the single
phase to earth current to a level that reduces the probability of iron damage. Normally the
resistor is chosen to reduce the current to a value between 5 and 20 A. Because of the low
fault current, the time delay of the protection can be longer (up to a few seconds).
1. High-resistance earth with a neutral transformer: with this technique, a low resistance
R is connected to the secondary of an earthing transformer with ratio N. The
resistance as seen from the primary is then N2*R.
2. High-resistance earth with a wye-earthed/broken-delta transformer: this earthing
technique is similar to the previous one and can be obtained by connecting a three
phase five-legged earthing transformer on the generator terminals with wye-
grounded primary and broken-delta secondary. In this case the load resistor is
connected across the broken-delta secondary winding.
3. High-resistance earth with a neutral earthing resistor: This technique is similar to the
previous ones with the same advantages. But in this case it is used a primary
resistor. Although this scheme would appear to be more expensive because of the
size of the required resistance, it is still the preferred solution in some countries
where it is considered as economical.
For delta connected generators only method 2 is possible.
1.6.2 Low and medium-resistance earthing
This technique will give a higher phase to earth current, normally from 200 A to 150% of
rated current, but will reduce the temporary over voltages during an earth fault. The time
delay of the protection must be small to reduce the iron damage.
1. Medium-resistance earth with a wye-earthed/delta transformer: With this method, a

wye-delta transformer is used. The primary windings are connected to the generator
leads with a resistor connected between the transformer neutral and the earth.
2. Low-resistance earth with a neutral earthing resistor: Similar as number 3, but with a
lower resistance.
For delta connected generators only method 1 is possible.
4
1.6.3 Reactance earthing
1. Low-reactance earth with a neutral earthing reactor: The single phase-to-ground fault
current will be many times higher than for high-resistance earthing.
2. Resonant earth with an earth-fault neutralizer: In this arrangement, a reactor is
connected to the secondary of an earthing transformer. The fault current in this
technique is limited to a very low value (less than 1A).
3. Capacitors connected between generator neutral and the earth. The capacitor divider
is used as the voltage transformer for zero sequence voltage measuring
1.6.4 Other earthing methods

1. Unearthed: No connection between the generator and the earth. The main advantage
is that fault currents are extremely low. Significant overvoltages can result from faults
or switching operations.
2. Hybrid earth: Both low- and high-resistance earthed systems are available in parallel.
If an earth fault is detected, the low-resistance path is removed by a breaker opening.
This scheme allows to have both low magnitude fault currents and to remove the
risks of transient overvoltages during a fault.
1.7 SYNCHRONOUS GENERATORS DURING A SHORT-CIRCUIT
The various impedances and time constants in Tables 1.2 and 1.3 appear in the
mathematical expression of the short-circuit current magnitude at the terminals of the
generator when a three-phase bolted short circuit is applied. If Uo is the open circuit voltage
at the output of the generator, the effective value of the armature short-circuit current AC
component will be:
Uo 1 1 −t / Td' 1 1 ''
I ac = +Uo( ' − )e + U o ( '' − ' ) e −t / Td (1.2)
xd xd xd xd xd
Expressing the steady state, transient, and subtransient currents as:
U0
Is =
xd
U0
I s + I 0' = (1.3)
x'd
U0
I s + I 0' + I 0'' =
x 'd'
we finally get for equation (1.2):
I ac = I s + I ' 0 e − t / T 'd + I "0 e − t / T "d (1.4)
The instantaneous value of the DC component (aperiodic) of the armature current can be
expressed as:
5
2U 0 cos α −t / Ta
idc = e (1.5)
x"d
where: α - angle that determines voltage phase in the fault moment
Combining the AC and DC current components and from equations 1.2 and 1.5, the
maximum effective value of short circuit current at the generator terminals will occur at time
zero and angle α zero and will be equal to:
2 2
U   2U 0  U
I a max =  0  +   = 3 0 (1.6)
 
 x" d   x" d  x" d
Figure1.2 shows the subtransient, transient and steady state periods of the short-circuit
current. Figure 1.3 shows the impact of the DC component on the three-phase short circuit
currents. The DC components in each of the three phases establishes itself so that the sum
of the three phase currents is always equal to zero.
Subtransient Period
Transient Period
Steady State Period
Current
Time
Extrapolation of
Steady-State Current
Extrapolation of
Transient Envelope
Figure 1.2 Generator three-phase short-circuit phase current
6
Current
Time
Phase A
DC Component
DC Component
Current
Time
Phase B
Current
Time
Phase C
DC Component
Figure 1.3 Typical generator three-phase short circuit currents with DC component
1.8 SYMMETRICAL COMPONENT REPRESENTATION OF GENERATORS
When making symmetrical component based fault studies, generators are represented as
voltage sources. It is then necessary to define the three types of source impedances
required: the positive, negative and zero sequence of the generator.
When a short-circuit occurs, as demonstrated by equation 1.2, the level of output current is
going to be determined by one of the three sequence impedances depending upon the time
after the short circuit inception. The highest possible short circuit current occurs at the
beginning of the short circuit and depends upon the subtransient reactance xd’’. For this
reason, for the purpose of short circuit studies, the subtransient reactance will be used as
the positive sequence reactance.
Negative sequence reactance can be measured by applying negative sequence currents to

the armature and by moving the field structure into a forward rotation with the rotor winding
short-circuited. The value of the negative sequence reactance x2 will lie between the direct
and quadrature subtransient reactances xd’’ and xq’’. For practical purposes it is very often
taken as the average of these two reactances or:
xd′′ + xq′′
x2 = (1.7)
2
7
If a single phase current is injected into the three armature windings connected in series and
the voltage measured across one phase winding then the ratio of the measured voltage
divided by the injected current will be equal to the machine zero sequence reactance x0.
This reactance has practically the lowest value of all the generator reactances. It varies
generally between 0.15 and 0.60 times the subtransient direct axis reactance depending
upon the armature winding pitch.
a) Positive sequence network
b) Negative sequence network
c) Zero sequence network
Figure 1.4 Synchronous generator sequence networks
The symmetrical reactance components can be used to calculate the generator short circuit
currents. (see Table 1.4)
1.9 GENERATOR STEADY-STATE PHASORS DIAGRAM
It is beyond the scope of this guide to derive the complete mathematical model of a
synchronous generator or to determine its transient state phasor diagram. However, for a
salient pole generator in steady-state balanced operation, the phasor diagram can be
derived as follows.
Figue 1.5 shows an idealized 2-pole synchronous generator with the location of the direct-
and quadrature axis with respect to the rotor. The Park transformation [1,2] allows defining
the generator output voltage and current with these two axis as the axis of reference. The
generator terminal voltage and the output current It phasors can be expressed as follows:
Ut = U d + jU q (1.8)
It = I d + j I q (1.9)
8
Table 1.4 Generator short circuit currents formulas
Kind of short Short circuit current

circuit on
the Subtransient state Transient state Steady state
generator
terminals
3 ph E01 E01 E01

I "sc ( 3 ph ) = I 'sc ( 3 ph ) = I sc (3 ph ) =
x"d x'd xd
2 ph 3E01 3E01 3E01

I "sc ( 2 ph ) = I 'sc ( 2 ph ) = I sc ( 2 ph ) =
x"d + x2 x ' d + x2 xd + x2
1 ph.-earth I"sc( 1 ph ) =
3 E01
I"sc( 1 ph ) =
3 E01
I sc( 1 ph ) =
3 E01
x"d + x2 + x0 + 3 Z N x'd + x2 + x0 + 3Z N xd + x2 + x0 + 3Z N
Figure 1.5 Idealized synchronous generator
The generator total power is provided:
S = U t I t* = ( U d + j U q )( I d − j I q ) (1.10)
S = P + j Q = ( U d I d + U q I q ) + j( U q I d − U d I q ) (1.11)
P = U d Id + U q Iq (1.12)
Q = U q Id − U d Iq (1.13)
The machine power factor is provided by the cosine of the angle between the phasor of the
output voltage and the phasor of the output current:
The total torque supplied to the machine is the real power produced plus the losses in the
armature:
Te = ( U d I d + U q I q ) + ra ( I d2 + I q2 ) (1.14)
9
The excitation voltage Ef is equal to the generator voltage at open circuit. Its relation to the
direct current if flowing into the field winding is provided by the next equation:
ω Laf i f
Ef = (1.15)
2
In this equation, Laf is the stator-to-rotor mutual inductance and ω is rated frequency in
rads/s.
The relation between the excitation voltage Ef (also called the internal voltage) and the
voltage at the generator terminal is provided by:
E f = U t + ra I t + j xd I d + j xq I q (1.16)
The angle δ between the internal voltage Ef and the output voltage Ut is the generator
internal angle. At no-load, this angle is obviously equal to zero.
Figure 1.6 represents the steady-state diagram of a generator with the following conditions:
P = 0.9, Q = 0.434, Ut = 1.0, It = 1.0, xd = 1.0, xq = 0.6
The current It phase angle θ can be computed by taking into account the fact that the
excitation voltage Et must lie on the quadrature axis and therefore its real part is equal to
zero. This leads to the following equation for θ:
 U t cos φ + ra I t 
θ = tg −1  

(1.17)
 U t sin φ + xq I t 
Figure 1.7 represents the steady-state diagram of the same generator absorbing time the
same amount of reactive power (lagging power factor or Q=-0.434).
By looking at Figure 1.6, one can see that in order for the generator to be overexcited
(lagging power factor), a greater magnitude of excitation voltage is required as compared to
Figure 1.7 where the generator is underexcited (leading power factor). On a generator fitted
with an automatic voltage control, the excitation voltage will be set automatically to the
proper value so that the generator voltage will be close to the rated value and the proper
reactive power will be supplied to the load.
The same reasoning can be applied to a round-rotor generator simply by putting xd equal to
xq.
10
quadrature
axis
j xq Iq
Ef
j xd Id
Uq Ra It
Ut
Iq It
φ=25.34
δ
θ=41
direct
Ud Id axis
Figure 1.6 Generator steady-state phasor diagram at lagging power factor
quadrature
axis
Iq j xq Iq
Ef It j xd I d
Uq ra It
Ut
δ φ=25.34
θ=53.85
direct
Id Ud axis
Figure 1.7 Generator steady-state phasor diagram at leading power factor
1.10 GENERATOR STEADY-STATE POWER-ANGLE CHARACTERITIC
Xd,Xq Xe
Ef Ut Es
Figure 1.8 Salient-pole machine connected to an infinite bus
11
Consider the system of Figure 1.8 consisting of a salient pole machine supplying an infinite
bus through an impedance Xe. Assuming that the excitation voltage Ef is constant, the
equation of the real power supplied by the generator to the infinite bus is provided by the
next equation:
E f Es Xd − Xq
P= sin δ + E s2 sin 2δ (1.18)
Xd + Xe 2 ( X d + X e )( X q + X e )
The equation of the imaginary power supplied by the machine is provided by the next
equation:
E f Es  cos 2 δ sin 2 δ 
Q= cos δ − E s2  +  (1.19)
Xd + Xe  X d + X e X q + X e 
Figures 1.9a and 1.9b show the plots of the per unit real and imaginary powers respectively
as a function of the excitation voltage Ef and the angle δ between Ef and Es (Es=1 pu, Xd=1
pu, Xq=0.6 pu, Xe=0.1 pu). It is interesting to note that when Ef goes to zero, the salient pole
machine can still supply real power to the system. Under the same condition, the machine
will absorb capacitive power irrespective of the angle δ.
3 2
Ef=3
Ef=3
2.5
1
Ef=2
Ef=2
2
0
1.5 Ef=1
Per Unit Power P
Ef=1
Per Unit Power P
Ef=0.5
1 -1
Ef=0 Ef=0
Ef=0.5
Ef=0.5
0.5
Ef=0 Ef=1
-2
0
Ef=2
-3
-0.5
Ef=3
-1 -4
0 30 60 90 120 150 180 0 30 60 90 120 150 180

Angle δ
Angle δ
a) Real power b) Reactive power
Figure 1.9 Plots of real and imaginary powers supplied by the salient-pole machine as a function of
the excitation voltage and rotor angle
Neglecting the effect of saliency and assuming that Xs, the synchronous reactance is equal
to Xd and Xq, the two equations can be simplified to:
E f Es
P= sin δ (1.20)
Xs + Xe
E f Es E s2
Q= cos δ − (1.21)
Xs + X e Xs + Xe
12
Figures 1.10a and 1.10b show the plots of the per unit real and imaginary powers
respectively in the case of a round-rotor machine as a function of the excitation voltage Ef
and the angle δ between Ef and Es (Es = 1 pu, Xs = 0.9 pu, Xe = 0.1 pu). When Ef goes to
zero this time, the round-rotor machine cannot supply any real power to the system. Under
the same condition however, the machine will still absorb a large amount of capacitive power
irrespective of the angle δ.
3.5 3
Ef=3 Ef=3
3 2
Ef=2
2.5 1
Ef=2 Ef=1
2 0
Per Unit Power P
Ef=0.5
Per Unit Power P

Ef=0
1.5 -1
Ef=0
Ef=1 Ef=0.5
1 -2
Ef=1
Ef=0.5
0.5
-3
Ef=2
Ef=0
0
-4
Ef=3
-0.5
-5
0 30 60 90 120 150 180
0 30 60 90 120 150 180
Angle δ Angle δ
a) Real power b) Reactive power
Figure 1.10 Plots of real and imaginary powers supplied by a round-rotor machine as a function of the
excitation voltage and rotor angle
1.11 GENERATOR CONTROLS
Synchronous generators connected to a power network should deliver the requested real
power (MW) at a constant output voltage (normal rated voltage), at constant frequency
(rated frequency) while they supply the reactive power (MVars) needed by the network.
The purpose of the constant frequency is to have all the generators run synchronously. The
real power requirement is to satisfy the network generation-load balance (total generation
equals total load plus the losses). The supply of reactive power is to satisfy the network
reactive equation: the sum of the reactive power produced must be equal to the sum of the
absorbed reactive power. The network reactive power balance is in turn closely associated
with voltage profile equilibrium.
In order to satisfy these requirements, two basic controls are normally implemented on a
generator: the load-speed turbine governor control and the controls associated with the
excitation system.
1.11.1 Governor controls
The turbine governor control can be of two types: speed-droop regulation or isochronous
control depending upon the unit operation.
13
Speed-droop regulation is a system where the generator speed is reduced as the load (or
generated megawatts) increases. Its purpose is to automatically share the load between
different generators connected to a network. Figure 1.11 provides an example of a speed-
droop characteristic. The speed-droop percentage R is the slope of the characteristic.
Isochronous governor control is a regulation type where the generator control is adjusted so
that the generator runs at synchronous rated speed irrespective of the load. An isolated
generator will normally have an isochronous control. On a network, one or more generators
will have an isochronous control so as to satisfy the requirement of constant frequency
requirement and the rest of the generators will have a speed-droop regulation so as to
automatically share the load between the generators.
Frequency (pu)
∆f
R= ∆P
f0
f
∆P
Power Output (pu)

1
Figure 1.11 Speed-droop characteristic
1.11.2 Automatic voltage regulation (AVR)
The excitation system is the system that supplies the DC voltage to the field winding. In
modern installation, this DC voltage is derived from the three-phase AC generator output by
way of a transformer and a solid-state rectifier bridge. The most common type of control
associated with the excitation system is the Automatic Voltage Regulator (AVR): the level of
DC voltage supplied to the field winding will be determined by a control loop so that the
generator output voltage magnitude will correspond to a preset value. This preset value in
turn is close to the rated output voltage. When the AVR has reached its quiescent point of
operation, the generator will supply or absorb an amount of reactive power corresponding to
the quantity required by load (as imposed by the network) so that the desired voltage
regulation is achieved. The generator could be operated so that it supplies reactive power
with a lagging power factor (overexcited state) or that it absorbs reactive power with a
leading power factor (underexcited state). In both cases, the amount of reactive power that
the generator can supply or absorb is limited by the generator capability characteristic. The
generator capability characteristic is imbedded into the controls generally by way of settings
accessible to the operator so that the AVR will automatically maintain the generator
operation within its capability.
Modern excitation systems use solid-state power components and have a very fast response
time. In critical conditions like short circuits, they help in increasing the synchronizing torque
so that the generator will remain in synchronism with the network. Figure 1.12 shows the
principle of a modern solid-state AVR that could switch to manual mode in case of the auto-
regulator failure. In manual mode, the field current is regulated in this example.
14
SCR Bridge
Field
Power
Circuit
Transformer
Voltage
Transformer
Manual Set Point If

Regulator
Auto Set Point Uset

Regulator
Figure 1.12 Self-excited automatic voltage regulator
1.12 GENERATORS OPERATION LIMITS
There are a number of physical principles governing the operation limits of a generator. The
most important of them are the thermal limits (Joule effect) of the armature (stator) and field
windings.
1.12.1 Generator capability curve in the P-Q plane
As shown in Figure 1.13, three thermal limits determine the operation of a generator in the
P-Q plane: the armature current (stator current) heating limit, the field current heating limit
and the stator end region heating limit.
The armature current limit is simply the maximum armature current that the generator can
carry on a continuous basis without exceeding the stator heating limitations. This limit in the
P-Q plane is simply a circle centered at the origin with a radius equal to the rated MVA of the
generator. On a per unit basis, the radius becomes simply equal to 1. It should be borne in
mind that the turbine has also a power limit and, as indicated in Fig. 13, the turbine power
limit could be smaller that the armature current limit so that the generator output will be
limited by the turbine capacity.
The field current limit is due to the heat dissipation in the field winding. It could be
demonstrated [2] that this limit translates in the P-Q plane as a circle with the following
center coordinates and radius:
U t2
Center coordinates = ( 0 , − )
Xs
(1.22)
X
Radius = ad U t i fd
Xs
As shown in Figure 1.13, the armature current limit and the field current limit circles intersect
at the point corresponding to the generator rated power factor. For the purpose of illustration,
this power factor has been represented as 0.85.
15
The stator-end heating limit is due to the localized heating taking place in the end-turn region
of a generator. The main magnetic flux in a generator is parallel to the stator lamination. The
armature end-turn leakage flux is an axial flux, perpendicular to the stator lamination. As a
consequence the produced eddy currents are the cause for the heating of the end region.
The field current is high when the generator is overexcited. The high magnetic field will
cause the retaining rings to be saturated. Under these conditions, the leakage flux will be at
a low value. For underexcited generators, the field current is low, the retaining rings are not
saturated and the leakage flux will be high. Also when the generator is underexcited, the flux
due to the armature current adds to the flux due to the field current: this has the effect of
enhancing the axial flux in the end region.
Assuming that the thermal energy produced by the eddy currents is proportional to the
square of the end region flux and that the end core leakage flux is proportional to the main
air gap flux, reference [4] has shown that the end region limit curve is a circle with the
following center coordinates and radius:
U t2
Center coordinates = ( 0 , K 1 )
Xd
(1.23)
U
Radius = K 2 t
Xd
In equation (1.23), the constant K1 and K2 are defined as follows:
N a N f − N 2f
K1 = (1.24)
N a2 + N 2f − 2 N a N f
∆θ
K2 = (1.25)
Kt ( N a2 + N 2f − 2 N a N f )
In equations (1.24-25), Nf and Na are the number of turns in the field and armature winding,
∆θ is the maximum permissible continuous temperature rise above the no-load temperature
in the end region and finally Kt is a proportionality constant relating the thermal energy with
the square of the end region magnetic flux.
Regarding thermal limits due to stator-end effects, the following should be borne in mind:
a) All generator types are not subjected to stator-end effect. Hydro machines do not
have in principle this limitation. Also, some units may be equipped with magnetic
shield that will limit the thermal effect. A user should always therefore refer to the
limit curves as provided by the machine manufacturer.
b) Hydrogen cooled generators exhibit multiple capability curves depending upon
the H2 pressure.
16
Reactive power (pu)
1 Field Current
Limit
Overexcited
Rated MVA pf
Armature Current
Limit
0.85 pf
0 1 pu
Real Power (pu)
Turbine Limit
Underexcited
Stator End
Region Limit
-1
Figure 1.13 Generator Capability Curve
1.12.2 Steady state stability limit
Consider the power system of Figure 1.8 consisting of a generator, where saliency is
neglected, connected to an infinite bus through an impedance Xe. Xs is the generator
synchronous reactance. We want to determine the generator steady-state stability limit in the
P-Q plane in terms of the generator terminal voltage Ut and the network impedances that are
assumed to be constant.
The steady-state stability limit or manual steady-state stability limit is the generator
operational limit with constant excitation voltage. It could be demonstrated [5] that the
stability limit is a circle, the center coordinates and the radius are provided by the following
equation:
 x d2 − xe2 − Re2
U t2 
Center coordinates = ( 0 ,   )
2 2
[
 x d Re + x e ( x e + x d ) ] 
(1.26)
U 2  R 2 + ( xe + x d ) 2 
Radius = t  2 e 2 
[
2  x d Re + xe ( x e + x d )  ]
If we assume that the system impedance resistance is zero, the equations simplify to:
17
U t2  1 1 
Center coordinates = ( 0 ,  −  )
2  xe x d 
(1.27)
U t2  1 1 
Radius =  + 
2  xe x d 
Figure 1.14 shows the location of the stability limit for different values of xd and xe at 1.0 pu
terminal voltage Ut.. The half-circle with 1.0 pu radius represents the machine rated power
capability. Figure 1. 14 indicates that instability of the machine will occur in the under-excited
mode for value of xd that are smaller than 1. Figure 1.14 indicates also that the limit of power
output is dependant upon the external reactance xe.
The steady-state stability limit shown here is applicable only to manual operation. Different
limits will exist when the AVR is taken into account. They can be determined using small-
signal stability techniques [6,7].
1.5
Unit capability circle

1
0.5
X =1.4 X =1.0
d d
Per Unit Q
Xe=0.9 Xe=1.0
0
-0.5
Xd=1.4
X =0.8
d
X =0.2
e X =0.2
e
-1
-1.5
0 0.5 1 1.5 2 2.5 3
Per Unit P
Figure 1.14 Manual steady-state stability limit with respect to the unit-circle capability
1.12.3 Underexcitation and overexcitation limiters
There exist a variety of techniques to insure that a generator operation will be restricted
automatically within the limits of its capability curve. One of the most commonly used
approach is to have limiters imbedded in the generator controls.
Very often, the Automatic Voltage Regulation will host these limiters so that the generator
capability limits will be implemented automatically and on a continuous basis. As shown in
Figure 1.15, there are two limiters in the AVR and their characteristics have to be entered in
the form of settings.
The first limiter is the overexcitation limiter (OEL) that is set just below and coordinates with
the field current thermal limit. It restricts the generator operation in the overexcitation region.
The second limiter is the underexcitation limiter (UEL) that is set just above and coordinates
18
with the stator end region thermal limit. It restricts the generator operation just above the
stator end iron limit.
Reactive power (pu)
1 Field Current
Limit
Overexcited
Rated MVA pf
Overexcitation
Limiter (OEL)
Armature Current
Limit
0.85 pf
0 1 pu
Real Power (pu)
Underexcitation
Limiter (UEL)
Underexcited
Steady-State
Stator End Stability Limit
Region Limit
-1
Figure 1.15 Generator capability curve and limits
Models of common underexciation limiters are described in [8]. Figure 1.16 shows the model
of an underexciter limiter of type II (UEL2) with a straight-line characteristic: when the point
of operation of the generator falls below the line, the limiter will in turn generate a positive
error signal that will be sent to the AVR summing point. The AVR will raise the generator
voltage until the point of operation will get above the line. Consequently in many
technologies, an UEL limiter will increase the AVR voltage setpoint whereas the OEL will
decrease the voltage setpoint in order to decrease the field current.
QxF1 KUQ
Q
1 + s TUQ
F2 KUV VUEL to AVR

F 2 = ( E t )k 2 Summing
1 + s TUV
Ut F 1 = ( Et )− k 1 Point
KUP
P
PxF1 1 + s TUP
Figure 1.16 UEL2 model with straight-line characteristic
Q
KUV
KUP
KUV ( k1 + k2 )
Ut
KUQ KUP
slope =
KUQ
Figure 1.17 UEL2 straight-line characteristic and settings
19
1.12.4 Generator capability curve in the R-X plane
Some generator protective functions or elements have a characteristic in the R-X plane.
Often, it could become important to verify that these characteristics coordinate with the
generator capability curve in the P-Q plane. This can be accomplished by converting the
characteristic in the P-Q plane into a characteristic in the R-X plane. The following
transformation can be used to convert the generator capability curve from P-Q plane to R-X
plane:
U2 P
R=
P2 + Q2
(1.28)
2
U Q
X =
P2 + Q2
In this equation U represents the rated line-to-line voltage of the generator.
Alternatively, the transformation for a characteristic in the R-X plane to a characteristic in the
P-Q plane can be accomplished using the following transformation:
U2 R
P=
R2 + X 2
(1.29)
U2 X
Q=
R2 + X 2
Another mode of representation of a function characteristic is the admittance or G-B plane. A

characteristic in the P-Q plane can be mapped into the G-B plane thanks to the next
transformations:
P
G=
U2
(1.30)
−V
B=
U2
Alternatively, a characteristic in the R-X plane can be mapped in the G-B plane following the
next transformations:
R
G=
R2 + X 2
(1.31)
−X
B=
R2 + X 2
1.13 GENERATOR-NETWORK INTERACTIONS
By way of the two basic regulation loops found on a generator, the excitation control
providing the regulated voltage output and the turbine governor control, a synchronous
generator participate to the two basic equilibriums determining the overall stability of a power
20
network: For the real power, the total generated real power must be equal to the total power
dissipated into the loads plus the line losses. For the reactive power, the net sum of the
generated reactive power (by all generators or synchronous condensers) must be equal to
the net sum of all the reactive power in all the passive elements in the power network.
When an unbalance develops in either of these two equilibriums, the generator automatic
controls (turbine governor or AVR) will be called upon so that a new point of stable operation
will be reached on the network. When a major disturbance occurs on a network, the
possibility exists that a new point of stable operation could not be found. A loss-of-
synchronism could occur between some parts of the network when the unbalance is related
to the real power. Alternatively, when the unbalance concerns the reactive power the
network voltage profile could be impacted leading to the most extreme cases where a
voltage collapse phenomenon can occur.
Different stability aspects and generator-network interactions are treated in Chapter 17.
1.14 REFERENCES
[1] Edward W. Kimbark, POWER SYSTEM STABILITY: SYNCHRONOUS MACHINES,

Dover Publications, Inc., New York, 1968.
[2] Prabha Kundur, POWER SYSTEM STABILITY AND CONTROL, McGraw-Hill, Inc.
1994.
[3] Westinghouse Electric Corporation, ELECTRICAL TRANSMISSION AND
DISTRIBUTION REFERENCE BOOK, 1964.
[4] S. S. Choy and X. M. Xia, “Under excitation limiter and its role in preventing excessive
synchronous generator stator end-core heating”, IEEE Trans. Power Syst., Vol.15, No.
1, pp.95-101, February 2000.
[5] J.T. Carleton, P.O. Bobo, and D. A. Burt, "Minimum Excitation Limit for Magnetic
Amplifier Regulating system," AIEE Trans., Vol.PAS-73, pp. 869-874, August 1954.
[6] Donald Reimert, “Protective Relaying for Power Generation systems”, Taylor & Francis
Group.
[7] Gabriel Benmouyal, “The Impact of Synchronous Generators Excitation Supply on
Protection and Relays”, 34th Annual Western Protective Relay Conference, Spokane,
WA, October 16–18, 2007.
[8] IEEE Recommended Practice for Excitation System Models for Power system Stability
Studies, IEEE standard 421.5-1992.
21
2. CURRENT STATE OF GENERATOR PROTECTION
2.1 INTRODUCTION
Generator protective relaying has gone through various generations of relaying technologies.
These include discrete electromechanical relays, static relays and now to numerical
multifunction protection systems. Several protection schemes in service today are either
discrete electromechanical or static relay types. These relays have a long history of
providing reliable protection. However, with the availability, additional performance,
economic advantages of numerical multifunction protection systems, this technology is being
incorporated into most new protection schemes. In most cases, new generators are being
protected with either dual multifunction generator protection systems or a single multifunction
generator protection, possibly backed up by some single function relays. Some modern
excitation systems contain protection functions which may be considered as backup.
This chapter details numerical multifunction technology which is being widely used around
the world for generator protection. It details its advantages as well as its limitations in terms
of protection redundancy.
This chapter also details typical applications based on various sizes of generators. It also
suggests various redundancy schemes with multifunction relays to improve protection
reliability.
2.2 MULTIFUNCTION GENERATOR PROTECTION SYSTEMS (MGPS)

Numerical relaying technology offers several additional features which were not available
with earlier technologies. These features include: metering of voltages, currents, and other
measurements; oscillography; sequence of events capture with time tagging; remote setting
and monitoring through communications; user configurability of tripping schemes and other
control logic; reduced panel space and wiring; low burden on the VTs and CTs; continuous
self-supervision.
The MGPS is made up of two or more functions implemented on a single hardware platform.
It has analog inputs (voltage and current signals), digital inputs (status inputs), and digital
outputs for sending trip and alarm signals. It may also have bi-directional communication
ports for communicating with the external world. Internal hardware consists of an analog
data acquisition system which converts analog signals to digital samples after signal scaling
and filtering (anti-aliasing). The digital subsystem consists of a microprocessor, flash
memory for program storage, random-access memory (RAM) for temporary storage of
information, and electrically erasable programmable memory (EEPROM) for storage of
setpoints.
The functional operation and performance of the MGPS are determined by both hardware
and software programs. Digital signal-processing algorithms are used to filter the voltage
and current input signals and calculate the parameters required for the relaying functions.
Digital filtering has significantly improve the accuracy of voltage and current phasors
computation. Conventional filtering system allow the acquisition of the fundamental
component (60 or 50 Hz) with the rejection of the harmonics and any DC component in the
waveform. Features like frequency tracking, that have become standard in most MGPS,
allow computing the voltage and current phasors across a broad frequency interval (typically
between 15 and 75 Hz). The relay logic program compares the setpoints to the calculated
parameters and implements the required time delay characteristics. The software program
also implements other features such as communications, oscillography, event recording,
and local user interface.
22
2.3 TYPICAL APPLICATIONS
There is no available standard which describes the required functions. Each country/utility
has its own recommendations for the necessary protection functions.
Some selection criteria are:
- Type of the power plant and operation conditions

- Size of the generator
- Importance of the generator
- Criteria imposed by the nature of the grid or the power system
- Costs for repair and energy lost if faults are undetected or trips with a too long delay
- Budget for the protection in relation to the cost of the generator (small generator
lower number of protection functions), etc…
Table 2.1 gives an overview of recommended protection functions in relation of the
generator size (generator/plant costs).
The following figures show some protective applications for different types of generators.
In figure 1 a typical bus bar application of a small generator is shown.
Figure 2 shows a unit application without a generator circuit breaker. This is a typical design
for medium sized generators.
Figure 3 shows a typical unit connection for larger generators. Via the auxiliary transformer
the generator provides the auxiliary power supply. The application is without a generator
circuit breaker.
Figure 4 shows an application for very large generators. Additionally to figure 3 a generator
circuit breaker increases the flexibility of operation, several large sized generators (>200MW,
between 200MW -550MW), were designed without a generator circuit breaker however. The
auxiliary power supply transformer is a three winding type.
23
Table 2.1 Overview of recommended protection functions
(X - recommended, o optional, Y- at pump storage station for motor operation)
Protection functions ANSI-No. Generator – Rated Power (MVA)

<5 5 - 50 50 - 200 > 200
Stator earth fault 90 % 64, 59N, 67N X X X X
Stator earth fault 100 % 64 (100%) X X
27TN
Differential 87G, 87T o X X X
Overcurrent time 50, 51V X X o o
Impedance 21 o X X
Interturn fault X1) X1)
Rotor earth fault 64R o X X X
Unbalanced load 46 X X X
Underexcitation 40 o X X X
Out of step 78 o
Stator overload 49 X X X X
Rotor overload 49R X
Overvoltage 59 X X X X
Frequency f > 81 X X X X
Frequency f < 81 X X X
Reverse power 32 X X X X
Undervoltage 27 Y Y Y Y
Overexcitation (U/f) 24 o X X
1)
recommended on generators with split phase winding recommended)
24
Q1 Q2 T
Q1
I1 87G Generator
I3 I3 Differential
I2 67N Directional Earth
U Fault Protection
46.1 Negative Phase Alarm
I3
46.2 Sequence
I1 51V Overcurrent with
U Voltage restraint
49S Stator Overload Alarm
I1
I3 40 Loss of Field
U
U 32R Reverse
I3
I4 U Power
50ET Excitation
I4 Transformer
Ins.Overcurrent
Excitation Note 3 51ET Excitation
Transformer I4 Transformer
I2
(ET) Overcurrent
If 64R.2 Fault
Q2
59 Over Note 1
U Voltage
+
27 Under Note 2
U
- T Voltage
81O Over Note 1
Uf U Frequency
If
81U Under Note 2
I1 U Frequency
Note 1: Only for hydro generators

Note 2: Only if required by independent generator producer
Note 3: Core balanced CT
Q1 – Generator Circuit Breaker Q2 – Field Circuit Breaker T – Prime Mover
Figure 2.1. Protection System for generator (<5MW) connected directly to a distribution
system
25
110kV (220kV)
Q2 Q5
Q1 Q1 Q4 T
I2 87G Generator
I3 Differential
I6
I1 87U Overall
I7 I4 Transformer
6kV 6kV
Unit I6 Differential
Transformer 59NG Stator Earth Fault
(UT) U1 (95%)
Q4 Q5
46.1 Negative Phase
I346.2 Sequence
I2 21.1 Under Impedance
U2 21.2
I4 59 Over Voltage
U2
51G Stator Overload
I2 Alarm
I3 40 Loss of Field
U2
24 Overfluxing
U2 U2 (V/Hz)
I5 I3' 32R Reverse
U2 Power
Note 1 51A Auxillary Supply
I3' I4 Overcurrent
Excitation
51N.1 HV Unit Xsformer
Transformer I7 51N.2
I3 Earth Fault
(ET)
50ET Excitation
I5 Transformer
Ins.Overcurrent
Q2
51ET Excitation
+ I5 Transformer
Overcurrent
- T 51oET Excitation
I5 Transformer Alarm
Uf I2 Overload
If
If 64R.2 Fault
I1 27 Under
U2
Voltage
81U Under
U2 Frequency
U1
B
Note 1: Metering Class CT

Q1 – Unit Circuit Breaker Q2 – Field Circuit Breaker Q4, Q5 – Auxiliary Circuit Breakers, T – Prime Mover
Figure 2.2 Protection System for generator transformer unit less than 100MW
26
110kV (220kV)
6kV 6kV
Q1
Q4 Q5 I8 A
B I9
B I7 Q2 Q5
Q5 Q2 A I10 Q1 Q4 T
Auxiliary Unit
T Q4 Q1 Transformer Transformer
B
A (AT) (UT)
I2 87U Unit
Generator 87G I1 I6 Differential
Differential I3 I9
A
I5 87TA Aux Xsformer
Unit Xformer 50TU I4 I7 Differential
Instantaneous I8 B I6
Overcurrent I2' 32R Reverse
U4 Power
Stator Earth 59NG B I5
Fault (95%) U3 40 Loss of field
I2
Reverse 32R I3' U4
Power U2 B U4 27 Under Voltage
U4
Under 21.1 I1
Impedance 21.2 U2 I2 51/27 Inadvertent
I13 Energising
Over Voltage 59 U2 A U4
A
U2
U1 59NG Stator Earth
Fault 95%
Alarm Stator 49S U3 A
Overload I1 46.1 NPS Thermal
I2
Excitation Note 1 46.2
Over Fluxing 24
V/Hz U2 Transformer (ET) A I3' 59 Over Voltage
U4
HV Unit 51N.1
Circuit Earth I10 A I3 81< Under
Fault 51N.2 U4 Frequency
Aux Xformer 50TA
Q2
Alarm
Uf 64R Rotor Earth
Instantaneous
Overcurrent
+ If Fault
Aux Xformer 50TA I4

- T
Overcurrent
Note 1
Alarm Aux Xformer 51 B I2'
Oveload TAo
Excitation 50TF Uf B I2
B
Xformer If
Instantaneous A I1
Overcurrent
Excitation 51TF
Xformer I13
Overcurrent U1
B
Alarm Excitation 51
Xformer TFo
Overload
Note 1 Metering Class CT

Q1 – Unit Circuit Breaker Q2 – Field Circuit Breaker Q4, Q5 – Auxiliary Circuit Breakers, T – Prime Mover
Figure 2.3 Protection system for Generator-transformer unit (100MW < P < 200MW)
27
28
2.4 REDUNDANCY ASPECTS
Traditional applications of generator protective relays involved separate relays performing

different functions with some overlap and backup where appropriate. In general, the
generator differential relay was connected to a dedicated set of CTs due to reliability,
burden, and CT characteristic matching issues. In the case of MGPS only one set of CTs are
connected and the MGPS offer very low burden. The differential and other protection
functions are implemented using this single set of current measurements without
deterioration of performance caused by CT burden. The use of single set of CTs is a
concern to many application engineers because CT inputs are not duplicated in this scheme
resulting in lower reliability. However, for schemes requiring high reliability two MGPSs with
separate CT and VT inputs are used.
Integrating many protective functions into one package raises issues concerning reliability.
For medium to large generators this issue has been addressed in a number of ways and
some of them are discussed here.
For the selection of the number of MGPSs and the relaying functions in each of them the
same arguments are valid as discussed in the beginning of chapter 2.3. Three scenarios are
discussed here.
a) Single MGPS (No redundancy)

The basic scheme is shown in Figure 2.5. All functions are integrated in the multifunction
numerical protection device (denoted as Protection). This device has all required protection
functions and additional features. The self supervision detects internal failures and gives an
alarm. In the case of a failure of MGPS the unit is without any protection. In such a case a
shut down of the unit is recommended.
Figure 2.5: Solution without redundancy
b) Two MGPS with single battery, CT, VT and trip circuit (Partial redundancy)
In this application at least two protection devices are used (see figure 2.6). If one relay fails
the necessary functions are available in the other device and there is no reason to shut
down the plant. Partly redundancy means that there is a cost saving in the power supply of
the relays and in the number of CT‘s and VT’s. The circuit breaker has one or two trip coils.
29
Figure 2.6: Solution with partly redundancy
c) Two MGPS with dual battery, two sets of CTs and VTs (Full redundancy)
This is the typical application for larger units and uses consequently the main 1 (protection
A) and 2 (protection B) protection. Sometimes the application is also called group A and
group B. The main difference to b) is that all important components are duplicated. In
addition to two separate protection systems the full redundancy scheme also has two battery
systems (sometime separate operation or decoupled via diodes), two sets of CT’s and VT’s.
In this principle the circuit breakers have two trip coils.
The basic concept of this application shows figure 2.7. The redundancy philosophy is only
shown for the generator. If we apply this concept for each of the protected zone (generator,
step-up transformer and auxiliary transformer) a total of 6 multifunction protection systems
are necessary. If the hardware and functionality is integrated in one larger module then two
modules are necessary.
Protection B
Protection A
Figure 2.7: Solution with full redundancy
30
3. STATOR PHASE FAULT SHORT CIRCUIT PROTECTION
3.1 INTRODUCTION
Generator stator phase faults are always considered to be serious because of the high
currents involved and the severe damage to the windings and the stator core. The phase
fault current level is not limited by the method of earthing at the generator’s neutral point.
The faults can occur between phase windings located in the same stator slot, or in the
winding terminals positioned side by side outside the stator slots. For insulation failure
between phases in the same slot, it will normally develop very quickly into an earth fault.
The extent of the damage depends upon the fault current level and the duration of the fault.
Furthermore, the fault current does not cease even after the field is tripped and the
generator isolated from the system. The energy stored in the field will continue to supply the
fault current for several seconds.
Another danger arising from the fault is fire, which could cause even greater damage. A
large proportion of the insulating material is inflammable. The risk of fire occurring is
dependent on the cooling system, more so for the air-cooled with forced ventilation and less
so for hydrogen-cooled machines.
Long repair times for severely damaged generators can be very expensive, especially
because of the loss of revenue and the cost of importing additional energy whilst the
machine is out of service.
Protection should be applied to limit the degree of damage in order to reduce the repair time
and to minimize the cost involved. For primary generating plant, high-speed disconnection
of the plant from the power system may be necessary to maintain system stability. The use
of rapid de-excitation methods may also be justifiable to produce faster decay of fault
currents in order to minimize the damage.
3.2 GENERATOR DIFFERENTIAL PROTECTION
For generators rated above 1 MVA, it is common to apply generator differential protection for
stator phase faults. This form of unit protection allows discriminative detection of winding
faults, with no intentional time delay, where a significant fault current arises. The zone of
protection, defined by the location of the CTs, should be arranged to overlap protection for
other items of plant, such as a busbar or a step-up transformer.
For some earthing arrangement of the generator neutral, such as low impedance earthing
where the earth fault current is limited from 200A to 400A, it is possible for the generator
differential to detect earth faults. Earth fault detection will be inadequate for high impedance
earthing, where the earth fault current is normally limited to 10A or in the worst-case
condition up to 20A.
Circulating current differential protection operates on the principle that current entering and
leaving a zone of protection will be equal. Any difference between these currents is
indicative of a fault being present in the zone. If CTs are connected as shown in Figure 3.1,
it can be seen that current flowing through the generator windings will cause current to
circulate around the secondary wiring. If the CTs are of the same ratio and have identical
magnetising characteristics, they will produce identical secondary currents and hence zero
current will flow through the relay. If a fault exists inside the windings, there will be a
difference between the outputs from each CT. This difference will be flowing through the
relay causing it to operate.
31
IF IF
A
If If
B
If
87G 87G 87G
Figure 3.1 Principle of circulating current differential protection
Heavy through current, arising from an external fault condition can cause one CT to saturate
more than the other, resulting in a difference between the secondary current produced by
each CT. It is essential to stabilise the protection for these conditions. Two methods are
commonly used. A biasing technique, where the relay setting is raised as through current
increases. Alternatively, a high impedance technique, where the relay impedance is such
that under maximum through fault conditions, the current in the differential element is
insufficient for the relay to operate.
3.2.1 Biased differential protection
In a biased differential relay, the through current is used to increase the setting of the
differential element. For heavy through faults, it is unlikely that the CT outputs at each zone
end will be identical, due to the effects of CT saturation. In this case a differential current
can be produced. However, the biasing will increase the relay setting, such that the
differential spill current is insufficient to operate the relay.
For numerical relays, the differential and bias currents are derived by algorithmic calculation,
after measurement of the individual CT secondary currents. The relay design will normally
have full galvanic separation of the neutral-tail and terminal CT secondary circuits. There are
two different solutions to calculate the differential and the bias currents, as shown in Figure
3.2.
1. The positive definition of the currents is according to the load currents
Differential Current: I diff = I1 - I 2 (3.1)
Biased (or restraint) current: I bias = I restr. = I1 + I 2 2 (3.2)
2. The positive definition of the currents is towards to the protective object (towards to the
generator)
Differential Current: I diff = I 1 + I 2 (3.3)
32
Biased (or restraint) current: I bias = I restr. = ( I 1 + I 2 )2 (3.4)
In both cases the calculation of the differential current is the same. The sign (- or +) is
determined by the positive definition of the currents. The main difference is in the calculation
of bias current and which leads to different operating ranges in the characteristic. From the
user’s point of view, the major consideration is in the selection of parameters for the
characteristic, mainly the inclination of the straight line.
Figure 3.2 Generator biased differential protection
A typical duel-slope operating characteristic of a biased differential protection is shown in

Figure 3.3. For general applications, the differential current threshold setting Is can be set as
low as 5% of the rated generator current, to provide protection for as much of the winding as
possible. The K1 slope would typically be set to 0% in order to maintain protection sensitivity
irrespective of the generator loading condition. This sensitive setting is only allowed if the
current transformers are well designed i.e., same types on both sides and good transient
behaviour even in the case of small fault currents with large DC time constants. The bias
slope break-point threshold setting Is2 would be set to a value above the generator rated
current, say 120%, to achieve external fault stability in case of transient asymmetric CT
saturation. Bias slope K2 would typically be set at 15-20%.
33
Figure 3.3 Generator biased differential protection
The threshold setting can be raised for applications where a relatively small unit auxiliary
transformer or excitation supply transformer is teed-off within the zone of differential
protection. In such cases, the differential setting should be set so as to prevent operation for
faults on the LV side of the transformer, or to grade the differential protection’s operating
time with the transformer fuse protection.
The protection should have additional counter measures to avoid mal-operation during
external faults or during switching on of an external load with large time constants, e.g., a
transformer. The main problem is the transient behaviour of the CT. There are two worst-
case situations:
1. External short circuit with a large fault current and a full DC offset
In this case, if one CT saturates a large differential current will be produced and it can reach
into the tripping zone. Therefore a CT saturation detector is required to avoid mal-operation.
2. External short circuit with small fault currents or additional load current during switching
on a large load.
In both cases it is assumed that there is a full DC offset. The large load can be for example
the inrush current of a transformer, which is switched on. In this case, there is no typical CT
saturation, but a phase shifting is possible, which leads to a small differential current. Since
the differential and the bias currents are in the sensitive point of the characteristic (the low
slope K1 region in figure 3-2), with a sensitive setting a trip is possible. Some manufacturers
design their protection so that it will detect such kind of “quasi saturation” and block the trip.
3.2.2 High impedance differential protection
The high impedance principle is best explained by considering a differential scheme where
one CT is saturated for an external fault, as shown in Figure 3.4.
34
Healthy CT Saturated CT
Zm
Zm
RCT RCT
IF IF
RL RL
RST
VS
I>
IF IF
RL RL
Figure 3.4 Principle of high impedance differential protection
If the relay circuit is considered to be very high impedance, the secondary current produced
by the healthy CT will flow through the saturated CT. If the magnetizing impedance of the
saturated CT is negligible, the maximum voltage Vs across the relay circuit will be equal to
the maximum secondary fault current IF multiplied by the connected impedance, (2RL + RCT),
where RL is the lead resistance and RCT is the current transformer secondary winding
resistance.
The relay can be made stable for this maximum applied voltage by increasing the overall
impedance of the relay circuit, such that the resulting current through the relay is less than
its current setting. As the impedance of the relay input alone is relatively low, a series
connected external resistor is required. The value of this resistor, RST, is calculated by the
formula as shown.
VS
R ST = − RR (3.5)
IS
Where
RR is the relay resistance
IS is the relay setting, and VS = IF (2R L + R CT )
Additional non-linear resistors may be required to limit the peak secondary circuit voltage
during internal fault conditions. The connection arrangement is as shown in Figure 3.5.
35
A
G B
C
NLR RST RST= Stabilizing Resistor

NLR = Non-linear Resistor
R R R
Figure 3.5 Relay connection for high impedance differential protection
The high impedance differential protection has the characteristics of maintaining a constant
sensitivity with varying load current and in being able to use smaller CTs. The latter can
provide advantages over the biased differential protection, especially for retrofit applications.
3.2.3 Self-Balancing differential protection
For small generators where the neutral and line end connections of the phase windings are
separately available, a core balance (or window) type CT can be mounted around the
connections of each phase winding, as shown in Figure 3.6. Under normal conditions, no
current will flow on the CT secondary because of zero flux. Therefore the CT ratio can be
low and there is no need to consider the nominal current of the generator.
I> I> I>
Figure 3.6 Self-balancing differential protection
36
A simple instantaneous overcurrent relay can be connected to the secondary of the CT to
measure the difference in the current entering and leaving the phase winding. Sensitive
protection of phase and ground faults can thus be achieved.
The small CT window limits the size of the conductor and hence the size of the generator
that can be protected. The relay should be of the low burden type to reduce the possibility of
CT saturation for internal faults.
3.3 INTER-TURN (OR TURN-TO-TURN) FAULT PROTECTION
For generators with multi-turn stator windings, there is the possibility of a winding inter-turn
fault occurring. Unless such a fault evolves to become a stator earth fault, it will not
otherwise be detected with the differential protection arrangements as discussed previously.
This is because the two currents on each side of the winding will be the same, with the fault
current circulating locally between the two turns involved.
3.3.1 Overcurrent protection
For generator stators wound with two or more identical windings connected in parallel, with
the neutral points brought out separately, overcurrent protection can be used as shown in
Figure 3.7. In the figure an ordinary differential protection is shown for phase fault detection.
In addition there is a sensitive overcurrent protection that detects inter-turn faults in the
stator winding. Any fault in one of the parallel windings will cause an unbalance current
flowing through the neutral point. This current will be detected by the overcurrent protection
which will trip the generator.
Figure 3.7 Overcurrent protection used to detect inter-turn faults for double wound machines
3.3.2 Biased differential protection
For generator stators wound with two or more identical windings connected in parallel, with
the windings brought out separately, differential protection can be used in a transverse
manner as shown in Figure 3.8. In this type of application a biased system should always be
used as it is not possible to guarantee in advance that exact current sharing between the
windings will take place. A small error in this sharing current would produce instability in an
unbiased system at high levels of through fault current. Any in zone fault, including an inter-
37
turn fault, will result in a circulation of current between the windings producing an output in
the relay operating circuit.
Stator W indings A
B
C
Bias elements
87G 87G 87G Operating

elements
Figure 3.8 Transverse biased differential protection for double wound machines
Another scheme that could be used on this type of generator is shown in Figure 3.9. This
arrangement is an attempt to get the benefits of inter-turn and differential protection with a
saving in CTs and relays. However, this arrangement is not as sensitive as other schemes
using separate inter-turn relays and differential relays. This scheme requires the neutral end
CTs having half the turns ratio of the terminal end CTs. The sensitivity of the protection for
inter-turn faults is limited by the fact that the two CT ratios applied must be selected in
accordance with the generator rated current.
Iph:1
Inter-turn
fault
87G
0.5xIph:1
CTs for other protection functions
NoteL Bias slope of differential protection must be kept low for up to full load current to give interturn
fault sensitivity
Figure 3.9 Generator differential and inter-turn protection
38
3.3.3 Split phase differential protection
One differential scheme using bushing type CTs that is commonly used for inter-turn
protection is shown in Figure 3.10. In this scheme the circuits in each phase of the stator
winding are split into two equal groups and the current of each group are compared. A
difference in these currents indicates an unbalance caused by an inter-turn fault. Since
there is normally some current unbalance between windings, the protection is set so that it
will not respond to this normal unbalance but will pick-up for the unbalance caused by a
single turn fault. In some cases the generator may run with a faulted turn until it is repaired
and therefore the current pick-up level for the faulted phase should be increased to allow
operation but still be able to detect a second fault. A time delay is used to prevent operation
on CT transient error currents that may occur during external faults. The problem of CT
transient error currents can be eliminated by using core balance (window) type CTs as
shown in Figure 3.11.
This method of inter-turn protection will detect phase and some ground faults in the stator
winding. However, because of the slow operating time of this protection it is common
practice to provide standard high speed differential protection for each phase and separate
earth fault protection.
I1
IL
I> IR
I2
I>
I>
Figure 3.10 Split-phase differential protection using separate CTs
39
I>
I>
I>
Figure 3.11 Split-phase differential protection using core balance (window) CTs
Differential Protection
22454 A -F11 (87G-A)
0.748 A
30000/1A CT Setting: 30000A
2⋅ 0.748 A CT Setting: 15000A
777.8 MVA
20 kV
22454 A
Interturn Protection
(Transverse Differential Protection)
-F12 (87GUP-A)
8981 A
0.748 A CT Setting: 30000A
12000/1A
13472 A
0.748 A
18000/1A CT Setting: 30000A
Relay
Redundant Interturn Protection
I> (not restraint)
Figure 3.12 An example of the differential protection for generators with split phase
Figure 3.12 shows the single line diagram of the hydro generator which has grouped the
winding into 3 and 2. The star point is separated for these two groups and finally connected
together. The figure shows the location of the current transformers. On the star point side the
40
CT’s are connected to transverse differential protection. In the case of an Inter-turn fault
(inside of one winding or between the windings) a current unbalance leads to a differential
current. This will be detected by relay F12. Additionally a circulating current flows between
both star points. An overcurrent relay can be used for a redundant Inter-turn fault protection.
The short circuit protection for phase to phase or 3-phase fault is realized with the relay F11.
In that case the star point CT’s are summated and connected to the relay. The other CT is
on the generator main. A classical biased differential protection is realized.
3.3.4 Zero sequence overvoltage protection
Inter-turn faults in a generator with a single winding can be detected by observing the zero
sequence voltage across the machine. Normally, no zero sequence voltage should exist but
a short circuit of one or more turns on one phase will cause the generated e.m.f. to contain
some zero sequence component.
External earth faults will also produce a zero sequence voltage on a directly connected
generator. Most of the voltage will be dropped across the earthing resistor, the drop on the
generator being small and the zero sequence component being limited to one or two
percent. It is preferable, therefore, to measure the voltage drop across the winding, rather
than the zero sequence voltage to earth at the line terminals. This can be done using a
voltage transformer connected to the line side of the generator, with the neutral point of the
primary winding connected to the generator neutral, above the earthing resistor or earthing
transformer. This arrangement is shown in Figure 3.13. The zero sequence voltage can be
measured directly from the voltage transformer broken delta winding. Alternatively, the zero
sequence voltage can be derived from the three phase voltage inputs to the relay.
For stator earth fault protection an independent neutral voltage displacement relay will be
required, as shown in Figure 3.13.
The 3rd harmonic component of the e.m.f. may be larger than the required setting, therefore
the relay must have third harmonic component rejection capability.
With a direct-connected machine, it is still possible that a close up earth fault will produce a
zero sequence voltage drop greater than that produced by the short-circuiting of one turn. It
is therefore necessary to apply a short time delay to the tripping element. With a generator-
transformer unit, an external earth fault cannot draw zero sequence current through the delta
winding of the transformer. Therefore, no residual voltage will be produced from the voltage
transformer and no time delay is required for the trip element.
41
VN>
VN>1
VN>
VN>2
For interturn
Protection
For earth fault
Protection
Figure 3-13 Inter-turn protection and earth fault protection by zero sequence voltage measurement
3.4 PHASE FAULT OVERCURRENT PROTECTION
A simple overcurrent protection can provide fast operation for machine phase faults. This is
applicable for small generators of less than 1MVA, where the differential protection is not
used. It can also provide local back-up protection for larger machines in the event that the
main differential protection fails to operate.
For a generator that operates in parallel with others and forms part of an interconnected
system, an instantaneous overcurrent element can be fed from the CTs connected either to
the generator terminals or on the HV side of the step-up transformer. The current setting
should be set to 120% of the maximum fault rating of the generator, typically 8 x full load
current. The protection will therefore be stable for external faults where the fault current
from the generator will be below the current setting. For faults within the machine, the fault
current will be supplied from the system and will be above the current setting, resulting in
fast clearance of the internal fault. A separate time-delayed element can be set to ensure
that the protected items of plant cannot pass levels of through fault current in excess of their
short-time withstand capabilities.
In the case of a single generator feeding an isolated system, current transformers at the
neutral end of the machine could energize the overcurrent protection to allow response to
winding fault conditions. Relay characteristics should be selected to take into account the
fault current decrement behaviour of the generator, with allowance for the performance of
the excitation system and its field forcing capability. An excitation system that is powered
from an excitation transformer at the generator terminals will exhibit a pronounced fault
current decrement for a terminal fault. To overcome this difficulty, voltage dependent
overcurrent protection will have to be used. This will be discussed in subsequent sections.
42
3.5 OVERALL GENERATOR TRANSFORMER DIFFERENTIAL PROTECTION
The most common form of backup phase fault protection for generator-transformer units is
the overall differential relay. The generator and the step-up transformer are protected by a
single zone of overall differential protection, as shown in Figure 3.14. Since a power
transformer is included within the protection, biased transformer differential with magnetizing
inrush restraint should be applied. Transient overfluxing of the generator transformer may
arise due to an overvoltage following generator load rejection. In such cases, consideration
should be given to applying protection with transient overfluxing restraint or blocking, based
for example on a 5th harmonic differential current threshold.
The current taken by the auxiliary transformer needs to be considered, by arranging the
generator differential protection as a three-ended scheme. The exception might be where
the auxiliary transformer rating is extremely low in relation to the generator rating, e.g., for
some hydro applications.
The third set of the CTs can be located either on the primary side or the secondary side of
the auxiliary transformer. If it is connected to the secondary side, interposing CTs may have
to be used because of the exceptionally high CT ratio required to balance the differential
currents.
The sensitivity of the overall differential protection to phase faults on the auxiliary
transformer would be inadequate because of its relatively low rating. Therefore, it should
have its own differential protection scheme. The scheme is not perfect because for faults at
the HV side of the transformer, the fault current can be extremely high. This could drive the
CTs into saturation, resulting in a failure to trip. The overall differential would provide backup
tripping in this situation.
I3
A
B
C
I1 I2
G B
C
I3(2)
I1(2) I2(2)
IA1 IA2
IB1 IB2
87U
IC1 IC2
Figure 3.14 Overall generator transformer differential protection
3.6 OTHER PHASE BACKUP PROTECTION
Under-impedance protection is an alternative to voltage dependent overcurrent protection for

phase fault backup and is often preferred due to its ease of setting. In order to provide
backup protection for phase faults when the machine is offline, the neutral CTs should be
43
used for this protection. One stage can be set to reach the step-up transformer and another
stage to reach further into power system. It normally needs to be set with a time delay to co-
ordinate with other downstream relays on the system. The first stage detects short circuits in
the generator or the LV-side of the step-up transformer. To coordinate with the differential
protection a short delay is used (close to 0.1s) to allow the differential protection to trip first.
The second zone must be graded with the line protection, both in impedance and in time
settings.
For sufficient impedance grading the relay should handle the measurement via the step-up
transformer (faults on the high voltages sides seen different on the generator side through
the vector groups Yd). A possible application considering the coordination of impedance and
distance relays is shown in Figure 3.15.
ZG ZT
ZL
I > (ZG)
t ZT (backward)
ZG ZL
Z
2. Stage (ZT) ZT
ZG: Generator ZT: Transformer ZL: Line

Transformer Line
Line
Figure 3.15 Scheme with impedance/distance relays and grading
44
4. STATOR EARTH FAULT PROTECTION
4.1 INTRODUCTION
Stator earth faults are the most common faults in generators. This type of fault occurs due to
a breakdown of insulation of the stator winding to earth through the core of the stator.
The great danger from earth faults is the possibility of damage to the laminations of the
stator core and to the stator winding due to the heat generated at the point of the fault.
Depending on the magnitude of the earth fault current, impact on the faulty conductor may
not be a serious matter but the damage to the core very often cannot be ignored. Fused
metal can be cut away, but in case of a severe damage, it may be necessary to rebuild the
core, which means dismantling of the winding.
Protection relays must trip the generator as soon as possible, tripping the main breaker,
disconnecting the excitation field and tripping the prime mover.
However the fault current will not disappear immediately after the breakers tripping, because
of the remaining magnetic field. Depending on the field discharge system, it is possible that
even after tripping, the fault could be supplied for a few seconds by the remnant energy in
the field.
4.1.1 Stator winding earthing system
The applied method of grounding will determine the generator performance during the fault
and also the damage to the stator core.
In a power station in which the generator is directly connected to the network, the methods
used commonly are:
1. Solidly earthed generator by a resistor.

2. Resistor earthed generator.
3. Ungrounded generator in a distribution grounded network.
In a power station with block arrangement, the generator is operating in an isolated network
and the methods are:
1. Direct grounding resistor.

2. Grounding transformer with a secondary resistor.
3. A grounding system on secondary open delta winding of voltage transformers.
In the case of a generator-transformer unit, the generator winding and primary winding of the
transformer can be treated as an isolated system which is less influenced by faults to earth
on the grid.
4.1.2 Effects of stator phase-to-earth faults
A stator phase-to-earth fault produces two effects:
1. Overcurrent on the affected phase
45
2. Overvoltage on the healthy phases.
4.1.2.1 Overcurrent effect
Phase-to-earth short circuit current flows from the damaged phase to ground through the
stator core.
Depending on the amount of dissipated power during the fault (kI2), damages to the core can
be very important, with the possibility of a lengthy and costly process of dismantling and
rebuilding of the stator. Based on tests and experience, there is a relation between the
damage level and the phase-to-ground fault current magnitude and the fault duration (Figure
4.1).
The duration of the phase-to-earth current depends on the discharge system of the rotor
winding. When the field breaker opens the excitation winding circuit, its inductive impedance
induces an overvoltage on the rotor winding, which in turn produces a voltage peak on the
stator winding. This effect must be mitigated, as soon as possible, by the field discharge
system, so as to reduce the voltage transient to a few seconds. The transient stator-to-
ground current will develop for a few seconds, depending on the field discharge system
design, typically less than 5 seconds.
The phase-to-ground fault current magnitude must be limited by the grounding impedance.
Depending upon the continent or country practice, the fault current can be limited from 10-20
A to 200-300 A.
Following figure 4.1, it is recommended that the grounding system be designed so as to

keep damages at a non-destructive level during stator earth faults, maintaining phase-to-
earth fault current below 20 A.
Figure 4.1 – Relationship Phase-to-ground fault current

/Fault duration / Generator damages
4.1.2.2 Overvoltage effect
During a phase-to-earth fault, the voltages are unbalanced. Depending on the type of stator
winding grounding system, phase-to-earth voltages on the healthy phases could get above
the nominal phase-to-earth nominal voltage. Depending on the location of the fault in the
winding, voltages on healthy phases and neutral voltage can be very high. During a phase-
46
to-earth fault at the terminal end of the winding, voltages on healthy phases reach phase-to-
phase nominal voltage level and neutral voltage equals phase-to-earth nominal voltage.
Thus when a phase-to-earth fault occurs on one phase, the resulting overvoltage on the
healthy phases could produce a new fault, creating a double phase-to-earth fault. In that
case, the fault current is higher than in a single phase-to-earth fault, because it is not limited
by the earthing impedance.
As a conclusion, stator winding earthing system must be designed in order to avoid

undesirable overvoltages on phases not affected by the initial fault.
4.2 STATOR EARTH FAULT PROTECTION SCHEMES
Earth fault protection schemes are based on the detection of zero sequence electrical
components, which develop and are always present during phase-to-earth faults.
Although phase-to-earth fault current would be limited to a level smaller than 20A by the
stator grounding system, it is recommended to reduce as much as possible the fault
duration. In doing so, the overvoltage duration on healthy phases will be reduced in order to
limit the risk of double phase-to-ground fault. The tripping time of the earth fault protection
must be lower than 1 second, typically in the range of 500 ms.
It is recommended that the earth fault protection trips the following elements:
1. The main breaker, to eliminate currents from the grid to the fault.
2. Disconnect the field, in order to eliminate the induction of the rotor winding to the
fault.
3. The prime mover at the same time, because during the fault could have been
separated any material from the stator core.
Depending on the generator size and the stator grounding configuration, the earth fault
protection schemes are different. There are two ways for connecting generators to the
network as listed below:
1. Generators directly connected to a distribution system.

2. Generators connected to a network through one step-up transformer.
Following these two categories, earth fault protection schemes depend upon the connecting
method and are described in the next paragraphs.
4.3 GENERATORS DIRECTLY CONNECTED TO THE GRID
4.3.1 Solidly earthed generators
Practically this is never the case because of the high magnitude of earth fault current. This
configuration can be found, however, on small size generators.
In this case, there is not limitation to the phase-to-earth fault current, and its value can
exceed the three-phase fault current. Following the description in paragraph 4.1.2, the risk of
severe damages on the stator core and the phase windings is accepted in this case. The
short circuit current can be taken advantage of to detect the stator earth fault using an
overcurrent relay (Figure 4.2).
47
This protection scheme has the disadvantage that any current transient flowing from the
neutral point to the ground, could produce an undesirable trip of the relay. The neutral
current could be caused by either some inducted voltages or some unbalances in the stator
winding
In case of small size generators with cables as phase conductors, external faults can be
filtered out by an adapted scheme of restricted earth fault protection (Figure 4.3). An
overcurrent relay measures the differential current between the summation of the three
phase currents and the neutral point to ground current. The protection trips only in case of
internal fault.
AB C
CT I>
Figure 4.2 – Stator earth fault Figure 4.3 – Restricted earth fault
overcurrent relay overcurrent relay
In case of generators with current transformers on one or both terminals of every phase,
different differential schemes can be applied (Figure 4.4).
For solidly earthed generators, the earth fault current can be very high with an important
mechanical stress for the generator windings. It is recommended to install any possible
grounding system, enough at least, to limit the ground fault current to the three-phase value.
Figure 4.4 – Earth fault protection in solidly earthed generators. Differential schemes
48
4.3.2 Resistor earthed generator
During the last decades, the development of semiconductors has improved the grounding
resistors designs. The size and cost of the resistors have been reduced for the same
dissipated energy.
Furthermore, using a grounding transformer can reduce the resistor impedance value. The
grounding transformer can be a standard dry single-phase transformer with very low
maintenance costs. In doing so, the money spent even in small stations may be worth
weighing against the possibility of increased damages to the machine.
According to this, generators can be grounded through resistors with two different systems:
1. With a direct (primary) resistor.

2. With a grounding transformer with secondary resistor.
Depending on the limited earth fault current there are two grounding schemes: low and high
impedance earthing systems.
In Europe, most of the machines are earthed by resistors, limiting the earth fault current to 5-
20 A. In U.K. and other continents, it is a practice to use resistors, limiting the fault current to
200-300 A.
According to experience (Figure 4.1), earth fault currents limited to 20 A have not produced
damages to the stator core, even if they were present over an extended period of time.
Currents greater then 20 A can produce severe damages even if they are present during a
short time period. Limitation to 20 A is recommended to reduce stator damages.
Different schemes of earth fault protections are possible on resistor-earthed generators

depending on the fault current limit value and depending upon the location of the available
current transformers (Figure 4.5).
Stator earth faults can be detected by an overcurrent relay installed on neutral point to
ground with a direct resistor (Figure 4.5-a) or in the secondary of a grounding transformer
with a resistor (Figure 4.5-b).
There are schemes of earth fault protection by overvoltage relays with voltage transformer in
a direct grounding resistor (Figure 4.5-c) or in the secondary of a grounding transformer
(Figure 4.5-d).
In all these protection schemes, some time delay is necessary in order to avoid inadvertent
trip of the relay for inducted transients on neutral point to ground. A delay smaller than 1
second, typically in the order of 500 ms is recommended.
49
Figure 4.5- Stator earth fault protection on resistor earthed generators.
In case of generators with a grounding resistor of low impedance value, the earth fault
current limited to 200-300A is high enough to be detected by differential protection (Figure
4.6).
R
GENERATOR GENERATOR
R
STABLISING STABLISING
RESISTOR RESISTOR
HIGH
87 87 87 87 IMPEDANCE
HIGH
RELAY
IMPEDANCE
RELAY
a) High impedance earth fault b) High impedance restricted

(phase) protection earth fault protection
Figure 4.6- Generator earthed with low impedance resistor.

Earth fault protection with scheme of differential relay
4.3.3 Ungrounded generators
When a generator is connected directly to an isolated distribution system, in most

applications the generator is ungrounded. Distribution stations are usually connected with a
cable network for distribution. This means that when an earth fault occurs in this type of
system, a capacitive current component will be present. In this case, the total capacitance
current 3Icc can be measured by current transformers connected for summation or a special
current transformer for earth fault detection.
When in a relay this current is compared to a voltage produced by a voltage transformer the
secondary of which is connected in open delta, a selective tripping can be performed for
50
earth faults in generator (Figure 4.7). The same reasoning can be done for a resistive
current produced by a loading resistor on the open delta winding of the voltage transformer.
Figure 4.7- Directional earth fault protection based on capacitance or resistive current.
If there is a possibility to create an earth fault current by means of a special current source
then different schemes can be used. The scheme will depend on the magnitude of the earth
fault current and the current transformers foreseen in the generator unit.
The most commonly applied practice is to earth the neutral connection of the primary
windings of the voltage transformers. The secondary winding in open delta is used to
energize an alarm device through a voltage relay and give the signal of the existence of an
earth fault. With the same relay during a reasonable time a contactor device will be
energized and a loading resistor at the open delta winding will be inserted to create the earth
fault current (Figure 4.8). The system for earth fault current creation is only in operation for
10 to 20 seconds, but the detection of an earth fault is always in service.
It is common practice to detect an earth fault by a voltage relay on open delta winding from
the voltage transformers (Figure 4.8). But when the voltage transformers are provided on the
terminals of the generator instead of the bus bars, it can be served as earth fault detection
(Figure 4.10).
In practice, however, the voltage relay will not provide exactly the location of the earth fault
in the system.
For this reason and also because the lack of earth current, a double time delay relay can be
applied for tripping first the circuit breaker and later on the generator.
51
Figure 4.8- Stator earth fault protection based on created resistive current.
Figure 4.9- Stator earth fault protection by percentage based differential relay.
Figure 4.10- Non selective stator earth fault protection
52
4.4 GENERATORS CONNECTED TO THE GRID THROUGH A STEP-UP TRANSFORMER
Most generators, of medium and large size, are connected to the grid through a step-up
transformer bank. The unit transformer has the low-voltage side connected to the generator
terminals being a delta and the high-voltage side as a star.
Phase-to-ground faults must be detected in any part of the generation voltage area,
including stator windings and generator bus. A ground fault on any part of the circuit
comprising the low voltage winding of the transformer, connection leads, or generator
winding, will produce an unbalance on phase to ground voltages and an increase of the
voltage from neutral to ground at winding ends.
Stator phase-to-ground protection schemes are designed and set according to the stator
grounding system.
4.4.1 Ungrounded generators
This protection measures the voltage present in the open delta secondary of a voltage
transformer connected to the generator bus (Figure 4.11).
20/√3 kV 110/3 V
R U>
UG = 20 kV
Figure 4.11 – Stator earth fault protection by low earth fault current
This scheme is used as main earth fault protection in the case of small ungrounded
generators. The resistor on the open delta secondary is used as load for grounding all the
low voltage system.
The voltage monitored by the protection is the summation of the three phase voltages. This
summation is the zero sequence voltage, and its value is zero in normal conditions (Figure
4.12).
53
Figure 4.12 – Bus ground protection voltages.
In case of a phase-to-ground fault, the inducted voltage increases to a value determined by

the transformer ratio. Usually the transformer ratio is determined in order to have a standard
voltage of 110 V in the open delta secondary.
4.4.2 High impedance stator earth fault protection
As mentioned before, the stator phase-to-ground protection schemes are designed and set
according to the stator grounding system. Most medium and large generators incorporate a
resistor between the neutral point and ground with two possible arrangements:
1. A direct grounding resistor.

2. A grounding transformer with a secondary resistor.
Following the indications in paragraph 4.1.2.1, this resistor must be calculated in order to
limit the earth fault current in a non-severe level lower than 20A.
There are two main types of stator winding earth fault protection. One covering only the 95%
of the winding down from the generator terminals and the other one, covering 100% of the
winding.
4.4.2.1 Stator winding earth fault 59G-(95%) protection
4.4.2.1.1 Neutral-to-ground overvoltage relay:
When a single stator ground fault occurs, the neutral-to-ground voltage increases from zero
to a maximum of nominal phase-to-ground nominal voltage. One way of detecting phase-to-
ground faults is to monitor the voltage between the neutral-point. (Figure 4.13)
In case the generator neutral is connected to the ground by way of a resistor, stator ground
protection 59G must be connected to a voltage transformer monitoring the voltage across
the resistor, which secondary nominal voltage is usually a standard such as 120 V or 220 V.
In case of a grounding transformer and a secondary resistor, the earth fault protection
monitors the voltage between resistor terminals, but at the lower voltage of the secondary
winding.
The transformer will be phase-to-phase nominal voltage on primary winding, and 120 V or
220 V on the secondary.
54
In both cases, the voltage at the relay is directly proportional to neutral-to-ground voltage
according to the transformer ratio.
The voltage between the neutral point and ground will depend on the position of the fault in
the stator winding. The most severe condition occurs when the fault is at the winding
terminal, the neutral to ground voltage will then be equal to the phase-to ground voltage.
When the fault occurs in the middle of one phase of the stator winding, at the 50% of the
winding, voltage applied between the neutral-point and ground will be a half of the phase-to
ground nominal voltage. When the fault is close to the neutral point, for example at only 1%
from neutral point, the voltage between neutral point and ground will be only 1% of the
phase-to ground voltage.
Figure 4.13- Neutral-to-ground overvoltage relay.
Consequently, the lower the trip setting of the overvoltage relay 59G is, the bigger is the
protected area from the winding terminal to the neutral point.
However, it is not recommended to set this overvoltage function too low because any
neutral-point to ground voltage transient of a few volts could produce an inadvertent
protection trip for several different reasons:
1. Inducted voltages.
2. Minor unbalance in the stator winding.
3. Phase-to-ground short circuit currents on the grid system, through capacitive
coupling between the high-voltage and low-voltage windings of the step-up
transformer.
Setting of the overvoltage relay determines the range of protected generator windings and
depends on the grounding resistor value.
A 5% of the phase-to-ground voltage is the lowest trip setting recommended. In this way,
95% of the stator winding from terminals will be protected.
Tripping time is typically from 0.5 to 1 second. Tripping the main breaker, disconnecting the
field and tripping the prime mover is usually done at the same time.
55
4.4.2.1.2 Neutral-to-ground overcurrent relay:
The short circuit current flows through grounding resistor to ground during a phase-to-ground
fault and this current is limited to 20 A by the resistor.
In this case, the stator winding ground protection is connected to a current transformer with a
standard ratio (i.e. 100/5 or 200/5).
A trip setting in the range of 20% to 40% of the limited fault current of 20A is recommended.
Sometimes this is a fixed setting in the relay. As previously mentioned, it is not
recommended to use a too low setting because any current induction or transient could
produce an inadvertent trip of the relay.
Tripping time is typically from 0.5 to 1 second.
Figure 4.14 – Neutral-to-ground overcurrent relay
4.4.2.2 Stator winding ground 59G-(100%) protection
The ground fault 59G-(95%) scheme described in the previous paragraph, will detect a
phase-to-ground fault in the major part of the generator winding, up to 95% of the stator
winding from terminals. If a phase-to-ground fault occurs in the 5% of phase stator winding
from neutral point, the neutral-to-ground voltage will be lower than trip setting and the fault
will not be tripped.
Under this condition the phase-to-ground voltages on the undamaged phases will be only
slightly increased, and fault current will be very low. Experience shows that any core
damage is likely to be acceptable.
In case of large generators, continued operation under these unbalanced phase-to-ground

voltages can produce permanent unwanted effects such as unbalanced stator currents,
increased vibrations, and a slight increase of voltage in the un-damaged phases.
Thus, in case of large generators (especially with water cooling system), it is required to
detect faults located at any position in the winding. There are two different types of 100%
stator winding protection available:
56
4.4.2.2.1 Third harmonic stator earth fault protection:
This protection is based on the measurement of the third harmonic voltage. This third
harmonic voltage is present in each of the three windings of the generator, it has the same
magnitude and phase angle (zero sequence voltages), Third harmonic voltage will change in
magnitude according to the load of the generator and depending on MW and MVAR
conditions.
Theoretically, third harmonic voltage generation depends upon the mechanical setup of the
stator windings. Its distribution depends upon the capacitive coupling of the generator
windings to earth (Figure 4.15).
Figure 4.15 – Third harmonic voltage behavior.
In normal conditions third harmonic voltage magnitude are equal on terminal and on neutral
point, but opposite in direction. Total third harmonic measured is zero.
In case of earth fault in the neutral point, the third harmonic level presents a different
behaviour depending on the winding position: the neutral point and the generator terminals.
a) In neutral point, the harmonic level decreases, and this effect can be detected by a third
harmonic undervoltage relay (Figure 4.16).
It is based on a third harmonic filtered relay which trips in case of a harmonic level below a
predetermined value.
Before using this protection scheme, it must be checked that third harmonic voltage level in
the generator under any operation conditions is large enough to exceed a minimum trip
level.
As third harmonic voltage reduction occurs between the neutral point and the ground during
a fault near to the neutral point, this relay helps to protect the 100% of the stator windings.
This protection should be an additional relay to the conventional stator earth fault 59G (95%)
protection.
The third harmonic voltage tripping level must be adjusted according to the manufacturer’s
design criteria and depends on the third harmonic voltage level measured between the
neutral point and ground. This option can sometimes provide practical difficulties in setting
this protection.
57
Figure 4.16 – Third harmonic undervoltage relay.
b) In generator terminals, the third harmonic level increases and this effect can be monitored
by an overvoltage relay. The third harmonic voltage is monitored on the open delta
secondary of three voltage transformers connected at the terminals of the generator.
Sometimes, the third harmonic variations can lead to false operation of the protection
because any increase of generator voltage. For that purpose, a compensation scheme is
recommended.
This compensation can be done in two different ways: by measurement of load current or by
the comparison of the third harmonic voltage.
In case of compensation by the load current (Figure 4.17), the protective function is only in
operation with load on the generator. That is the reason why this protection scheme is not
recommended for large generators.
A comparator system is based on a measurement of the differential third harmonic voltage

between neutral point and terminal voltages (Figure 4.18). It is based on a measurement of
the differential voltage between neutral and terminal voltages.
An interruption in voltage of the measurement on the terminals could also lead to false
tripping, because of the lack on third harmonic compensation.
For any case of third harmonic stator earth fault protection, tripping time are usually adjusted
to be between 1 and 5 seconds.
58
Figure 4.17 – Third harmonic load current Figure 4.18 – Third harmonic comparator
compensated Earth fault protection Earth fault protection
4.4.2.2.2 Stator earth fault protection with voltage injection:
Due to an incomplete protection of the stator winding by a high impedance protection and
different practical difficulties in setting of the third harmonic stator ground protection, the
stator ground fault protection based on a sub-harmonic frequency injection was introduced.
It is based on the injection of a pulse voltage wave into the generator winding. The voltage
has a sub-harmonic frequency of 12,5Hz or 20Hz. The voltage injection can be done in
neutral transformer system as well as in the transformers connected to the terminals with a
secondary open delta winding, and is based on an impulse frequency generator independent
from the generator which should be protected. The produced frequency can also be coded
with time interruption, so that more security is available.
The voltage injection system operates with the same sensitivity for all earth faults over the
entire winding and provides a 100% earth fault protection.
There are different technical solutions for earth fault detection.
In some cases, voltage injection is done in a part of the grounding resistor (R1) by an
independent source of frequency. (Figure 4.19)
59
Figure 4.19 – Stator earth fault 59G.
The injected pulse voltage produces a pulse waveform current from windings to ground
through insulation, and returns through the grounding resistor. The relay works by monitoring
the voltage induced in the rest of the grounding resistor (R2).
In normal operation conditions, the pulse current ratio is related to insulation conditions. In
case of phase-to-ground fault, the pulse current increases rapidly through the fault, and the
protection trips.
In other cases, the injected voltage, through a filter system for the generated frequency,
generates a current by an earth fault which can be measured. These current can be
compared to the injected voltage so that, the setting must be higher than the disturbance
current created by all capacitive coupling current without an earth fault (Figure 4.20).
From his dependency on other sources, for creation of his frequency, this protection will be a
reserve protection to the high impedance stator earth fault protection. This way, this
protection is available when the generator is coupled on the network but can also be
available during stand-by and/or start-up situation.
Trip adjustment is usually expressed in ohms, and trip time is typically between 1 and 10
seconds.
60
4.5. REFERENCES
[1] IEEE C37.101-1993 “Guide for Generator Ground Protection”
[2] IEEE C62.92.2-1989. “Guide for the Application of Neutral Grounding in Electrical Utility
Systems. Part II- Grounding of Synchronous Generator Systems”
[3] “Lecture Notes of Generator Protection Relay Course” BBC-March 1975
61
5. ROTOR EARTH FAULTS PROTECTION
5.1 INTRODUCTION
The field circuit of the generator, comprising the rotor winding, the exciter and the field circuit
breaker is a DC circuit that is floating and therefore not connected to the earth. If one point in
the field circuit gets connected to the earth, an earth fault occurs and no fault current will
flow.
If a second earth fault occurs at a separate location in the field system, the current will be
diverted from the intervening turns. The field current from a large machine can be high,
causing serious damage on the rotor and the exciter.
If a large part of the field winding is short-circuited, the resulting flux could create an
attracting force that will be strong on one pole but weak on the opposite one. The outcome is
an unbalanced force, which for large machines may be of the order of 50 to 100 tons. This
will cause violent vibrations, which in turn may damage the bearings or even displace the
rotor to foul the stator.
One single earth fault causes no harm to the rotor. But when one earth fault has occurred,
the risk for a second separate fault rises, since the first earth fault establishes an earth
reference for voltages induced in the field by stator transients. These transients increase the
stress to earth at other points on the field winding. Therefore it is important to get the
protection as sensitive as possible. Typical sensitivity is from 1 to 30 kΩ, but there is also
protection available up to 100 kΩ. It is recommended to use two levels, one for alarming and
one for trip.
In the starting sequence of the generator, the station battery is used to provide the field
circuits for with DC-current for an interval of time of a few seconds. In this situation, the
protection must be removed or blocked during the same interval of time to prevent a false
tripping.
5.2 VOLTAGE DIVIDER METHOD
This method uses a resistor connected in parallel with the field winding. The centre-point of
the resistor is connected to the station earth through a sensitive relay. The voltage across
the relay will be at its maximum if the fault occurs at one of the two winding extremities.
If there is an earth fault in the middle of the field winding, this method is ineffective because
the voltage across the relay will be zero. To avoid a fault in this area to remain undetected, it
is customary to switch between the tapping point in the centre and another one at ¾ end of
the winding. The station personnel can do this manually. It is recommended to do this check
at least once per week.
Another method to avoid an undetected area is to use a voltage dependent resistor or

varistor. At nominal excitation voltage the varistor has a value equal to R2-R1, in this case an
earth fault at the center point of the rotor winding is undetected. When the excitation voltage
varies, the varistor resistance value will vary accordingly, and the relay will detect the earth
fault. When the varistor scheme is used, a push button for manual testing is very often
provided.
62
Varistor
Sensitive
Sensitive R1
relay
relay
Rotor Rotor
Re<
winding Re< winding
R2
Figure 5.1 Voltage Divider Method
5.3 AC INJECTION METHOD
This method uses an external AC voltage source, supplied normally through the secondary
of a voltage transformer, connected to the generator outlets. The voltage is connected to the
field-circuit via coupling capacitors and coupling resistors. The earthing side of this voltage is
connected via an earthing brush to the shaft. Under normal conditions there will be a small
current flowing continuously in the system because of the rotor windings capacitance to
earth (CE), this current is however harmless. This injection method is designed for field
systems up to 1800 volts.
RS CS
Re<
Rotor
winding
CE
RE RS CS
Figure 5.2: AC Injection Method
The AC injection method can be applied using two types of measurements. .
1. Rotor earth current measurement

In that case only the current input is connected to the relay. The sensitivity is
limited by the rotor-to-earth capacitance and by the disturbances from excitation
system. The trip stage is set near 1 to 1.5 kΩ. In current, that is close to 23 mA. A
second stage can be used for alarm.
It’s also possible to connect the additional equipment on one slip ring. Then it
must be connected in parallel.
2. Rotor earth resistance measurement
The relay measures both the voltage and the current and calculates via a replica
the fault resistance. The negative influence of the rotor earth capacitance is
neglected. Only the harmonics from existing in the excitation circuit can affect the
accuracy. The advantage of this principle is a higher sensitivity (max. 30 kΩ). For
practical applications, the trip resistive setting is between 2 and 5 kΩ and the
alarm setting between 10 and 20 kΩ.
63
5.4 DC INJECTION METHOD
The capacitance current associated with the ac injection can be avoided by rectifying the
injection voltage. This principle is shown in figure 5.3. The DC output from the transformer-
rectifier power unit is connected to the positive side of the field winding. This biases the
positive side of the field winding to negative voltage relative to the station earth. Thus an
earth fault anywhere in the field winding will cause a current to flow through the power unit.
The power unit could be replaced by a power supply with a DC voltage source Udc.
This injection method is designed for field systems up to 1200 volts. Typical injection voltage
is 25-50 Volts.
-
Rotor
winding
From secondary
side of VT
Sensitive
Re<
relay
Figure 5.3: DC Injection method
Ufg
x RE
Udc
IRM
Vtrip
RM
Fig. 5.4 Electronic circuit equivalent of the DC injection method
The electronic equivalent of the circuit of figure 5.3 is shown in figure 5.4. Ufg is the exciter
voltage applied to the rotor winding. Udc is the injected DC voltage. The fault occurs at a
distance x (comprised between 0 and 1) from the positive lead. RE is the fault resistance.
The current through the sense resistance RM when a fault occurs on the field winding is
equal to:
U dc + x • U fg
IRM = (5.1)
RM + RE
64
Assume that the relay will trip when the voltage across the resistance RM gets above a
threshold value Utrip. This will occur therefore when:
U trip U dc + x •U fg
≤ (5.2)
RM RM + R E
The fault resistance below which the relay will assert is therefore:
 U dc + x •U fd 
R E ≤ RM •  − 1 (5.3)
 U trip 
In this last equation, Udc, Utrip and RM are constant. One can see therefore that the
sensitivity of the relay (or the maximum fault resistance for which the relay will pick up)
depends upon the distance to the fault and the excitation voltage. The maximum sensitivity
occurs when the fault distance is one.
5.5 DC SWITCHING INJECTION METHOD

(SQUARE-WAVE INJECTION METHOD WITH A LOW FREQUENCY)
A low frequency square-wave generator with a frequency range between 1 to 3 Hz charges

the rotor to earth capacitance CE via a series resistor (RS) as shown in figure 5.5. After the
capacitance has been fully charged, the only DC current that could flow is determined by the
fault resistance RE. If there is no fault, the fault resistance is close to infinity and no current
will flow. With a prior knowledge of the voltage levels of the square-wave generator and the
value of the coupling resistors RS, the fault resistance RE can be estimated by measuring the
voltage across the resistance RM. In order to avoid overreaching due to DC offsets caused
by the excitation voltage, the circuit uses only the difference voltages. The rotor-to-earth
capacitance will always be charged in both directions. If two successive voltage
measurements are the same, the fault resistance estimation is declared valid. This
minimizes the measuring errors if there is a change in the excitation voltage. Through the
use of the low frequency square-wave, the noise margin due to the harmonics of the
excitation system is much higher compared to the AC injection method and therefore the
disturbance influence is smaller. Sensitivity up to 80kΩ (or 100 kΩ) can be reached. The
normal setting for the alarm stage is between 40 and 80kΩ. The trip stage is set
approximately at 5kΩ.
Figure 5.6 gives an indication of the measuring principle. This was recorded during a
commissioning on a 1000 MVA generator with rotating diodes. The first curve shows the
control voltages. This voltage is used for triggering the measurement and calculating the
actual e.m.f.-voltage. The second curve shows the voltage drop via the measuring shunt RM
of figure 5. The change in the square waves indicates different fault resistances that were
simulated by test resistors. On the bottom there are the binary traces with the pickup and the
trip signals. The trip delay was set to 0.5 s.
65
RS
Rotor
winding
Re<
CE
RE RS RM
Uout Umeas Umeas

∆U ∼ 0 ∆U ∼ 1/RE
t t t
τ=RSCE
Figure 5.5 DC Switching injection method
Figure 5.6 Record during a commissioning test (frequency of square-wave voltage is 1.5 Hz
The principle of figure 5.5 can be represented by the electrical circuit of figure 5.7. In this
circuit, where the capacitance between the rotor and the earth is neglected, RE is the value
of the fault resistance between the rotor winding and the earth when a loss of insulation in
the rotor winding occurs. The purpose of the DC switching injection method is therefore to
evaluate RE, which under normal condition should be close to an infinite value.
66
RS
(1-x)Ufg
Ufg
RE RM
x Ufg
RS
Figure 5.7 Equivalent electrical circuit of the DC switching method
In figure 5.7, Ufg is the field voltage and x is the distance to the fault. Let us assume that the
square wave generator injects a signal at frequency Fdc with maximum value UDCp and
minimum value UDCn. Tdc is the signal period. If we measure the voltage across the
measurement resistance RM at the end of two successive half-period (Tdc/2) we would get
by analysis of figure 6 the following voltages:
2 ⋅ RM ⋅ U DCp + RM ⋅ (2 x − 1) ⋅ U fgp
U RMp = (5.4)
RS + 2 ⋅ RE + 2 ⋅ RM
2 ⋅ RM ⋅ U DCn + RM ⋅ (2 x − 1) ⋅ U fgn
U RMn = (5.5)
RS + 2 ⋅ RE + 2 ⋅ RM
where:
URMp : voltage across the measurement resistance at the end of the positive half-period
URMn : voltage across the measurement resistance at the end of the negative half-period
Ufgp : field excitation voltage at the end of the positive half-period
Ufgn : field excitation voltage at the end of the negative half-period
If we assume that the field voltage remains constant and we subtract the two measurements,
we get finally:
2 ⋅ RM ⋅ (U DCp − U DCn )
U RMp − U RMn = (5.6)
RS + 2 ⋅ RE + 2 ⋅ RM
from which we can extract the value of the fault resistance:
RM ⋅ (U DCp − U DCn )  RS 
RE = − + RM 
(U RMp − U RMn )  2 
(5.7)
67
In this last equation, RM, RS, UDCp, UDCn are constant and URMp and URMn are measured
voltages. This last result indicates that the DC switching method allow calculating a fault
resistance that is independent from the fault location and the field voltage.
5.6 FIELD PROTECTION FOR BRUSHLESS GENERATORS
The inaccessibility of the main field circuit makes the direct detection of rotor earth faults
impossible, if earth fault occur this must be detected by measurement of the excitation level.
A severe inter-turn fault or a short-circuited diode could be detected because of an

increasing AC current in the exciter. We must have in mind normal regulations due to the
power system for the setting parameters of this function.
5.7 IMPACT ON DIGITAL RELAYS IN THIS AREA
The latest generation of digital relays are based on the injection principles described in this
chapter.
Because modern digital relays allow making direct measurements of voltages, currents and
resistances at different points of the field system, settings closer to the real physics can be
entered into the relay. The same measurements can be stored in the relay and retrieved with
an event report. Better analysis of faults can be performed then. This helps in reducing the
risk of malfunction and improving the protection reliability.
5.8 ROTOR EARTH FAULT TRIPPING PRACTICES
There exists a variety of practices applied depending upon the power utility with respect to
tripping and alarming applied to rotor earth fault protection.
Because of the difficulty to locate some intermittent earth faults in the field-system that could
be cleared at the next generator start without any reasonable explanation, some utilities may
choose to provide only an alarm.
Another practice is to supervise a vibration detector by a rotor-earth fault relay and to trip the
generator and field breaker when the vibration detector picks up. The purpose of the rotor-
earth fault relay is to sense a first rotor-to-earth fault and the purpose of the vibration
detector is to detect a second rotor-to-earth fault at another location.
Recommendation is made to carefully install the rotor-to-earth protective relay and

particularly to take into account the rotor winding to earth capacity when required by the
relay type. Two detection levels should be set on the relay, one between 20 to 100 kΩ for
alarming and another one between 1 and 5 kΩ for tripping.
68
6. UNDER-EXCITATION (LOSS-OF-FIELD) PROTECTION
6.1 INTRODUCTION
Under-excitation or total loss of excitation can be the consequence of a short circuit or open
circuit in the excitation circuit, mal-operation of the automatic voltage regulator, incorrect
control of generators and transformers, or the case of a generator connected to a system
with capacitive load. Under-excitation corresponds to the condition where the excitation of
the synchronous machine is less than required for stable operation at a particular power
level. This excitation limit determines the steady state stability characteristic of the generator.
If the excitation is not sufficient to provide the power required from the generator then this
stability limit is exceeded. As a result the machine will slip poles and thereby obtain the
required excitation from the connected three phase system.
For example when a loaded synchronous generator loses excitation, it accelerates and runs
above synchronous speed operating as an induction generator. Large induced currents are
then produced in the axial direction of the rotor iron. Operation as an induction generator
causes slip frequency current to flow in the rotor, damper winding, slot wedges and the
surface of the solid rotor body. Dangerous overheating can occur, particularly at the ends of
the rotor where the currents flow across the slots through the wedges and the retaining ring.
Therefore rotor acceleration, local overheating in rotor and stator, over-voltages on the rotor,
mechanical impact on the foundation and power swings could impact the generator itself or
the power system depending upon the construction of the generator, the nature of the
excitation circuit, system conditions, the amount of supplied power as well as the influence of
voltage and power regulators (AGC). Furthermore large reactive power consumption can
trigger area wide voltage collapse. Under-excitation protection is required to detect these
conditions in order to prevent or at least limit the duration of these harmful effects and initiate
tripping the machine off-line.
6.2 THEORY - CAPABILITY DIAGRAM OF GENERATORS
The definition of fundamental electrical quantities is provided here to help the following
discussion:
• The sign convention whereby exported power (P, Q > 0) is positive is used
(generator reference-arrow system).
• The apparent power is described by the symbol S and has the unit of VA (kVA,
MVA).
• The active power has the symbol P and has the unit of W (kW, MW).
• The reactive power has the symbol Q and has the unit of Var (kVar, MVar)
• The generator nominal values UN (voltage) and IN (current)are used for per unit
(p.u.) conversions.
Following the Cartesian coordinate system convention (x-axis = real component and y-axis =
reactive component), the first quadrant defines the operating range (P > 0 and Q > 0) of the
generator. Operation is in the 4th quadrant (P > 0 and Q < 0) in the event of under-excitation.
Graphic representation of the under-excitation protection takes place in the 4th quadrant. The
diagram is often flipped to avoid “neck strain” such that real power is along the y-axis (refer
to figure 6.1).
69
Figure 6.1: Alternative representation of the capability diagram
The stability limit is derived from the equations for the active and reactive power of the
machine. Assume that the generator is connected to an infinite bus; that is, the terminal
voltage of a generator is equal to the network voltage and the equivalent system reactance
(Xe) is zero. Equations (6.1) and (6.2) are the general equations and may be used directly for
the salient pole generator, which has different direct axis and quadrature axis reactance. A
xd - xq
reluctance response circle with the diameter ( U 2 ) results due to the difference in xd
xd xq
and xq. This circle indicates the steady state power that the generator can produce with zero
excitation (E = 0).
EU U2 xd - xq
P =3 sin ϑ + 3 sin2ϑ (6.1)
xd 2 xd xq
EU U2 xd - xq
Q =3 cosϑ − 3 (1 + sin 2ϑ ) (6.2)
xd xd xq
where:
E excitation voltage (phase to star point voltage - field e.m.f.), p.u.*
U terminal voltage (phase to star point voltage) of the generator, p.u. *
xd synchronous direct axis reactance, p.u.
xq synchronous quadrature axis reactance, p.u.
ϑ rotor angle (angular displacement)
Note that if the phase-to-phase voltage representation is used in equations (6.1) to (6.4) then
leave out the factor of 3.
The equations are simplified as the direct axis (xd) and quadrature axis (xq) reactance are
approximately the same for the case of cylindrical generators.
EU
P =3 sin ϑ (6.3)
xd
70
EU U2
Q =3 cosϑ − 3 (6.4)
xd xd
The theoretical stability limit is ϑ equal to 90° for the case of an ideal cylindrical generator.
Accordingly the limiting value in the derived representation is given by the direct axis
reactance xd. This limit is dependent on the reactances xd and xq, as well as excitation and
the terminal voltage for the case of a salient pole generator. The base point for the
theoretical limit on the Q-axis is determined by the quadrature axis reactance xq. The
allowed rotor angle ϑ (theoretical stability limit) is less than 90°. Th ese limits are graphically
visualized in figure 6.2 and 6.3 by means of the voltage and current vectors as well as the
capability diagram for both machine types.
Figure 6.2: Vector and capability diagram of the cylindrical generator with xd = xq
(Eexc excitation voltage; I stator current)
Figure 6.3: Vector and capability diagram of the salient pole generator (xd ≠ xq)
It is apparent that the operating range of the generator is limited as follows:
71
1. Over-excited range - by the power supplied by the turbine and the excitation
(rotor values)
2. Under-excited range - by the power supplied by the turbine, the stator limits or
the stability limit.
The actual (practical) stability limits apply for operation in an interconnected system. These
take into consideration the superseding reactances (for example, unit transformer) which are
always present, and a security margin (10 percent reserve at changing load). The actual
admissible value for stability is therefore smaller than the theoretical value. Figure 1.14 in
Chapter 1 shows the steady-state stability limit of the generator at constant excitation
voltage when the external impedance Xe of the line connecting the generator to the system
is taken into consideration.
The manufacturer of the generator specifies the limits that must be adhered to via their
capability diagram. Various representations and scales of the axes can be found as in figure
6.1. These may have dimensions (MW, MVar) or be dimensionless (p.u.). The latter
representation is preferred. The result is the p.u. representation if the values in a
dimensioned characteristic are divided by the nominal apparent power.
Figures 6.4 and 6.5 provide examples for cylindrical generator and salient pole
generator capability diagrams.
Figure 6.4: Capability diagram of a cylindrical Figure 6.5: capability diagram of a salient pole
generator generator
In the case of the cylindrical generator the stability curve is shifted to the right due to the
superseding reactances when compared to the theoretical curve. On the other hand in the
case of the salient pole generator the theoretical stability characteristic is shifted to the right
as a whole. The intersection with the reactive power axis is approximately at the centre of
the reluctance power circle, as the theoretical intersection (-U2/xq) is instable as a result of
the missing directional force.
Representation in the capability diagram is applicable with nominal voltage and current (UN,
IN). However constant voltage may not always be assumed. The following example
illustrates the influence on the stability limits by varying the voltage. The theoretical stability
limit of the cylindrical generator (refer to figure 6.2) is used to illustrate the influence of a 10
percent change in voltage.
72
If the excitation is equal to 0 then the rotor voltage E is equal to zero as well. According to
equation (6.4) the maximum reactive power that can be imported is as follows:
Q = -U2/xd
U2 0,9 2 0.81
At : U = 0.9 : Q = − =− =−
xd xd xd
(6.5)
U2 1,12 1.21
At : U = 1.1 : Q = − =− =−
xd xd xd
The stability limit is shifted to the right during under-voltage conditions and this further limits
the amount of reactive power that may be imported compared to the value at nominal
voltage. The over-voltage conditions are not critical as the stability limit is shifted to the left in
this case.
The foregoing explanations apply to slow variations of system conditions, A transient

response will occur during sudden changes of load in general or load changes during
underexcited operation. Therefore a dynamic stability limit also exists. To reach a simplified
approximation the transient values (x’d, x’q and E‘) are applied to equations (6.1) and (6.2).
The basic result is shown in figure 6.6. It was assumed that the steady state and transient
quadrature axis reactance is the same for this purpose. It is apparent from the diagram that
the machine may even remain stable in the “dynamic” condition with a rotor angle greater
than 90°. The maximum dynamic stability limit is de termined by the transient direct axis
reactance.
theorectical
dynamic
stability
limit
P [p.u.] stator limit

1
maximum
theorectical turbine limit
steady-state E2exc
stability xd
P =1 rotor
limit
limit
Q [p.u.]
U2
xd
U2
xq
U2
x’d
Figure 6.6: Dynamic stability limit (salient pole generator)
The statements regarding the limits during under-excitation are summarized in figure 6.7.
They are as follows:
• The practical (steady state) stability limit is to the right of the theoretical value and
is given by the capability diagram of the generator. It applies at nominal voltage.
• If the generator is operated with a voltage U less than UN the limit is shifted to the
right.
73
• A dynamic stability limit is introduced to consider “dynamic conditions”.. If it is
exceeded the machine must be disconnected from the system immediately as
pole slip will most likely take place.
Figure 6.7: Summary of the statements regarding stability limits
The stability limits (steady-state and dynamic) for both the salient pole and cylindrical
generator - are calculated below in an example using equations (6.1) and (6.2).
The following per unit data are used:
Salient pole generator:

Generator terminal voltage U =1,0
Rotor voltage E = in steps (0; 0.25; 0.5; 0.75; 1.0; 1.68)
Reactances: xd = 1.0; xq = 0.6; x’d = 0.3
Safety margin for the practical stability curve is 9 percent from theoretical stability limit.
Figure 6.8 shows the calculation results. The “half” circles represent the P= f (Q) for different
angles (0° to 180°) and the selected rotor voltage E. The stability limt is reached at
maximum active power (dP/d ϑ = 0). The practical steady-state stability is calculated with a
safety margin of 9 percent. The same calculation was done with the transient direct axis
reactance for the dynamic stability limit.
Cylindrical generator:
Generator terminal voltage U =1.0
Rotor voltage E = in steps (0, 0.4; 0.8; 1; 1.5; 2.51)\
Reactances: xd = 1.919; xq = 1.885 x’d = 0.35
Figure 6.9 shows the calculation results for the cylindrical generator.
74
Dynamic Theoretical Practical
stability limit steady-state steady-state
stability limit stability limit
2
2.0
P[p.u]
P1
i
P3 1.5
i
P5 E=1,68
i
P6
i
1
P dy1 E=1,0
i E’=0
P dy2
i
E=0,5
Sthy 0.5
a
Spry
a
Stdy
a
E=0 Q[p.u]
3.5 3 2.5 2 1.5 1 0.5 0 0.5 1
cosφ 1/xd
E’=0,25
1/xq
0.5
1/x’d 0.5
3.5 Q 1 , Q 3 , Q 5 , Q 6 , Q dy 1 , Q dy 2 , Sthx , Sprx , Stdyx , Q N 1

i i i i i i a a a
Figure 6.8: Calculated stability limit of a salient pole generator
2
2.0
Dynamic Theoretical P[p.u]

stability limit steady-state
Practical
stability limit
steady-state
1.5 stability limit
P1
i
P3
i E=2,51
P6
E’=0
i
1
P dy 1
i
P dy 2
i
Sthy
a
0.5
Spry
a
Stdy E=0,8
a
cosφ
3 2.5 2 1.5 1 0.5 0 0.5 1
E’=0,4
1/xd Q[p.u]
1/x’d
0.5 0.5
Figure 6.9: Calculated stability limit of a cylindrical generator
The loss of synchronism by salient pole and cylindrical generators is illustrated in figures
6.10 and 6.11. The increase of the rotor angle following loss of excitation can be seen in
both diagrams. The imported reactive power increases and the steady state stability limit is
exceeded due to constant turbine power (that is, the real power does not change). An
additional flux appears in the excitation circuit or an additional induced rotor voltage appears
which attempts to maintain the armature reaction of the machine at a constant level as a
result of the slip. This is apparent from figure 6.10b where in the range between 90° and
180° the excitation current increases significantly again. Only shortly before 180° is the rotor
accelerated towards the stator pole (zero load state with inverted pole). This large
acceleration causes large slip and thus increased influence by the damper windings. The
flux changes and therefore the (no longer measurable) rotor voltage becomes equal to zero
75
when ϑ equals 180°. The zero crossing of the excitation c urrent and the measured rotor
voltage is delayed as this takes place in an inductive circuit. The rotor is now decelerated
until it almost reaches synchronous speed as the synchronizing torque becomes very large
shortly after ϑ equal to 180°. The result is a strong torque impul se that is also noticeable as
a significant real power impulse (refer to 6.8a). The mechanical power driving the machine is
however too large to allow a recovery and the machine will continue slipping. This sequence
is repeated between ϑ equal to 180° and 360° and also between n •180° and (n+1) •180°.
Some deviations apply during the transient state before the steady state slip condition is
reached. The reactive power minimum following the first torque impulse therefore has a
different value compared to the following swing cycles. The swing and transient conditions
are particularly severe in the salient pole machine due to the difference in direct axis and
quadrature axis reactance. The slip changes dramatically during one cycle.
The cylindrical generator with xd ≈ xq on the contrary has relatively mild transient conditions,
at least for small real power, resulting from the slip oscillation as can be seen in figure 6.11.
Accordingly the excitation current also has a more even course.
a) b)
Figure 6.10: Loss of synchronism on a 30-MVA salient pole generator (caused by rapid loss
of excitation) with P= 0.8, ie =1.3
a) course of power flow
b) excitation signals (ie =excitation current, ie0= zero load nominal excitation current; ve =
excitation voltage, ve0 = zero load excitation nominal voltage
76
a) b)
Figure 6.11: Loss of synchronism of a 30-MVA cylindrical generator (caused by rapid

loss of excitation) with P= 0, 19, ie =1, 2 ie0
a) course of power flow
b) excitation signals (ie =excitation current, ie0= zero load nominal excitation current;
ve = excitation voltage, ve0 = zero load excitation nominal voltage, ie0 = zero load
excitation nominal current
6.3 PROTECTION METHODS
6.3.1 Field Current Measurement
Figure 6.12: Measurement of field current
77
The application of loss of excitation protection is not standard on smaller machines. If
applied on smaller machines in some cases it is designed to trip and others merely to give
an alarm.
The loss of excitation condition is detected by monitoring the field current and operates when
this signal drops below a pre-set value. The relay must have a setting below the minimum
exciting current.
Finally this method has limitations and is not recommended for medium-sized and larger
generators as it can only operate if there are failures in the excitation circuit (interruption or
short circuits).
6.3.2 Impedance Measurement

6.3.2.1 Scheme 1 - Offset Mho Circle Method (Negative Offset)
The impedance measurement is a widely applied measurement principle. This technique is

used widely and simple to set using impedance measuring elements in electro-mechanical
relays or numerical relays.
The generator capability diagram must be transformed (mapped) into the impedance plane
to obtain the setting parameters for the impedance principle. Use equations (6.6) and (6.7) to
convert complex power to impedance.
U
Z= Z = R + jX (6.6)
I
U ⋅ U* U 2 U2 ⋅ S U2 ⋅ P U2 ⋅ Q
Z= = = = + j (6.7)
I ⋅ U* S* S* ⋅ S P2 + Q2 P2 + Q2
This transformation mathematically corresponds to the mathematical inversion of a locus.

Consequently the direct reference to the capability diagram is lost. In accordance with the
locus theory - a line that just misses the origin becomes a circle that touches the origin when
inverted - the transformation of the theoretical stability characteristic to the impedance plane
results in the section of a circle. This transformation is illustrated in Figure 6.13. All points to
the left of the stability limit in the capability diagram are located inside the semi-circle
(shaded area) in the impedance plane.
Figure 6.13: Transformation to the impedance plane
78
The protection function uses the terminal voltages and currents of the machine to measure
the impedance as viewed from the machine terminals and asserts when the impedance falls
inside the circular characteristic (shaded area).
The protection function is offset from the origin by one half of the direct transient axis
reactance (x’d/2) to prevent misoperation during system disturbances and other fault
conditions. The diameter of the circle is adjusted to be equal to the direct axis synchronous
reactance (see figure 6.14). Characteristic 1 in figure 6.14 supervises the steady state
stability limit of the machine and tripping is normally delayed with a time from 0.5 to 1.5
seconds. This is active for low load operation of the generator and partial field voltage
failure. Characteristic 2 with a diameter of 1.0 pu on the machine base to trip faster with a
time delay less than 0.3 seconds and sometimes with no time delay. Characteristic 2 is
adapted to the dynamic stability limit and responds from full load down to 30 percent load.
Figure 6.14: Two offset mho protection characteristic
The transformation of the offset mho characteristic back into the generator p.u.-capability
diagram is shown in figure 6.15 - the adaptation on both stability limits is clear. Also see
figure 6.7 for a visual explanation.
Figure 6.15: Transformation of offset mho characteristic back into the p.u. - capability diagram
79
6.3.2.2 Scheme 2 - Impedance Circle (Positive Offset) with Directional Element
The impedance measurement function supervises two circles in this protection application
(see figure 6.16). Characteristic 2 has a positive offset Xe (or Xe - system impedance, see
chapter 1.10) and is controlled by a directional element to avoid unwanted operation during
close-in system faults. An alarm is immediately given for operating personnel. Normally a trip
occurs after a 1 minute delay . Additionally an undervoltage function is used for accelerated
tripping with a shorter time delay (0.25s to 1s). Use the shorter time if there is no
characteristic 1. The longer time delay is used when both characteristics are applied. The
undervoltage element is typically set from 0.80 p.u. to 0.90 p.u. of the machine rating. Figure
6.17 gives an overview of the logic.
A modification of the undervoltage element is applied for hydro generators. If the generator
is in a synchronous condenser application then the machine is operated underexcited and
draws reactive power from the system. The voltage on the generator terminals becomes
lower than the system voltage (90 to 95 percent of the rated voltage) and can cause the
undervoltage element to operate. In such an application the undervoltage pickup blocks the
trip and a drop-out permits the trip.
Characteristic 2 is coordinated with the steady state stability limit of the generator and the
settings of characteristic 1 are similar to characteristic 1 of figure 6.14. The negative offset is
x’d/ 2.
Figure 6.16: Two impedance characteristics plus directional element
80
Figure 6.17: Logic of scheme 2
6.3.3 Admittance Measurement
The transformation of the capability diagram into the admittance plane has the distinct
advantage that when using the p.u. representation a direct reference to the generator
capability diagram is provided that is independent of the actual generator voltage. The
transformation is explained as follows. Equations (6.8) and (6.9) describe the fundamental
definition of the complex power and the complex admittance.
S = U ∗ I* S = P + jQ (6.8)
I
Y= Y =G + jB (6.9)
U
with: Y admittance
G conductance (real component of the admittance)
B susceptance (reactive component of the admittance)
The relationship for the transformation can be derived by multiplying equation (6.8) with the
conjugated complex voltage.
I ⋅ U* S* P − jQ P Q
Y= = = = −j (6.10)
U⋅U *
U 2
U 2
U 2
U2
Comparing the coefficient in equations (6.9) and (6.10) leads to the definition of the
admittance values.
P Q
G= B=− (6.11)
U2 U2
The values from the axis in the generator capability diagram must simply be divided by the
square of the voltage. If subsequently the sign of the reactive component is inverted then the
transformation is complete. The per unit numerical values in the capability diagram are
identical with those in the admittance diagram (refer to figure 6.18) when:
81
U = UN = 1
Directly derive the setting values of the under-excitation protection from the per unit
capability diagram.
Figure 6.18: Capability diagram of the generator and admittance diagram
The protection measuring algorithms are based on equations (6.8) and (6.11). The vector
signals are derived from the sampled instantaneous values in the three phase to ground
voltages and three phase currents. The positive-sequence components are calculated from
these vectors. According to the definition in equation (6.8), the positive-sequence voltage
and current components are used to calculate the active and reactive power. Division by the
positive-sequence voltage U1 according to equation (6.11), results in the transformation from
the power plane into the admittance plane.
Use figure 6.7 to extract the characteristics required for these protection functions. The given
steady state stability limit must be monitored. Generally two lines are sufficient for this
purpose. A further threshold value which depends on the dynamic stability limit applies as
well. It is highly unlikely that the machine will recover to the stable operating range from the
area of extreme under-excitation (that is, on the left of characteristic 3). Therefore fast
tripping is required for this particular case. This is different than if the steady state stability
limit is exceeded (characteristics 1 and 2 in figure 6.19). In this case if the excitation voltage
is still sufficiently large then a recovery by the machine to the stable operating range is
possible. Monitoring the excitation voltage (Uexc <) is therefore introduced as an additional
criterion. This criterion controls the tripping time of characteristics 1 and 2. Over functions as
result of transient transgression of the steady state stability limit due to dynamic impulses
that are followed by recovery to a stable condition are thus prevented.
The setting parameters that consist of the intersection with the B-axis as base point of the
line and the inclination are applied to set these characteristics. The angle of inclination may
be derived from the generator capability diagram. The setting should be such that it is close
to the given stability characteristic. The excitation circuit controller characteristic supersedes
this condition.
82
Figure 6.19: Characteristics of the admittance protection
The protection response resulting from the characteristics in figure 6.19 is as follows:
a) characteristics 1 and 2 are exceeded and the excitation voltage monitoring (Uexc <) is
not picked up.
First alarm for this condition must be alarmed and then trip following a long time delay
(approximately after 10 seconds) if it persists.
b) characteristics 1 and 2 are exceeded and the excitation voltage monitoring (Uexc <) is
picked up.
Trip with a short delay (approximately 0.5 to 1.5 seconds) for this condition.
c) characteristic 3 is exceeded.
Trip quickly with less than 0.3 seconds or no time delay at all.
Two basic controls are normally implemented on a generator in order to satisfy these
requirements - the load-speed turbine governor control and the controls associated with the
excitation system.
6.3.4 Measurement of the Reactive Power
Some applications use underexcitation protection to supervise the level of negative reactive
power. This is popular for synchronous motors. The characteristic is a straight line and is
constant for various levels of active power. Use two characteristics to supervise both the
steady state and dynamic stability limits. Figure 6.20 shows an example. The disadvantage
is this measurement is not very sensitive during low voltage operation of the generator.
Improvements are possible:
83
• Inclination of the characteristic
• Conditioning of the characteristic in the case of undervoltage (undervoltage

restraint)
Figure 6.20: Characteristics of reactive power measurements
6.3.5 Directional Current Principle
When generator losses its excitation it starts to draw reactive power from the connected
power system. Such excessive reactive power flow into the generator can be detected by
using a directional current relay. The directional relay measures the generator current
towards the power system and it is set in accordance with the generator capability diagram
as shown in Figure 6.21 (e.g. relay current pickup is arranged to coincide with the line CD on
the machine capability diagram at rated voltage). The directional relay characteristic shall be
practically independent of the magnitude of the polarizing voltage (e.g. directional relay is
only blocked once the polarizing voltage magnitude fails below 5% of the rated voltage).
With numerical technology it is easily possible to use the positive sequence current and the
positive sequence voltage for this directional measurement. To prevent directional relay
operation during power swing conditions a definite time delay of 2s is typically used.
Alternatively inverse time characteristic can be utilized.
84
Figure 6.21: Characteristics of directional overcurrent relay in P-Q and U-I planes
Operation of the directional element only initiates an under-excitation alarm. In order to get
the loss-of excitation trip signal a simultaneous pickup of either an overcurrent element
(typically set at 115% of machine rated current) or an under-voltage element (typically set at
85% of the machine rated voltage) is also required. Overall scheme tripping logic is shown in
figure 6.22.
Figure 6.22: Alarming and tripping logic
85
6.4 IMPACT OF DIGITAL RELAYS ON UNDER-EXCITATION PROTECTION
The loss-of-field function in the presently available numerical relays is based on the
principles mentioned before. The most commonly applied principles are the impedance
measuring principle according to figure 6.14 (scheme 1- offset mho circle) and admittance
measuring principle (figure 6.19).
The advantages of digital technology include a high accuracy and flexible adaptation of the
of the protection function for specific generator conditions via the setting parameters. In the
case of an under-excitation condition all the necessary information is stored (event and
oscillographic recording, measurement) in the protection device. Analysis of the event report
allows making a better understanding of the faults and helps a deeper study of the
phenomenon. This allows the relay settings to be fine tuned which reduces the risk of
misoperation and the protection function becomes more secure.
6.5 REFERENCES
[1] IEEE tutorial on the protection of synchronous generators. 1995, IEEE catalog number:
95 TP102.
[2] IEEE Guide for AC generator Protection, IEEE Std C37.102-2006.
[3] “Numerical generator protection REG 316”, Manual of company ABB.
[4] “Multifunction generator, motor and transformer protection relay 7UM62”, Manual of
company Siemens.
[5] Herrmann, H-J.: Increased Sensitivity of Underexcitation Protection Based on

Admittance Measurement. Technical conference of “Politechinika Slaska” in Ustron
Poland, 19 – 20 May 2005.
[6] Herrmann, H-J.: Underexcitation Protection based on Admittance – Excellent Adaptation

on Generator Capability Curves. 1st International Conferences on Hydropower
Technology & Key Equipment 2006, Beijing 28- 29 Oct. 2006.
[7] Reimert, D.: Protective Relaying for Power Generation Systems. Taylor & Francis, 2006.
86
7. ABNORMAL FREQUENCY OPERATION
7.1 INTRODUCTION
When a power system is in stable operating condition, all generating units will run at a speed
corresponding to the rated frequency, 50 or 60 Hertz. The total mechanical power input from
the prime movers to the generators equals the sum of all connected loads plus the real
power losses in the transport system. Abnormal frequency operation occurs when this
condition of balance is lost.
Underfrequency condition on a power system will be the result of the sudden reduction of
the input power as would occur with the loss of one or more generators or an intertie
importing power. Because of the overload condition, the remaining generators will undergo a
reduction of their speed with the result that the system frequency will decline.
Overfrequency condition will most likely be the consequence of a sudden loss of load or
intertie exporting power. The additional energy available from the prime mover outputs will
be converted to accelerating power by the generating units with the associated increase of
the system frequency.
Abnormal frequency operation of a power generating unit can be detrimental to the key
major pieces of equipment such as the turbine, generator, main transformer, and auxiliary
equipment. In addition abnormal frequency operation can be detrimental to system operation
if proper actions are not taken. This can lead to generator trips, tie-lines to open due to
overload, or parts of the system can become islanded due to power swings that result in
instability. Figure 7.1 represents the various conditions of operation and consequences as
the frequency departs from rated.
Both generator and turbine have limited operation capability outside the rated frequency.
The operation capability is more restrictive at speed below rated. At reduced speed, there
will be reduction in the output capability of the generator. Turbines, particularly steam and
gas turbines, are considered to be more limited than the generator during underfrequency
operation because of the potential mechanical resonances of the turbine blades. Departure
from rated speed under overload will bring stimulus frequencies closer to one or more of the
natural resonant frequencies of the various blades. As vibratory stresses increase, damage
is accumulated, which may lead to cracking of some parts of the blade structure.
Underfrequency protection for turbine depends upon a utility philosophy. If turbine safety is
the prime concern, a primary underfrequency protection scheme will be applied to the
turbine. Alternatively, the implementation of a load-shedding scheme could be designed to
be the turbine primary underfrequency protection scheme. The load-shedding program
should be designed to shed enough load under various system conditions so that system
frequency will be restored close to normal as soon as possible. Back up turbine
underfrequency protection could supplement the load shedding automatic program and it
goes without saying that the two schemes should be designed so that they will coordinate
properly.
87
Frequency
Generator tripping
Activation of other emergency power actions (HVDC support, etc.)
fn: Nominal Frequency Frequency band during normal operation
Activation of spinning disturbance power reserves
Activation of other emergency power actions (HVDC support, etc.)
Frequency controlled load shedding
Generator tripping
Dead System
Fig.7.1 Consequences for system operation with frequency deviation from rated
7.2 GENERATOR AND TURBINE CAPABILITY AND PROTECTION DESIGN CRITERIA
It should be borne in mind that much of the information concerning the operation of turbines
and generators at off-nominal frequency often do not belong to the public domain and may
vary from one manufacturer to another. In view of that situation, it is recommended that the
relevant information for a particular equipment or installation regarding its frequency
operation capabilities should be obtained from the manufacturer. It is now customary
practice for the equipment manufacturer to give the final approval of the frequency protection
scheme for a turbine or generator.
7.3 CONFORMANCE TO IEC 60034:1996
Some turbine generators are designed to accommodate the IEC 60034-3:1996 frequency-
voltage characteristics. IEC 60034-3:1996 requires generators to deliver continuously rated
output at the rated power factor over the ranges of ±5% in voltage and ±2% in frequency, as
shown by the shaded area in 2.
IEC 60034-3:1996 recommends that operation outside the shaded area “be limited in extent,
duration and frequency of occurrence.” A manufacturer could, therefore, impose severe time
restrictions for the generator itself, particularly for operation below 95% or above 103% of
rated frequency (respectively 57 or 61.8 Hz on a 60 Hz basis) and, to a lesser extent, for
operation outside of the continuous range of 98% to 102% of rated frequency.
In view of these considerations, a manufacturer may require, for the generator only,
frequency operational limits in the form of time-frequency characteristics. In such situations,
the principal goal of frequency protection schemes is to return the frequency to the
continuous IEC operating frequency range (98% to 102% of rated frequency) as soon as
possible and to minimize operation outside of this range, both in extent and in duration, and
in concert with load-shedding practices.
88
V%
106
104
102
94 96 98 100 102 104 f%
98
96
94
Figure 7.2 Operation over-ranges of voltage and frequency (IEC 60034-3:1996 Figure1)
7.4 LOAD SHEDDING SCHEMES
7.4.1 Purpose of underfrequency load shedding schemes
Load shedding schemes are used to prevent a significant unbalance between generation
and load in a power network from building into major events and restore the system
frequency to near normal. These practices should be designed so that during maximum
possible overload, the load shedding can quickly restore the system to near normal
frequency following a disturbance. Utilities have instituted some type of automated load
shedding to prevent total system collapse and attempt to minimize the possibility of
equipment damage.
The steps in a typical load shedding program are:
1. Shed sufficient load to minimize the overloads on the connected generators.

2. Eliminate the possibility of a cascading event that causes the unit(s)
underfrequency protection to trip the unit(s).
3. Restore the system frequency as soon as possible.
4. Prevent damage to the turbine generators and other plant equipment.
7.4.2 Coordination of system and turbine-generator underfrequency protection
Load shedding programs and underfrequency protection should coordinate in the following
manner. If the load shedding scheme fails to return the system to an acceptable frequency of
operation, then and only then should the underfrequency protection take action. Traditionally
applying a combination of frequency relays in series with timers has provided this protection.
A first level detection might only alarm with later stages of the protection scheme initiating a
unit trip.
89
Table 1 shows a typical set of underfrequency settings for a 50/60 Hz machines.
Table 1 Typical underfrequency settings
Frequency Band (Hz) Time Delay Comments

50.0 – 49.5 / 60.0 – - No relay action – the
59.5 turbine can operate
continuously.
49.5 / 59.5 None Frequency recorder
alarms.
49.5 – 48.5 / 59.9 – - Load must be shed
58.5 within 30 minutes,
otherwise isolate the unit
48.5 – 47.0 / 58.5 – 5.0 Minutes
57.0
47.0 / 57.0 and below 1.0 s
7.5 STEAM GENERATION PLANTS
7.5.1 Generator limits
During reduced frequency operation of the generator there will be a subsequent required
reduction in KVA output. This is due to the reduced cooling capability of the unit and the
resulting higher temperature. Due to the deficit in generation during this event there will most
likely be a high value of load current required which could result in exceeding the short-time
thermal capability of the generator. The limits on the generator are normally less restrictive
than those that are on the turbine. The thermal overload aspect of this event is protected
against by use of overcurrent and temperature protection (RTDs and alarms).
An overfrequency condition occurs when a significant reduction in load (load rejection)

results. If rated KVA and voltage are not exceeded and operation is maintained within the
turbine limits, generator overheating should not occur.
7.5.2 Turbine limits
With regard to the turbine limits the concern is the large last stage blades operating at other
than the normal frequency of the system. Operation of this stage at load at a low speed can
cause the blading to resonate at a natural frequency. This will lead to blading fatigue
damage and eventual failure. If significant negative sequence current occurs, an exciting
torsional frequency blade resonance is likely to cause severe problems.
Governor runback controls and/or operator actions are generally depended upon to provide
overfrequency protection and correct the overspeed. This is in lieu of overfrequency
protection. However, thought should be given to how a unit overspeed protection is impacted
during the isolation of the unit such that could occur during an overfrequency condition. The
turbine blading must be protected for an overfrequency condition, since there are operating
limits for this event as well. Figure 3 shows a typical 60 hertz steam turbine partial or full-
load operating limitations during abnormal frequency.
90
Figure7.3 Steam turbine partial- or full-load operating limitations during abnormal frequency [1]
7.5.3 Turbine underfrequency protection
In many applications, system load shedding programs provide the primary underfrequency
protection for turbine-generators. When a load shedding program is applied, underfrequency
protection is used to back up the load shedding. As already noted in the introduction, if the
primary concern is the safety of the turbine, the underfrequency turbine protection will be the
main abnormal frequency protection of the plant.
The first step in designing an underfrequency protection scheme is determining the turbine’s
abnormal frequency operating characteristic. As recommended in paragraph 7.2,
consultation with the manufacturer should provide the initial design parameters. Once the
number of frequency steps is known, the time delay for each step must be determined.
Some margin should be included in the time delays and should be discussed with the
turbine’s manufacturer. The time-delay margins should consider the importance of the unit,
the susceptibility of the system to an underfrequency event, and the operating agreements
with local or regional power authorities. This allows for application of underfrequency
protection, even if the unit has been in operation for many years without having accumulated
previous underfrequency operational data.
The following stepwise method is suggested for developing the scheme.
1. Use the manufacturers frequency limits to determine time delays and trip set
points
2. Introduce some margin in the time-delays.
3. Coordinate with the load shedding program.
4. There should be no single point failure that would cause lack of security (tripping
needlessly).
5. A single point failure should not jeopardize the reliability of the protection.
6. Quality protection should be selected (accuracy, speed of operation and speed of
reset).
91
7. The protection should be in service whenever the unit is synchronized or
supplying the auxiliaries off line.
8. Staged alarms should be used to alert operators prior to tripping the unit during
an underfrequency event.
Figure7.4 demonstrates an underfrequency relay setting. The example indicates the turbine
is capable of continuous operation at frequencies above 58.5 hertz and is limited to a
maximum of 10 minutes accumulated over the blading life at 56.0 hertz. These are operating
conditions with the turbine at load. An alarm should be provided when the underfrequency
relay begins to time out, providing operating personnel a warning of the impending
underfrequency trip.
Frequency (Hz)
Figure 7.4. Example of turbine abnormal frequency limitations and settings
If IEC 60034-3:1996 is applicable, as described in paragraph 7.3, time-frequency protection

could be required by a manufacturer for the turbine generator. In such situations, the more
restrictive of the generator and turbine frequency requirements should be used to determine
the appropriate frequency settings for underfrequency relaying. The frequency capability of
the generator, when applicable, must be considered in the design of the frequency relay
protection scheme.
7.5.4 Plant Auxiliary Considerations
Adequate steam supply to the unit for acceptable operation is predicated on the margin in
capacity of the auxiliary equipment. In most cases the limiting equipment is the boiler feed
pumps, circulating water pumps, and condensate pumps. The reason being that each
percent of speed reduction causes a greater reduction in application output. Therefore the
critical frequency limit will vary from plant to plant.
Typical protection for this equipment against underfrequency operation is provided with
thermal protection, or a frequency-sensitive relay or volt per hertz protection.
7.6 COMBUSTION TURBINE UNITS
Limitations for combustion turbine-generators are similar in most respects to the steam units.
However some differences do exist. In general the combustion units have greater capability
92
than steam units to tolerate underfrequency operation, especially if the control system
includes a load runback feature.
Also a combustion turbine may lose air flow if full output is attempted to be maintained
during an underfrequency event. Loss of air flow would eventually cause the unit to trip on
blade over-temperature. Combustion turbines are equipped with a control system that
automatically unloads the unit by reducing fuel flow as speed decreases. This protects both
the turbine blade and the generator from overheating during underfrequency operation.
The following steps should be used to protect the combustion units:
1. A single underfrequency relay per unit connected to the unit voltage transformer.
2. A second relay can be used to supervise tripping for security. It may be common
to several units.
3. User should be aware of supplied protection in the manufacturers control system.
Coordination may be required to avoid conflict.
Figure 7.5 shows a typical example of the combustion turbines under/overfrequency limits
and associated times.
Figure 7.5. Typical combustion turbine operational limits
7.7 COMBINED CYCLE UNITS
A combined cycle unit is a combination of a combustion and a steam unit. The associated
underfrequency limits should set per the recommendations for each type of unit. The
protection should be implemented separately.
7.8 HYDRO UNITS
In general hydro units can tolerate a much wider frequency range than the other types of
generating units. Typically, underfrequency protection is not required for turbine protection.
Pressure in the penstock is generally the limiting parameter.
Overfrequency protection may be applied to hydro units to backup overspeed devices.

These relays are set at a frequency that is lower than the maximum occurring during a load
rejection, but with enough time delay to ride through normal governor action. The
overfrequency protection will act if the governor fails to operate in an appropriate amount of
time.
93
7.9 NUCLEAR GENERATING UNITS
In general the issues discussed in the previous section apply to nuclear generating units. As
discussed previously the output capability of the cooling pumps is reduced due to a
reduction in frequency. Similar to the steam plant discussions the coolant pump flows in the
nuclear steam system will be reduced. This reduced flow can cause detrimental impacts on
plant equipment.
The two types of nuclear plant designs are discussed separately because their responses to
abnormal frequency operation are different.
7.9.1 Pressurized water reactor (PWR) plants
In the case of the PWR plant, abnormal frequency operation impacts the reactor coolant
pump speed. These pumps' output varies as a function of speed. If a significant
underfrequency event occurs, the reactor will trip when reduced coolant flow through the
core occurs. Safe operation is of utmost importance for the nuclear unit, therefore
accommodations for underfrequency must be made in the protection scheme. If the coast-
down rate of the pump is exceeded by the frequency decay, underfrequency relays must be
implemented to trip the reactor and generator at a frequency level that will allow the isolated
reactor coolant pump to meet its coast-down operational performance requirements. The
following issues should be investigated when implementing an underfrequency scheme:
1. Calculated departure from nuclear boiling ratio of the plant.

2. Sizing of the coolant system versus the requirements of the reactor core.
3. Rate of maximum frequency decay that could occur.
4. Rating of the core with respect to its loading.
5. Power system load shedding scheme should be coordinated with.
6. Pre-event system voltage conditions before decline.
7.9.2 Boiling water reactors (BWR) plants
Some BWR units employ non-seismically qualified motor-generator sets to supply power to
the reactor protection systems. To ensure that these systems can perform their intended
safety functions during a seismic event for which an underfrequency condition of the motor-
generator sets or alternate supply could damage components of these systems, redundant
underfrequency relays are provided. This protection is provided between the alternate power
source and the reactor protection system buses. Operation of either or both of the
underfrequency relays associated with a reactor protection system will cause a half scram of
the unit. If one or both of the underfrequency relays operate on each of the reactor system
protection buses, a full scram of the unit occurs.
Factors to be considered for application of this scheme are:
1. Underfrequency relays' characteristics tolerance.

2. Motor-generator sets' slip characteristic.
3. Load-shedding scheme characteristics.
94
7.10 REQUIREMENTS ON FREQUENCY PROTECTION
7.10.1 Frequency accuracy
The difference in frequency for activation of different preventive and protection actions can
be relatively small. The set frequency difference between such actions can be within the
range 0.05 – 0.2 Hz. Therefore the accuracy of frequency measure measurement can vary
between different power systems. A minimum requirement on accuracy should be 0.025 Hz
(half the setting difference between different protections).
7.10.2 Setting range
The underfrequency protection shall be set within the range 0.80 – 1.00 * the rated
frequency.
The overfrequency protection shall be set within the range 1.00 – 1.20 * the rated frequency.
In case of hydro units, the upper limit could go as high a twice the rated frequency to take
into account situations of load rejection.
7.10.3 Speed
Frequency disturbances are normally slow compared to “normal” faults (short circuits and
earth fault). The time window of the frequency measurement should span several periods in
order to obtain sufficient accuracy. Therefore the speed requirements are moderate.
7.10.4 Voltage influence
During major disturbances both frequency and voltage levels are changed. It is important
that the frequency measurement does not change significantly in case of voltage deviation.
7.10.5 Rate of change of frequency
If a generator is disconnected from the grid the frequency can rise rapidly. Rate of change of
frequency can be useful in situations where the frequency is changing rapidly and it is
necessary to predict the change and trip the generator before dangerous levels of frequency
are reached.
7.11 IMPACT OF DIGITAL PROTECTION SYSTEMS APPLICATION
Many benefits are derived when applying digital protection to this function.
In modern multi-functions generators protection package, frequency measurement is

implemented with a response time of a few periods (2 to 6) and accuracy very often better
than 0.005 Hz. Some systems track the frequency during operation and adapt the digital
protection system accordingly.
The same generators relays will incorporate a number of over/under frequency elements
with an associated definite-time delay. This offers the user the ability to create a protective
function for the specific time-frequency capability of the equipment. The flexibility in selecting
the number of set-points and in setting both the pickups and time delays is an important
advantage. This is far superior to a combination of discrete component relays. The
95
configurability available in digital system allows integrating the outputs into the trip and
control logic. It also allows for modification at a later date with changes in protection
philosophy.
Another aspect is the reduction in the number of devices used to implement these functions.
Using discrete component devices to create the desired scheme is complex and costly. In
some cases it has been unreliable. The digital protection is easily programmable and can
provide more steps in the scheme than were traditionally implemented. Frequency response
of the digital systems is superior to that of the component relays.
7.12 REFERENCES
[1] IEEE Guide for Abnormal Frequency Protection for Power Generating Plants, IEEE
Standard C37.106-2003, 2004.
[2] Rotating Electrical Machines – Part 3: Specific Requirements for Synchronous

Generators Driven by Steam Turbines or Combustion Gas Turbines, IEC Standard
60034-3, Ed. 6, Nov. 2007.
96
8. BREAKER FAILURE PROTECTION
8.1 INTRODUCTION
Breaker failure occurs when one or more breaker poles remain closed after the breaker has
received a trip order and after the normal opening breaker time has elapsed. Breaker failure,
following a trip signal, could be the result of several causes affecting the two conditions that
must be fulfilled in order to declare state that the breaker is in an open state:
• The breaker completely opens mechanically its poles.
• The flashover that occurs during the breaker opening is extinguished.

If a fault or abnormal operating condition occurs inside or outside of the generator zone,
there is an absolute necessity to remove the generator from the system. Within the
generator zone, three different breakers can be identified as shown in figure 8.1.
Unit breaker
Transformer
Generator breaker
breaker
Figure 8.1 – Breakers on generator zone
Breaker failure function needs to be initiated in other situations than the one where a
protective relay operates as a result of a fault or an abnormal condition. As an example, in a
normal operating condition, when the shut down sequence starts, the trip signals sent to the
generator breaker and field breaker are simultaneous. It could happen that the field breaker
will trip normally and the generator breaker will remain closed. The excitation will become
disconnected but the generator will remain connected to the network. As a result, the
generator would change to asynchronous operating mode and several electrical problems
will develop such as:
• Great consumption of reactive power when I = In (In = rated current).
• Possibility of losing synchronism between generator and network.
• Induced currents by the stator flow through the rotor winding and de-excitation
resistor.
Breaker failure protection will provide the signals for tripping of backup breakers if a fault is
detected by the protective relays and the associated breaker does not open after trip
initiation. Depending upon the configuration, these backup breakers could be located in the
97
same station as the generator breaker, or it could be necessary to transfer tripping to
breakers in other stations. Figure 8.2 illustrates an example of such a situation that could be
encountered in multi-breaker arrangements such as ring bus or breaker-and-a half
configurations.
Figure 8.2 – Example of a ring bus configuration
A fault in the protection zone of generator G1 requires tripping of two breakers at station A (1
and 2). If any breaker fails to clear the fault, for example number 1, breaker failure protection
will initiate tripping of the local breaker (3) and transfer tripping to the remote breaker (5).
The breaker failure protection should be fast enough to maintain stability but not too fast to
compromise tripping security. This means that the breaker failure function have to wait
before sending the tripping signal an interval of time equal to the minimum time required by
the breaker to complete its opening sequence. When this latter is completed, the breaker
failure function will be able to operate if it is necessary. In accordance with this criterion,
operating times must be shorter than the line protections second stage tripping times.
8.2 BREAKER FAILURE LOGIC
Figure 8.3 is a logic diagram depicting a basic breaker failure protection scheme.
Figure 8.3 – Breaker failure logic
When the protective relays detect an internal fault or an abnormal operating condition, they
will attempt to trip the generator breaker and at the same time they will initiate the breaker
failure timer. If the generator breaker does not clear the fault in a specified time, when the
timer asserts, it will send a signal to trip the necessary breakers for removing the generator
from the system.
98
Figure 8.3 shows that the start of breaker failure timer requires an initial protective relay
operation together with an asserted current detector or a closed breaker auxiliary contact
must indicate that the breaker has failed to open.
Since there is no clear-cut justification for the presence of a breaker auxiliary contact in the
breaker failure logic scheme, a discussion is needed in order to justify its use.
On the one hand, there are faults and/or abnormal operating conditions such as stator or bus
ground faults, over excitation, negative sequence, reverse power flow, etc..., which may not
produce the overcurent condition required by the breaker failure. This situation points to the
necessity of using another indication different from the current flow. For this reason, an
auxiliary contact that provides information about the breaker position is included in the
breaker failure logic. So, if the auxiliary contact remains closed when the breaker opening
time has expired, the breaker failure protection will operate.
On the other hand, what happens if the auxiliary contact is not included in the logic scheme?
When a trip occurs with the machine unloaded, without current, the generator breaker could
not open although the field breaker would. Under these conditions, and right at the
beginning, if the logic did not include the auxiliary contact, the current detector would not
detect any current and the breaker failure protection would fail to operate.
Under this condition, several seconds later, the generator will lose completely its excitation,
and starts operating in an asynchronous mode and presents to the network a reactance with
the same value of the machine synchronous reactance, Xs. In this situation, the generator
consumes a great amount of reactive power and currents, close to their nominal value, will
flow through the breaker. When the current reaches a magnitude greater than the current
detector setting, the current detector will pickup and will triggers the breaker failure
operation.
The difference between having or not the breaker auxiliary contact lies in the fact that with
the auxiliary contact the trip occurs immediately, while without the contact the trip will be
delayed until the loss of excitation causes the current to increase to values that the current
detector will sense. This seems undesirable, but it is necessary to take into account that with
no current flow, the operating time of the breaker failure protection is not critical. Also, the
use of a breaker contact involves other risks due to the limited reliability of the information
provided by the contact. Let us mention the following situations:
1. The contact opens although the breaker pole has not opened completely.
2. The contact opens and the breaker pole does not open, due to a failure in the
opening mechanism.
3. The contact does not open although the breaker pole has completed its
opening.
In any of these cases, the result would have been the absence of operation or a
misoperation of the breaker failure protection. These possibilities indicate that, except in
specified cases, the sole use of current detectors is preferable.
Current detectors could detect any of the three phase or the neutral currents as a first option
or any the three phase currents only as a second option. As applied to the situation
described earlier, the choice between the two options is not necessarily clear, because both
of them have advantages and shortcomings that depend upon the circumstances.
Normally, the current detector s would be set at the next values shown for each of the
options:
99
1. Detection of any of three phase or neutral currents:
• I phase: > In
• I neutral: ≈ 0,2 In
2. Detection of any of the three phase currents:
• I phase: ≈ 0,5 In
where In is the nominal current.
The theoretical advantages associated with the detection of any of the three phase or neutral
currents are: the current detectors do not start in normal operating conditions and the
sensitivity remains for single-phase and bipolar faults.
The disadvantage of this first option is that breaker failure could misoperate if the breaker
opens normally but there is a residual current in the neutral current detector. This could
occur because of to several possible causes such as saturation of current transformers,
ground contacts in the current transformers circuit, etc...
The advantage of using the second option (detection of any of the three phase currents only)
is that there is no possibility of undesirable trips caused by the neutral current.
The disadvantage of this second method is that the current detector unit starts in normal
operating conditions.
When the moment comes to select one scheme or the other, it is very important to take into
account the consequences of an undesirable trip. One has to distinguish between the cases
of a breaker failure scheme that imposes the tripping of all the lines in a substation from the
case where it is sufficient to transfer the trip to another breaker, without shutting down other
lines. Generally speaking, it would not be wise to recommend one method or the other for
any breaker failure protection. Depending on the specific application of the breaker failure
protection, the analysis of the consequences associated with an undesirable trip, would
ultimately impose one solution or the other.
8.3 OPEN GENERATOR FLASHOVER PROTECTION
Another form of breaker failure that can occur and significantly damage the generator is an
open breaker flashover, that is, an internal or external flashover across the contacts of one
or more breaker poles that will lead to the energization of the generator.
Breaker flashover will most likely occur just prior to synchronizing or just after the generator
is removed from service when the voltage across the generator breaker contacts
approaches twice the normal value as the generator slips in frequency with respect to the
system. Although circuit breakers are rated to withstand this voltage, the probability of a
flashover occurring during this period is increased. It is very rare that this kind of flashover
will affect simultaneously the three phases.
If one or two poles of a breaker flashover, the resulting unbalanced current will generally
cause the generator negative sequence relay to trip. The most commonly applied protection
schemes are therefore designed to detect the flashover of one or two breaker poles.
When this situation occurs, the resulting unbalanced current will generally cause the
generator negative sequence relay or possibly ground overcurrent backup relays to operate,
100
which will initiate a tripping of the flashed-over breaker, if the current detector has enough
sensibility.
Another approach used to speed up the detection of a breaker flashover is to modify the
breaker failure scheme as shown in figure 8.4.
Figure 8.4- Modified breaker failure logic
An instantaneous over current relay (50N) is connected to the neutral of generator step-up
transformer. The relay' output is supervised by the generator breaker auxiliary contact and
provides an additional initiation of the to the breaker failure logic. When a flashover occurs,
the resulting residual neutral current triggers the neutral overcurrent relay operation and
initiates the breaker failure trip.
Generator flashover can also be detected by breaker pole disagreement relaying. This relay
monitors the three phase currents flowing through the breaker and detects whether any
phase is below a certain threshold level (indicating an open breaker pole) at the same time
that any other phase is above a substantially higher threshold level (indicating a closed or
flashed over pole). If this condition is fulfilled, the relay will operate.
8.4 REFERENCES
[1] IEEE tutorial on the AC protection of synchronous generators. January 2006.
[2] M. Gutierrez, J.A, Gonzalez, C. Lopez. "Protecciones de centrales electricas".
IBERDROLA S.A. June 1994.
101
9. UNBALANCED LOAD (NEGATIVE SEQUENCE CURRENT) PROTECTION
9.1 INTRODUCTION
The negative sequence component of the unbalanced current in a generator’ stator winding
produces an air-gap flux that induces an electromagnetic force (EMF) in the rotor. The air-
gap flux caused by negative sequence current rotates with synchronous speed in an
opposite direction from the rotor, therefore, a double-frequency EMF is induced in the rotor
surface. The EMF produces a current flow in the generator damper winding, in the surface of
the rotor, the slot wedges and on the retaining ring. This negative sequence component of
current can cause high and dangerous temperatures in a very short time. It is common
practice to provide protection for the generator for external unbalance conditions that might
damage the machine.
9.2 RELATIONSHIP OF NEGATIVE SEQUENCE AND UNBALANCED THREE-PHASE CURRENT
All three-phase currents can be resolved into a positive-sequence component having the
same direction as phase rotation, a negative sequence component having the opposite
direction from the phase rotation and a zero-sequence component.
The relation between the three phase current phasors quantity of I L1 , I L2 , I L3 and
symmetrical current phasor quantity of I 0 (zero-sequence), I 1 (positive sequence), and I 2
(negative sequence) is provided by the next equations:
I0 =
1
(I L1 + I L 2 + I L3 )
3
1
(
I 1 = I L1 + a I L 2 + a 2 I L 3
3
) (7.1)
1
(
I 2 = I L1 + a 2 I L 2 + a I L3
3
)
where;
2
j π
a = e j120° = e 3
(7.2)
2
−j π
a 2 = e j 240° = e 3
When the generator three-phase currents are perfectly balanced, I 1 is equal in magnitude
to each phase current, and I 2 and I 0 become zero.
When the generator three-phase currents are not balanced, I 2 and I 0 will exist with a
magnitude in proportion to the level of unbalance..
In general, the generator circuit with high resistance neutral earthing has negligible I 0 .
Moreover I 0 gives no influence against the generator rotor because total interlinkage flux to
the rotor circuit becomes zero, theoretically.
102
9.3 SYSTEM CONDITION CAUSING NEGATIVE SEQUENCE CURRENT
As mentioned above, unbalanced current follows the negative sequence current. The most
common causes leading to three-phase unbalanced currents are:
1. System asymmetries (due to non-transposed line)

2. Unbalanced loads
3. Unbalanced faults in the grid
4. Open phases
5. Application of single phase step-up transformer with different impedance
9.4 OTHER
Harmonic components in the phase currents have an impact equivalent to the negative
sequence current. Harmonic current in the generator stator winding produces rotating air-
gap flux with the same speed as the harmonic number, the dominant harmonics will normally
be the (6n+1)th and the (6n-1)th. The (6n+1)th harmonics rotate in the same direction as the
phase rotation, while (6n-1)th harmonics rotate in the opposite direction from the phase
rotation. Therefore, 6n-th harmonic EMF is induced in the surface of the rotor. This influence
of the harmonics shall be considered as the equivalent negative phase sequence current.
9.5 EFFECT ON GENERATOR BY NEGATIVE PHASE SEQUENCE CURRENT
9.5.1 Negative sequence generator damage
For balanced system conditions with only positive sequence current flowing, an air gap flux
rotates in the same direction and in synchronism with the field winding on the rotor. During
unbalanced conditions, negative sequence current is produced. The negative sequence
component of the current rotates in the opposite direction from the rotor. The flux produced
by this current as seen by the rotor, has a frequency equal to twice the synchronous speed
as a result of the reverse rotation combined with the positive rotation of the rotor. The main
effect of the skin effect of the double-frequency rotor current is that it will be located in the
surface elements of the rotor.
Skin effect causes the twice frequency currents to be concentrated at the surface of the pole
face and teeth. The rotor wedges and the metallic strips below the wedges, which are
located near the surface of the rotor, conduct the high frequency current. This current flows
along the surface to the retaining rings. The current then flows across the metal to metal
contact of the retaining rings to the rotor forging and wedges. Because of the skin effect,
only a very small portion of this high frequency current flows in the field windings.
9.5.2 Limit and against negative phase sequence current
Negative sequence heating in synchronous generators is a well-defined process that

produces specific limits for unbalanced operation. Except for a small stator loss, the losses
due to the negative sequence current appear in the machine rotor.
Negative sequence heating beyond rotor limits results in two failure modes. First, the
wedges are overheated to the point where they are brought to the point of rupture. Second,
the heating can cause the retaining rings to expand and float free from the rotor body, which
103
results in arcing at the shrink fits. In smaller machines, the failure of the shrink fit occurs first,
and in larger machines, the rupture of the wedges after they have been annealed from
overheating occurs first. Both failure modes will result in significant equipment downtime for
repairs to the rotor body.
Because excessive negative sequence current produces high induced voltage in the rotor, it
can cause arcing pits on the rotor wedges and at inner surface of the rotor teeth.
Subsequent, this can lead to severe damage.
The energy input to the rotor and the rotor temperature rise over an interval of time are
2
closely proportional to I 2 t , where I 2 is the negative sequence current from the stator and
t is the interval of time in seconds. The following rating method was developed based on the
concept of limiting the temperature to the rotor components below the damaged level. The
limit is based on the following equation for a given generator:
K = I2 t
2
(7.3)
where;
K is the a constant depending on generator design and size
I2 is the RMS value of negative sequence current in p.u.
t is the time in seconds
The limiting K value is determined by placing temperature sensors on the rotor of the
generator along the negative sequence current path while negative sequence current is
supplied to the stator. This monitoring is used to determine the limiting negative sequence
currents the rotor can withstand.
The value of K is provided by the generator manufacturer for each specific unit in
accordance with standards.
9.5.3 Generator negative phase sequence capability
The continuous unbalance current capability of a generator is defined in standards IEEE

C50.12, IEEE C50.13 and IEC60034-1. A generator shall be capable of withstanding,
without any damage, the effects of a permanent negative sequence current I 2 the maximum
value of which is provided in table 9.1 and 9.2 for various generators type.
Tables 9.3 and 9.4 provide the maximum values for the limit factor K for the same types of
synchronous generators.
104
Table 9.1: Salient pole machines [1], [3]
Permissible I2 [%]
IEEE C50.12 IEC 60034-1
Nonconnected amortisseur winding 5
Connected amortisseur winding 10
Indirectly cooled 8
Directly cooled 5
Table 9.2: Cylindrical rotor synchronous machines [2], [3]
Permissible I2 [%]
IEEE C50.13 IEC 60034-1
Indirectly cooled 10
Directly cooled, ≤ 350 MVA 8
Directly cooled, 351 – 1250 MVA S [MVA] − 350
8−
300
Directly cooled, 1251 - 1600 MVA 5
The short time (unbalanced fault) negative sequence capability of a generator according to
IEEE C50.12, IEEE C50.13 and IEC60034-1 are described below.
Table 9.3: Salient pole machines [1], [3]
K (I2t) [s]
IEEE C50.12 IEC 60034-1
Indirectly cooled 20
40
Directly cooled 15
Table 9.4: Cylindrical rotor synchronous machines [2], [3]
K (I2t) [s]
IEEE C50.13 IEC 60034-1
Indirectly cooled, air-cooled 30 15
Indirectly cooled, hydrogen-cooled 10
Directly cooled, ≤ 350 MVA 8
S [MVA] − 350
10
Directly cooled, 351 – 800 MVA
8 − 3⋅
Directly cooled, 801 – 900 MVA S [MVA] − 800 550
10 −
Directly cooled, 901 – 1600 MVA 160 5
105
9.6 NEGATIVE SEQUENCE CURRENT PROTECTIVE SCHEMES
9.6.1 Principle of negative sequence current protection
With the unbalance current capabilities of the generator defined by the negative sequence
current as measured at the stator, a negative sequence inverse-time protection can be used
to protect the generator. This protection becomes active if the negative sequence current is
higher than the continuous permissible negative sequence current given by the generator
manufacturer.
Figure 9.1 shows a typical negative sequence protection application. This protection consists
of three portions:
1. Negative sequence segregating circuit from the three phase currents. This
measurement can be located on the star point side or generator main leads.
2. Constant current detection circuit (usually for alarm) that responds to amount
of permissible negative sequence current.
3. Negative sequence detection circuit that has inverse time characteristic to
match as closely as possible the generator’s negative sequence capability
( I 2 t ≤ K ). With an integration of the I2 current a thermal behaviour can be
2
realised.
I 2>
Figure 9.1: Typical Connection of Negative Sequence Protection
Typical values of negative sequence current for various generator operation are as follows:
Table 9.5
Generator operation Negative sequence current (I2 / In)

a) Negative sequence current under normal operation Maximum 0,01 – 0,03 per unit
b) Unbalanced load condition, including one phase open Approximately 0.2 - 0.6 per unit
c) Unbalanced system fault (for instance two phase short Above 1.0 per unit
circuit in the grid)
106
9.6.2 Electromechanical protection
The electromechanical protection normally uses an inverse time characteristic. This type of
2
protection has not enough sensitivity to coordinate with generator I 2 t withstanding
capability especially at low I 2 level, making it more difficult to protect continuous
withstanding capability of the generator due to the sensitivity limitation of the mechanical
means. This protection will have difficulties to protect against operation event b) in table 9.5.
Coordination would be more difficult for large-capacity generators because the continuous
capability and K value of such generators becomes small.
9.6.3 Digital protection
Present types of negative sequence protection are numerical protection relays with an
improved sensitivity and functions compared to electromechanical devices. These relays
calculate from the phase current directly the negative sequence currents. Figure 9.2 shows
the typical characteristic of this protection. This protection has normally a pick-up range of
2
0.03 to 0.3 per unit, and can protect the generator suitable for the inverse time I 2 t
withstanding capability, and the continuous negative sequence withstanding capability. The
inverse time characteristic should be considered the thermal heating by integration the
current over time.
It is recommended that the protection have possibility to combine inverse time and definite
time characteristics to achieve the best protection for the generator.
Because the operator can, in many case, reduce negative sequence current caused by
unbalanced conditions (such as by reducing generator load), it is advantageous to provide
an alarm when the machine continuous capability is exceeded.
4
Per unit Negative Sequence Current
Def. time stage

Minimum time delay
0.1
Max Continous I2
K=2s
K=5s
K = 10 s
K = 15 s
K = 20 s
K = 30 s
K = 40 s
0.04
1 10 100 1000 10000
Time [s]
Figure 9.2: Typical Characteristics for Numerical Protection
From the generator manufacturer general two parameters are given, which shown in figure
9.3.
107
10
I22 x t = 10 s
Per unit Negative Sequence Current

1
I2 = 0.08 continous
0.1
0.01
0.1 1 10 100 1000
Time [s]
Figure 9.3: Negative sequence limit for a generator
9.7 IMPACT OF NUMERICAL PROTECTION IN THIS AREA
Use of a numerical type negative sequence protection would offer the following operating
advantages:
1. Improved frequency characteristic.

2. Improved electrical information linkage to the unit monitoring system.
3. Improved information by showing data like unbalanced current (I2) and heat
content in the display.
4. Use of alarm level to give the operator the possibility to reduce the generator
load.
9.8 CONCLUSION
The recent application of the electromechanical, static and numerical negative sequence
protections has been summarized.
It is recommendable that an electromechanical protection will be retrofitted to the
numerical protection, thereby improving sensitivity for negative sequence current
from the point of view to protect the withstand capacity of large capacity generators.
9.9 REFERENCES
[1] IEEE Std C50.12-2005, IEEE Standard for Salient-Pole 50 Hz and 60 Hz Synchronous
Generators and Generators/Motors for Hydraulic Turbine Application Rated 5 MVA and
Above.
[2] IEEE Std C50.13-2005, IEEE Standard for Cylindrical-Rotor 50 Hz and 60 Hz
Synchronous Generators Rated 10 MVA and Above.
[3] IEC 60034-1:2004, Rotating electrical machines – Part 1: Rating and performance
[4] CONCORDIA, CHARLES, Synchronous Machines -Theory and Performance.
108
10. SYSTEM BACKUP PROTECTION
10.1 INTRODUCTION
Backup protective functions are provided for faults away from the primary generator
protective relaying functions that are not cleared in time. These backup protective functions
protect the generator against system faults that could not be detected by the primary
relaying systems for transformers, lines, busses, motors, etc.
Two types of protective relays are used to provide system backup protection: voltage-
dependent time-overcurrent relays or distance relays. The choice of relay is based on the
type of protective relays used in the power system to coordinate and be selective. In
industrial plants, for example, where most of the protective relaying is done using
overcurrent relays, generator system backup will use a voltage-controlled / restrained
overcurrent relay to coordinate. Large, unit connected generators, connected to transmission
lines most likely will have distance elements to coordinate with transmission line protection
for backup.
10.2 GENERATOR FAULT CURRENT
Generator system backup functions are applied with a long time delay to allow the primary
relaying functions to operate for faults in the power system. As discussed in Chapter 1,
generators exhibit a time-varying fault current behavior denoted by Eq. 1.2 (Chapter 1) and
shown graphically in Figure 10.1.
60
40
I 20
n
20
0 0.2 0.4 0.6 0.8 1
t
n
.
Figure 10.1 – Generator fault current decrement characteristics
The behavior is called “Generator Decrement Characteristic” [1] and describes the
decreasing fault current magnitude from the point of view of the generator. This
characteristic is determined by sub-transient, transient and steady state generator constants
for the AC component and the short circuit time constant for the DC component.
The generator excitation system for a short circuit will try to boost the generator terminal
voltages; therefore allowing for more fault current than the one with constant excitation.
Techniques and mathematical equations are available to calculate the behavior of the
decrement curve for a generator [1]. Figure 10.2 is an example of a derived generator
decrement curve for a particular generator with particular sub-transient, transient and steady
state constants.
109
Fig. 10.2- Example of a generator decrement curve [1]
The figure illustrates the fault current magnitude relationship with time. The time axis is
plotted up to 20 seconds; although at 2 seconds the generator has reached is steady state
value. The steady state value, with a no-load constant excitation is found to be below the
generator full load current (denoted by GEN FLA).
Generator system backup relays should recognize the fact that the fault current could be
below the generator full load current magnitude. Moreover, generators are designed to
handle certain levels of emergency load above their nominal capability and backup relays
should not operate during this condition.
10.3 GENERATOR SYSTEM BACKUP RELAYS
Simple inverse or definite time overcurrent relays are not suitable for generator system fault
backup relays. The backup relays should be able to differentiate between high load
conditions and fault conditions with considerations of the generator decrement
characteristics.
Voltage-controlled and Voltage-restrained overcurrent relays are used to incorporate voltage

information in the fault detection. Distance relays incorporate the voltage information
inherent in their characteristics.
10.3.1 Voltage dependant overcurrent elements
The voltage-controlled and voltage-restrained overcurrent relays are denoted by the IEEE
functional number 51V. Figure 10.3 illustrates a typical application of a bus connected
generator. The 51V relay should provide backup for system faults and coordinate with
downstream overcurrent relays.
110
Figure 10.3 – Typical 51V application
10.3.1.1 Voltage controlled overcurrent elements
The voltage-controlled overcurrent relay is enabled once the controlling voltage falls below a
threshold. Typically the threshold is around 80% of the nominal generator terminal voltage.
When there is a system fault, a depressed voltage is expected at the terminals of the
generator; therefore enabling the overcurrent measurement. The voltage threshold controls
the operation of the inverse overcurrent unit. The pickup threshold of the element can be set
below load levels.
Fig. 10.4- Desired coordination with a voltage controlled overcurrent relay
Figure 10.4 illustrates a desired coordination scheme with a voltage-controlled inverse-curve

overcurrent relay. Being a backup function, the unit should operate for faults with the
generator at constant excitation. Selecting the relay pickup value below the calculated
sustained fault current ensures the operation. The time-dial and any additional delay
associated with the coordination with downstream devices will determine the actual position
of the curve with respect to the decrement characteristics of the generator.
Although in Fig.10.4 the generator current decrement curve and the relay curve are plotted
on the same time-current plane; these two curves are of different nature. The relay curve
111
indicates the operating time for a constant current. The generator current-decrement
characteristics denote the magnitude of the fault current over time. As in many other
techniques in protective relaying (motor inrush curves, for example) using the two curve in
the same time-current graph allows for coordinating devices for selectivity. It is also
noticeable that the trip times of the relays are not the ones denoted by the curves in the
graph. The trip times will be faster and the actual time can be found by integrating the
operation of the relay in time steps with the different current magnitudes [1].
10.3.1.2 Voltage restrained overcurrent elements
Voltage-restrained over-current relays have a variable pickup depending on the measured

voltage. Figure 10.5 is an example of a typical voltage to pickup characteristic. The
numerical nature of modern protective relays allow for a linear characteristic as shown in the
figure. Older electromechanical devices for example, may not exhibit a linear relationship;
but, the concept is similar. The lower threshold is typically chosen between 12.5% and 25%
depending on the design.
% Pickup
Fig. 10.5- Typical voltage to pickup characteristic of a voltage restrained relay
There are typically more considerations to make when applying voltage-restrained relays
compared to voltage-controlled relays. Two extremes are checked to satisfy coordinating
with the generator decrement characteristic. The 100% voltage pickup should be higher than
the maximum load expected from the generator. (Settings of 150 to 250 % of the nominal
load current are typical for the pickup of this relay). The 0% voltage pickup should cover the
expected sustained fault current with no load constant excitation.
Figure 10.6 illustrates a desired coordination scheme with a voltage-restrained overcurrent

relay. The curve location, shape and any additional delay will be determined by the
downstream devices to coordinate with. As discussed earlier for the voltage-controlled relay,
the times observed from the curves on the graph are not the exact tripping times due to the
nature of the generator fault current and impedance states. The trip time would be faster and
could be calculated using time steps.
112
Fig. 10.6 Desired coordination with a voltage restrained overcurrent relay
10.3.1.3 General Considerations
Either the voltage-controlled or the voltage-restrained overcurrent relays are considered

backup protection for phase faults. Ground fault system backup will be discussed in a later
section.
For industrial installations, for bus connected generators serving at the same voltage level
than the load, voltage-dependent over-current protection will most likely be used. This
implies that this function is most likely found in smaller size generators. It is the nature of the
downstream over-current relays that facilitates the coordination of this type of relay.
10.3.2 Distance relay backup
When the output of a generator is stepped up to a transmission level voltage and the
transmission lines are being protected with line distance relays, a phase distance relay is
used to provide to the generator the needed system fault backup.
Coordination with downstream distance protective relays is simpler than in the over-current
case. The generator decrement characteristics do not influence the application of distance
elements since these are measuring the instantaneous impedance to the fault.
113
Fig. 10.7 Unit connected-generator distance backup scheme
Figure 10.7 illustrates a typical application of distance backup relays for system faults.
Present numerical generator protective relays are provided with two zones of protection.
In some applications Zone-1 is set looking into the transformer impedance and
instantaneous. A second time-delayed Zone-2 is set long enough to coordinate with adjacent
line distance elements. In other applications, a single time-delayed Zone-2 is provided.
Fig. 10.8 Typical generator distance system-backup scheme
Figure 10.8 illustrates a typical distance element scheme for system fault backup. The two
zones of protection are illustrated. If the generator is connected through the step-up
transformer to a very long equivalent transmission system, the required Zone-2 setting may
be encroaching into the generator load equivalent impedance. To avoid unnecessary
tripping, either a shorter Zone-2 reach is provided or a ‘Load Encroachment” logic is used,
which is provided in present numerical generator protection packages.
The location of the voltage transformers determines the origin on the R-X diagram, in Fig.
10.8. This phase distance backup relay scheme is intended for system ground faults;
however, it is not uncommon to require the distance element to have a reverse reach, as
shown in Fig. 10.8. The reverse reach includes the generator reactance, and the unit is
providing additional fault detection capabilities. Depending on the design of the distance
114
element, to accommodate properly system phase faults, the distance elements may need to
compensate for the vector group of the step-up transformer.
10.3.3 Ground fault system backup
The voltage-overcurrent and phase distance schemes described in the previous sections
provide backup for phase faults in the system. They are not meant to provide backup for
ground faults, although they may be responding to ground faults.
For bus connected generators, as illustrated in Figure 10.3, when the generator grounding
allows for sufficient ground fault current for ground faults, a neutral overcurrent relay (51N) is
used to coordinate with downstream devices.
For unit-connected generators, as shown in Figure 10.7, a ground distance relay on the
generator side will not operate properly due to the zero sequence isolation due to the
transformer delta side. An inverse neutral overcurrent relay in the transformer neutral is used
to provide system ground fault backup as shown in Figure 10.7.
10.4 REFERENCES
[1] Baker, D. S., “Generator Backup Overcurrent Protection,” IEEE Transactions on

Industry Applications, vol. IA-18, no. 6, pp. 632-640, Nov/Dec 1982.
[2] J. L. Blackburn, Applied Protective Relaying, Principles and Applications, New York, NY:
Marcel Dekker. Inc. 1987.
[3] W. Elmore, Protective Relaying Theory and Applications, New York, NY: Marcel Dekker.
Inc. 1994.
[4] D. Ventruella, P. Steciuk, “A Second Look at Generator 51V Relays”, IEEE Transactions
on Industry Applications, vol. 33, no. 3, pp.848-856, May/June 1997.
[5] IEEE C37.102-1995 “Guide for Generator Protection”.
115
11. OVEREXCITATION (VOLTS PER HERTZ) AND OVERVOLTAGE
PROTECTION
11.1 INTRODUCTION
When voltage and frequency are allowed to deviate outside specified limits, thermal and
dielectric stresses occur on equipment, damage and/or failure can occur in a matter of
seconds. Generators and transformers are subjected to overexcitation when the ratio of
voltage to frequency, generally expressed as volts per hertz exceeds the design limits.
Figure 11.1 shows the established limits from the IEC 60034-1, IEEE C50.12 and
IEEE C50.13 standard. When voltage and frequency are within the limits of zone A, the
machine shall be capable of performing its primary function continuously. It does not need to
comply fully with its performance at rated voltage and frequency, and it may exhibit some
deviations. Within zone B the machine shall be capable of performing its primary function,
but may exhibit greater deviations from its performance at rated voltage and frequency than
in zone A. Note that Zone B is different in the IEC standard and the IEEE standard.
Voltage [p.u.]
1.08
Zone B (IEC)
1.05
Zone A (IEC and IEEE)
1.03
Zone B (IEEE)
Frequency [p.u.]
0.95 0.98 1.02 1.03
0.98
0.97
0.95
0.92
Figure 11.1 Voltage and frequency limits for generators [1], [2], [3]
For transformers, the established standard IEC 60076-1 [4] states that a transformer should
be in continuous service without damage if the volt per hertz is below 1.05 per unit.
Unless otherwise specified by the generator and transformer manufacturer these limits will
apply. Physically what occurs during this event is the magnetic core iron of the generator
and transformer can become saturated beyond rating. Additionally stray fluxes will be
induced onto other non-laminated components. These components are not designed to carry
flux. A volt per hertz protection is applied to protect against this condition. Generally this is
supplied in addition to the volts per hertz controller/limiter provided in the excitation system.
116
Faraday’s law of electromagnetic induction links the induced voltage in an electrical circuit to
the variation of the flux flowing in the turns belonging to this circuit as expressed in the next
equation:
dφ
u = −N (11.1)
dt
Assume that the voltage measured at the output of the generator is sinusoidal with
frequency f:
u (t ) = U max sin( 2π f t + θ ) (11.2)
The maximum flux can then be evaluated by integrating the voltage:
1 U
ϕ max = −
N ∫
U max sin( 2π f t + θ ) dt = − max
2π f N ∫ sin(2π f t + θ ) dt (11.3)
The interval of integration can be chosen so that its outcome will be a constant. It can be
inferred then that the maximum flux is proportional to the ratio of the measured maximum
voltage over the frequency:
U max
ϕ max ≈ (11.4)
f
This last equation indicates that measuring the generator output voltage and dividing this
voltage by the measured frequency accomplishes an indirect flux measurement. The relay
that will perform this kind of indirect flux measurement is generally called a volt-per-hertz
relay. It is common practice to measure the flux then in per unit value by performing the
measurement in comparison to the rated voltage and frequency as in the next equation:
1 U max f rated
ϕ pu ≈ • (11.5)
2 f U rated
From this formula it can be seen that the induction is direct proportional to U/f. The iron
losses increase if the induction is higher than the rated induction.
Overvoltage stress is a concern when the insulation dielectric strength is approached.

Volts per hertz protection will not be adequate for all cases with events that have a
proportional increase in frequency. It is generally the practice to provide overvoltage
protection to protect the generator for this event. It is also good practice to provide surge
arrestors and surge capacitors at the machine terminals to protect the machine against
severe transient voltages. Surge arrestors limit the severe overvoltage magnitude and the
surge capacitors grade a steep rate of rise on the front end of a transient overvoltage.
If the overvoltage condition is allowed to go unchecked, dielectric failure is possible.
An event that can cause an overvoltage or volts per hertz condition is operation of the unit
under regulator control at reduced frequencies during start-up and shutdown. As an example
if the regulator is maintaining rated voltage when the unit is at 95% or lower speed, the volt
per hertz is 1.05 p.u. or greater. Damage could occur if left unchecked. Another condition
that can cause overexcitation is a complete load rejection with the transmission system left
connected to the unit. If the excitation control is in service, the condition should be corrected
quickly. Having the control out of service may sustain the condition and damage could occur.
117
Loss of the voltage transformer signal can also cause an overexcited condition due to the
misoperation of the regulator in the excitation system with loss of the voltage signal.
11.2 OPERATION LIMITS AS DEFINED BY THE INDUSTRY STANDARDS
As stated previously, standards present the generator and transformer limits. Generally the
manufacturer’s limits are obtained individually for the transformer and the generator. The
limits may come in the form of a capability curve. In some cases the generator manufacturer
may chose to provide recommended protective settings instead of capability curves.
In Figure 11.2 it is shown a typical example of limit curves for volts per hertz for a turbo
generator and a step up generator transformer, these are joined together in the same
diagram to show the overall limit that is the basis for the protection settings.
140
135
130
Transformer limit curve
125
V/Hz [%]
Generator limit curve

120
115
110
105
100
0,1 1 10 100 1000
Time [min.]
Figure 11.2: Generator and transformer limits
11.3 VOLTS PER HERTZ PROTECTION
To provide volts per hertz protection traditionally two types of characteristics have been
applied; namely definite-time and inverse-time. With newer digital and numerical protection
user selectable inverse time curve settings are available. Either the user can select specific
point on the desired volts per hertz-time curve or the protection provides sets of volts per
hertz-time curves, from which the user can select a curve that best fits the application.
The three common protection schemes that are used for volts per hertz protection are:
1. Definite-Time (Single-level or Dual-level)

2. Inverse-Time
3. Combination of definite-time and inverse time
Figure 11.3 shows a typical definite–time relay characteristic. Figure 11.4 presents an
inverse-time characteristic. Figure 11.5 presents a typical combined characteristic in a
118
numerical protection. In this case several points on the characteristic can be chosen to
achieve the best coordination with the capability limits.
Volt/Hertz
Protection Zone
Time
Figure 11.3: Typical Definite-Time Relay characteristic
Volt/Hertz
Protection Zone
Time
Figure 11.4: Typical Inverse-Time Relay characteristic
Volt/Hertz
Protection Zone
Time
Figure 11.5: Typical Combined Relay characteristic in a numerical protection
119
A difficulty arises when strictly using dual–level definite–time since there is an unprotected
area where equipment limits could be exceeded. This is illustrated in Figure 11.6. Figure
11.7 shows an optimized protection making use of both inverse–time and definite-time
characteristics.
Volt/Hertz
Unprotected Area
Protection A
Protection B
Time
Figure 11.6 Typical Relay Characteristic for Dual Level, Definite-Time V/Hz Protection
Volt/Hertz
Protection A
Protection B
Time
Figure 11.7 Optimal Protection provided by Using Both Inverse and Definite-Time relays
11.4 OVERVOLTAGE PROTECTION
Generator overvoltage can occur without exceeding the generators volts per hertz limit. This
is especially a problem with hydro generators, due to load rejection. In these cases both the
voltage and speed can increase, and the volts per hertz is approximately constant. In such
cases an over voltage protection is needed.
For protection against overvoltage, a function should be implemented that is set above the
maximum normal operating voltage. Typically this is set around 110% of nominal voltage or
higher. Either an inverse–time or definite–time characteristic can be used. This allows for
coordination with the voltage regulator (AVR), to allow it to respond to a transient condition
first. In some cases an instantaneous device may be used for extreme overvoltage (130-
150%).
11.5 COORDINATION WITH AUTOMATIC VOLTAGE REGULATOR
In some automatic voltage regulators (AVR) there is V/Hz limiters. When this functionality is
used, the AVR will limit the total V/Hz on the generator to a set value no matter what the
speed (frequency) of the unit is. Normally this limiter is only active in automatic modus. It is
important to coordinate the protection setting with this limiter, a trip from the protection can
120
only be allowed if the set limiter value is exceeded. Some limiters have a dynamic
characteristic, which means that a short time over excitation can be allowed. The
coordination with the protection must therefore be both on level V/Hz and on time.
Also the over voltage protection must be coordinated with the AVR. In case of power system
disturbances, the voltage may be high for a short time period. The AVR must be allowed to
control the voltage to the set value before the over voltage protection is tripped. Some AVR’s
have a inverse-time characteristic, in this case several seconds time delay must be allowed
on the over voltage protection.
In case of loss of measuring voltage the AVR may increase the generator voltage. In this
case it is of significance that the over voltage protection is intact. Therefore it is
recommended that the AVR and the protection use different voltage transformers. As an
alternative way to increase the security, different fuses/mcb’s and separate secondary wiring
can be used.
11.6 IMPACT OF NUMERICAL PROTECTION IN THIS AREA
With numerical protection it is easier to set the inverse-time characteristic to comply with the
generator and transformer manufacturers capability curves. A flexible characteristic has
advantages, because sometimes the given Volt per Hertz curve can not be described by a
standard inverse time curve.
Another aspect that must be considered is the wide frequency operating range, especially for
hydro generators which starts with a constant Volt per Hertz.
11.7 REFERENCES
[1] IEC 60034-1:2004, Rotating electrical machines – Part 1: Rating and performance
[2] IEEE Std C50.12-2005, IEEE Standard for Salient-Pole 50 Hz and 60 Hz Synchronous
Generators and Generators/Motors for Hydraulic Turbine Application Rated 5 MVA and
Above.
[3] IEEE Std C50.13-2005, IEEE Standard for Cylindrical-Rotor 50 Hz and 60 Hz

Synchronous Generators Rated 10 MVA and Above.
[4] IEC 60076-1:1993, Power transformers – Part 1: General
121
12. OUT-OF-STEP PROTECTION
12.1 INTRODUCTION
An electrical network operating in stable conditions is characterized by the balance of power

between the mechanical power input to the generator and the electrical power output to the
grid. However, sudden occurrences in an electrical grid, such as large changes in load,
faults, fault clearing and asynchronous closings which disrupt the balance between
generated power and load are usually followed by large oscillations known as power swings.
Several different terms are commonly used to describe these oscillations associated with
generators:
pole slipping
loss of synchronism
out-of-step
These terms are used interchangeably in this chapter..
Another influential factor is the fault clearing time. The size of machines has increased along
with generator reactances while the inertia constants have decreased. The culmination of
these factors has resulted in reduced critical clearing times required to isolate a system fault
near a generating plant before the generator loses synchronism with the power system.
Power swings can remain stable if the disturbance is cleared quickly enough, otherwise an
out-of-step event can occur. A synchronous machine runs at an asynchronous speed
producing successive pole slips with load angles greater than 90 degrees while excitation is
maintained. Large oscillations in the active and reactive power can occur during power
swings and have a negative effect on the generator and network.
When a generator loses synchronism the resulting high peak currents and off-nominal
frequency operation can cause winding stress, pulsating torque and mechanical resonance
that can potentially damage the generator and turbine-generator shaft. The electrical system
stability is also adversely affected and can experience abnormal voltages.
Pole slips have caused major incidents to occur in high voltage electrical grids. Therefore, it
is necessary to provide a protection system that can quickly detect the loss of synchronism
and perform the necessary actions required to prevent a cascading event (reliability). The
protection must also be able to distinguish between stable and unstable swings (security).
12.2 THEORY OF STEADY STATE AND TRANSIENT STABILITY
It is very helpful to revisit the basic principles of power system stability so as to better
understand out-of-step events. A synchronous generator connected to an electrical network
can be represented by a two-machine system (see figure 12.1). The synchronous generator
is modelled as a constant voltage source behind a transient reactance and the network by a
second machine, Only the reactance of the generators and network are considered in this
model for the sake of simplicity.
122
a) Replica b) ohmic-inductive load c) ohmic load
Figure 12.1: Simple Power System Model for Stability
With: EA excitation voltage (field e.m.f)

EB network equivalent source voltage
UG generator terminal voltage
Xd synchronous direct-axis reactance
XN network equivalent reactance
Figure 12.1b and c show the vector diagram for different levels of load where δ describes the
phase angle between the voltages. This angle is mainly influenced by the load current and
the reactance.
The active power can be expressed by the following well known formula:
EA EB
P= sinδ (12.1)
X
With: X = Xd + XN
Figure 12.2 shows the graphical representation of equation (12.1). Point A is the normal
operating point determined by the actual load condition and is supplied by the mechanical
power of the turbine (one third is used as a result of single phase calculation). In the case of
an angle deflection from point A the active power Psupply becomes greater than active power
Pm/3 delivered by the turbine. The differential active power ∆P is taken from the centrifugal
mass of the rotor. As a result the speed of the rotor slows down and the machine returns to
the stable point A. During a deflection from point B the active power Psupply becomes smaller
than the mechanical power Pm/3 and the differential power ∆P is absorbed directly into the
rotor. The rotor accelerates as a result and the generator goes out-of-step..
The conclusion is that the steady state stability limit has a maximum angle of 90°. Therefore
the risk of instability rises as the load is increased. To guarantee stability normally a grid
runs with voltage angles no larger than 60° and mos t of the time the angle δ is close to 30°.
123
Figure 12.2: Active power in relation to the voltage angle between the machines
A second important part is the discussion of the power system transient stability. The
transient condition can be well illustrated by the case of a short circuit on a parallel line
connected to a generator. Figure 12.3a shows the simple network model and figure 12.3b
represents the active power during the different stages of a power swing (that is, pre fault,
fault and post fault).
a)
b)
Figure 12.3: Transient conditions during a fault
(a) Fault on a double line; b) Active power at different stages
The starting point is a healthy system with a voltage angle δbefore and the active power Pbefore.
The short circuit current is mainly determined by the reactance during the three-phase fault
and the active power is dramatically reduced. So as a result the mechanical power (Pm/3) is
124
larger than the active power Pduring. Finally the rotor accelerates and results in an increased
angle δ. After the fault is cleared, the faulty line is tripped (breakers open at both terminals)
and only one line is in service. The active power curve Pafter is now valid. The power Pafter is
greater than mechanical power Pm/3 and it starts the braking (deceleration) of the rotor. The
angle δ comes to rest at the new stable angle δafter..
The system is stable following the disturbance if area A (acceleration) is smaller than area B
(deceleration) otherwise the generator falls out-of-steps and loses synchronism with the
power system. The transient stability limit can also be described by the permissible short
circuit time which correlates with the maximum permissible angle δperm. Stability is
guaranteed if the angle δ during the short circuit is lower than δperm.. The equations for a
simple estimation are as follows:
2 Tm S N (δ perm − δ before )
t SC perm = (12.2)
ω ∆P
With: Tm mechanical time constant
∆P = Pm −
3
2
[
Pduring (δ before ) − Pduring (δ perm ) ] (12.3)
Tm can be derived from the flywheel effect (GD2), where nN is the rated speed and SN is the
rated power.
GD2
2
Tm  n  2
= 2.74  N -1  ⋅ Nm (12.4)
s  min  SN
MVA
The conclusion from the discussion above is that stability problems can be avoided with
countermeasures such as:
1. Increase EA (overexcited operation of synchronous machines)

2. Increase EB (high voltage level in the grid)
3. Increase EA during a fault (field forcing)
4. Decrease X (high degree of meshed networks)
5. Smaller value of δperm (fast fault clearing time - especially for faults close to the
power plant; dead time of a auto reclosing cycle should be so optimised such that
the change in the rotor angle is very small)
6. Decrease Pm (mechanical power) during the short circuit (total or partly closing of
valves; i.e., fast valving)
7. Increase P with additional load impedance during a short circuit (e.g. breaking
resistor)
125
12.3 PROTECTION DESIGN
12.3.1 Measuring Principle
Out-of-step protection relies upon impedance measurements as viewed from the generator
terminals. The measured impedance is compared to specific characteristics to detect if a
power swing is in progress. Figure 12.4 shows the basic model. The generator is connected
to the grid via a unit transformer. The critical situation for a machine occurs when the power
swing trajectory crosses into the generator or unit transformer impedance. The out-of-step
protection should not operate for power swings out in the grid. Therefore two zones are often
used.
Zone 1 Zone 2
EA X’d XT ZN EB
Relay B
A
Location
Figure 12.4: Replica for impedance calculation

(X’d transient reactance of the generator, XT short-circuit reactance of the unit transformer,
ZN reference impedance of the network)
The impedance seen from the generator terminals can be expressed by the following
equation:
EA Z Z
ZR = - jX 'd → Z R = - jX 'd
EA − EB 1+
EB - jδ
e (12.5)
EA
with : Z = jX 'd + jX T + Z N
The graphical representation of formula (12.5) is a circle determined by the voltage ratio
(EA/EB), the actual angle δ and the total impedance Z. The basic curves are shown in figure
12.5. Note that the displacement depends on the generator speed with respect to the
equivalent network.
n(EA) > n(EB) ≡ right to left displacement
n(EA) < n(EB) ≡ left to right displacement
126
jX
B
ZN Impedance
trajectory at the
EA/EB > 1 relay location
EA/EB = 1
XT
EA/EB < 1 X’d
Figure 12.5: Locus of the impedance measured at the generator terminals during a pole slip
The rate of change of impedance trajectory is also important as the impedance must be
measured accurately during time varying impedance conditions. Under the assumption that
ZN is reactive and an additional simplification that the rate of change can be estimated by the
following equation.
dZ(t ) dR(t) X π fP Ω
≈ = in (12.6)
dt dt 2 δ s
2 sin ( )
2
An example calculation is shown in figure 12.6. The rate of change of impedance is at a
maximum at the top of the circle (360°) and at a mi nimum at the bottom of the circle (180°)..
The protection must be able to accurately measure the rate of change of impedance across
the range.
127
Figure 12.6: Speed of the impedance trajectory
(slip frequency fp = 1 Hz and a total reactance of X =10 Ω)
The various types of protection schemes are mainly influenced by the history of relay design
and experience.
12.3.2 Blinder Schemes
A commonly used scheme to detect the generator loss of synchronism is the single blinder
scheme as illustrated in figure 12.7. The blinder units are supervised by a mho unit which is
set to only permit tripping for swings that cross into either the generator or unit transformer
impedance plus a limited portion of the system just beyond the high side of the unit
transformer. The blinders prevent operation during stable swings since the trajectory must
enter the region of the mho unit that resides between the pair of blinders by crossing one of
the blinders and exit the region from the other blinder. The scheme only operates if there is a
progressive change in impedance over time and not a sudden change in impedance which
happens during a fault. The mho unit also prevents unwanted operation during unstable
swings out in the system that do not pass through the mho characteristic.
128
Impedance
Locus during
Power swing if
EA > EB
EA = EB
Figure 12.7: Single blinder scheme [1]
The double-blinder scheme performs in the same manner as the single-blinder scheme. The
supervisory mho element is included in double blinder system to provide the same security
features as discussed in the single blinder scheme.
Referring to figure 12.8 the outer element operates when the swing impedance enters its
characteristic, such as point F. Note that in this scheme the mho unit picks up before the
outer blinder unit. If the swing remains between the outer and inner element characteristics
for longer than a pre-set time then it is recognized as an out-of-step condition. The swing
impedance will continue crossing through the inner element and will reach the outer
characteristic after some time interval. Tripping does not occur until the swing impedance
passes out of the outer characteristic or the mho unit..
The swing angle is defined by points D, F and C in figure 12.8 and is constantly changing as
the swing moves. Once the swing has been detected and the impedance has entered the
inner characteristic the swing can then exit the inner and outer characteristics in either
direction then tripping takes place. Therefore, the setting of the inner element must be such
that it will respond only to swings from which the system cannot recover (i.e., unstable). This
restriction does not apply to the single-blinder scheme because the logic requires that the
apparent impedance enters the inner area from one direction and exits back out from the
opposite.
129
Figure 12.8: Double blinder scheme
12.3.3 Lens Scheme
Figure 12.9 represents a lens scheme. As shown by equation (12.5) and figure 12.5 the
theoretical results of power swings are impedance circles with different diameters. It was
found out that a lens characteristic is an acceptable compromise with the goal that all the
circular trajectories enter into the out-of-step characteristic with the same angle
(approximately 90°).
The diameter of the lens is given by the maximum angle between the voltages EA and EB. A
trip angle of 120° is used in most cases. An additi onal directional element separates the
power swing into two zones, unstable swings passing through the generator-transformer
impedances and those outside.
130
Figure 12.9: Lens scheme
12.3.4 Two Zone Schemes with Straight Lines
A technical modification of the blinder scheme is the application of rectangular

characteristics as shown in figure 12.10. Characteristic 1 detects power swings inside the
generator-unit-transformer impedance and characteristic 2 detects swings out in the
network. The setting parameters are made according to the reactance axis as shown in the
figure 12.4 – they consist of the generator transient reactance, short-circuit impedance of the
transformer and the equivalent impedance of the network. The value on the resistance axis
is given by the maximum permissible angle and in most cases 120° is used.
Zd ≈ X T + Z N
Zc ≈ 0,9 X T
ϕ p ≈ 90° ... 70°
Za ≈ 0,289 ZSum
at δ = 120°
Zb ≈ X 'd
Figure 12.10: Rectangular scheme
131
If the impedance trajectory passes through either characteristic and also crosses the
reactive axis then either counter 1 or 2 is incremented. A counter is incremented after the
trajectory exits a characteristic. There are different rules for counting as illustrated in figure
12.11.
Figure 12.11: Rules for incrementing the counters
Additionally current supervision is used. Operation can only occur if there are symmetrical
fault conditions; that is, no negative-sequence current and a positive-sequence current
higher than the maximum load current (approximately 1.2 IN,G). A short pulse is generated
after each swing (see figure 12.12) and can be used to alarm. Not shown in the logic is the
reset criterion for the counters. After each swing an additional timer is started. If the following
swing cycle is not within the time which was set then stable conditions are assumed and the
counters are reset. The setting value for this timer depends on the lowest power swing slip
frequency plus a safety margin.
Figure 12.12: Logic of an out-of-step protection
12.4 IMPACT ON DIGITAL RELAYS
Numerical relays use the same principles as mentioned here. Again the main measuring
principle is the impedance measuring method.
132
Protection philosophies are influenced by the history and experiences of each company. The
main advantages in using numerical relays are the flexibility in design of the characteristics
and logical combinations. The task priorities are important for multi-function relays because
the out-of-step protection has to react to a swing trajectory in a very short period of time. The
calculations must be executed at least every half cycle. Note that the theoretical upper limit
for slip frequencies is roughly 18 Hz.
The speed of the impedance measurement is critical for double blinder schemes so as to
properly distinguish between system faults and power swings. It is not so easy to determine
the best settings and this requires sophisticated grid analysis.
Relay testing with steady-state test equipment is a critical issue. Transient files generated
with a dynamic network model solve this problem and can be easily played back to the
generator relay with numerical test sets. The oscillographic and event record captured by
numerical generator relays are excellent tools to analyze these events.
12. 5 REFERENCES
[1] IEEE tutorial on the protection of synchronous generators. January 1994.
[2] IEEE Guide for AC generator Protection, IEEE Std C37.102-2006
[3] M. Gutierrez, J.A, Gonzilez, C. Lopez. "Protecciones de centrales de eléctricas”

IBERDROLA S.A. June 1994.
[4] “Numerical generator protection REG 316”, Manual of company ABB
[5] “Multifunction generator, motor and transformer protection relay 7UM62”, Manual of
company Siemens
133
13. ANTI-MOTORING PROTECTION
13.1 INTRODUCTION
A number of situations exist where a generator could be driven as a motor. Anti-motoring

protection will more specifically apply in situations where the prime-mover supply is removed
for a generator supplying a network at synchronous speed with the field normally excited.
The power system will then drive the generator as a motor.
A motoring condition may develop if a generator is connected improperly to the power

system. This will happen if the generator circuit breaker is closed inadvertently at some
speed less than synchronous speed. Typical situations are when the generator is turning
gear, slowing down to standstill or has reached standstill. This motoring condition occurs
during what is called generator inadvertent energization. The protection schemes brought up
to respond to this situation are different that the one dealt with in this chapter and are treated
in more details in the chapter XV on Inadvertent Off-line Energizing.
Motoring will cause adverse effects, particularly in case of steam turbines. The basic
phenomenon is that the rotation of the turbine rotor and the blades in a steam environment
will cause windage losses. Windages losses are function of rotor diameter, blades length
and directly proportional to the density of the enclosed steam. In any situation where no
steam flow is present, harmful windage losses could occur.
The effect of the steam flow through a turbine is first to cause rotation of the turbine and
second to carry away the heat of the turbine elements. In a motoring condition, the steam
flow no longer exists and consequently the heat of the windage losses is not carried away.
Because of this, parts of the turbine could be heated to an abnormal level during motoring.
Maximum permissible time that a turbine can sustain motoring condition is normally available
from manufacture’s data and is normally a function of rated speed.
For gas turbines, gear problems may develop if driven in a motoring condition. In hydro-
turbine, cavitation of the blades may occur in low-water level as possible in a motoring
condition. In Kaplan turbines reverse power can cause mechanical pressure on the bearings
and thus causing damages. Also, hydro-units operated as synchronous condensers will
basically be driven as motors.
From the preceding discussion, one may conclude that the anti-motoring protection is more
of a prime-mover protection than generator protection. There exists at least one condition
however where motoring could have an adverse effect on the generator, that is if the
generator excitation supply where operated under manual mode prior to the motoring
condition. If the generator excitation voltage were supplied by the action of an AVR
(Automatic Voltage Regulator) prior to the motoring condition, the AVR would smoothly settle
to the new motoring condition by changing the field current to the value required by the new
electrical equilibrium. Under manual control, the field current would remain at the same value
and would not correspond to the value required by the new electrical equilibrium. This could
be the cause for the generator to exceed its output voltage and reactive power normal
operation limits.
The most obvious means of detection of motoring is to monitor the flow of real power i n t o t h e
g e n e r a t o r . If the power flow becomes negative below a preset level then a motoring
condition is detected.
134
13.2 REVERSE POWER LEVEL FOR VARIOUS TYPE OF PRIME MOVER
Sensitivity and setting of the power relay depends upon the energy drawn by the prime
mover considered now as a motor.
With a gas turbine, the large compressor represents a substantial load that could reach as
high as 50% of the unit nameplate rating. Sensitivity of the power relay is not an issue and
is definitely not critical.
With diesel type engine, with no firing in the cylinders, load could reach as high as 25% of
the unit rating and sensitivity, once again is not critical.
With hydro turbines, if the blades are below the tail-race level, the motoring energy is high. If
above, the reverse power gets as low as 0.2 to 2 % o f the rated power and a sensitive
reverse power relay is needed then.
With steam turbines, operation at full vacuum and zero steam input, motoring will draw
0.5 to 3% of unit rating. A sensitive power relay is then required. A consequence of this is
that should the turbine be operated with the valves not completely closed, the power
measured at the generator's output could be near zero and the power relay could not
detect the motoring condition. Overheating of the turbine could still occur and the protection
scheme should he supplemented with additional functions. Primary mechanical protection for
detection of motoring would then consist of the following four possibilities with increasing
level of reliability:
a) Measurement of exhaust hood temperature: A temperature-sensing device located in the

exhaust hood is often used as an alarm rather than a primary protection because of the
variations of temperatures in the hood.
b) Valve limit switches: Limit switches on the turbine valves provide an indication on
when steam flow is shut off. This scheme has become too complex to become a
primary protection because of the increased number of switches in modem steam
units. Also one single switch failure may render the protection inoperative.
c) Control oil pressure: Two oil systems are present in a steam turbine. A turbine-trip
system and a governing oil system. Pressure switches from these two systems
provide a reliable detection of a motoring condition. This system can he supervised
by a reverse power relay.
d) Turbine steam flow: Steam flow equal to or greater than the synchronous speed no-load
steam flow is an indication that the unit is not being motored. Therefore use of a
differential pressure switch across the high-pressure element is a very accurate method
of assessing a motoring condition.
Therefore a reverse power relay should be used as a primary protection for hydro, gas and
diesel generator and as a back-up protection for only steam turbine. Because of the difficulty
in measuring a low-reverse power with steam turbines, an additional low forward power
element could be added.. Table 13.1 provides ranges of typical motoring power for different
prime movers in percentage of the turbine power. It should be borne in mind that there exists
some turbine types for which a motoring condition will have little effect or damage on the
turbine: A hydro Pelton turbine is an example of such an exception. Some utilities will not
even apply a motoring protection for this situation.
The most stringent requirement for reverse power relays is their sensitivity at low power
level. Motoring may occur at very low levels of real power and much higher values of
reactive power, that means at a power angle approaching 90 degrees. Reverse power relays
135
must therefore be able to yield a good accuracy in case of low levels of current magnitudes
and high values of reactive power. The reverse power protection sets therefore extra
requirements on the instrument transformers, especially regarding phase angle error. In
certain applications requiring maximum accuracy, metering CTs rather than relaying CTs
should be used. With steam turbines, if valves leakage becomes significant, the motoring
power could get so small that motoring condition could become undetectable.
In order to override conditions of power swings due to disturbances or synchronizing,

reverse power is normally implemented with a long delay that could reach as high as 30 s.
Table 13.1 Typical motoring power for different prime movers
Prime Mover Motoring Power

(% of rated power)
Diesel Engine 5-25
10-15 (split shaft)
Gas Turbine --------------------------------------------
> 50 (single shaft)
0.2-2 (blades out of water)
Hydro --------------------------------------------
> 2 (blades in water)
Steam Turbine 0.5-6
13.3 SEQUENTIAL TRIPPING OF GENERATORS
In the case of steam generators, if following detection of a generator' fault, the generator
breaker is tripped; a condition of over-speed may develop. Overspeeding then may cause
damages to the steam turbine (see abnormal frequency operation of generators in
Chapter 7). In order to prevent this and provided the generator tripping can be delayed,
sequential tripping has been devised and follows the next steps:
• First, the turbine valves are tripped following detection of a fault. Condition for
overspeeding is therefore removed.
• Following the removal of the prime mover, a motoring condition will be detected
by a sensitive reverse power relay and subsequently, a tripping signal will be sent
to the generator and the field breaker.
A contact providing the valve close position will be monitored in series with the reverse
power condition to insure that the steam flow has been reduced to an amount not sufficient
to produce overspeeding.
When applying sequential tripping, the risk of maintaining the generator under the possible
damaging conditions of the current fed by the external network should be evaluated.
136
Figure 13.1 represents a possible logical scheme for the implementation of a sequential
tripping function. If the following three conditions are met:
• The real power is below a pre-set threshold SET_1 and the steam valve or a
differential pressure switch is closed. Note that either condition indicates the
removal of the prime mover. Also the threshold SET_1 could be a positive value
so that, as already mentioned, the element could be a low-forward power element
in order to remove the difficulties in measuring low negative power values.
• The sequential tripping function is enabled

A trip signal will then be sent to the generator and field breakers after a time delay
corresponding to TL1 as shown in the timer on figure 13.1.
P < SET_1
Valve Closed or TL1 0

Pressure Switch
& & Trip Signal
Sequential Trip Enable
Figure 13.1 Implementation of a sequential tripping function.
It should be borne in mind that there would not be any trip signal following a failure of either
the valve switch or the reverse power relay. For that reason, this scheme should be
supplemented with a back-up protection consisting of an additional reverse power relay.
13.4 IMPACT OF DIGITAL PROTECTION
The most important impact of digital protection is on the versatility of the protection schemes
that could be implemented.
It should be borne in mind that in a generator digital protection relay, instantaneous values of
all three output generator voltages and three output currents are acquired at a high
sampling rate (typically above 720 Hz). Two means of computing the power are available:
a) By taking the real part of the sum of the products of the each voltage fundamental
phasor by the conjugate of each corresponding current fundamental phasor as in:
real( P + j Q) = real (VA • conj( IA) + VB • conj( IB) + VC • conj( IC)) (13.1)
This solution could be frequency independent provided special signal processing techniques
are implemented like frequency tracking. Also, it does not include the energy of the
harmonics.
b) By performing the sum of the product of the instantaneous values of voltage by the
corresponding current on all three phases at the same sampling time ti as in:
P = va(t i) · ia(t i) + vb(t i) · ib(t i) + vc(t i) · ic(t i) (13.2)
The advantage of this solution is that it is frequency insensitive and it includes the power of
the harmonics energy.
Whatever the means chosen to compute the real power, the accuracy of the result depends
upon some basic design specifications, namely:
• the accuracy of the VT’s and CT’s
137
• the. dynamic range of the current measurement
• the resolution of the A/D converter
• the discrepancy in the change of phase involved by the analog acquisition

circuitry between different channels (voltages and currents)
• the accuracy in the simultaneous multiplexing of the different instantaneous

values
The two first factors are the most important and in order to get a good measurement of
power at low level of current, a higher resolution with a smaller current dynamic range is
desirable. Theoretically at least, accuracy can be brought to a high level.
As far versatility is concerned, once the measurement of power has been performed, it can
he used in different protection schemes with different delays and logic conditions.
Particularly, the measurement of reverse power can be can be used in three different
schemes within the same system:
• specific implementation of anti-motoring function with a long time delay to trip

by measurement of reverse power
• use of the same measurement of reverse power for the purpose of sequential
tripping function
• use of the same measurement of power for the purpose . of protection against
inadvertent energization but with a much smaller time delay to trip with the
anti-motoring function.
One of the reasons why anti-motoring detection is not used in inadvertent energization
schemes is the longer time-to-trip in conventional component relay. This is no longer an
issue with digital systems.
138
14. VT APPLICATIONS AND LOSS OF VOLTAGE SIGNALS
14.1 INTRODUCTION
The loss of a voltage phase signal can be due to a number of causes. The primary cause
for this nuisance is a blown-out fuse in the voltage transformer circuit. Other causes can be
a wiring error, a voltage transformer failure, contact opening, a misoperation during
maintenance etc.
Since the purpose of the voltage transformers, is to provide voltage signals to protective
relays and voltage regulator, the immediate effect of a loss of VT signals will be the possible
misoperation of some protective relays and the generator over excitation caused by the
voltage regulator.
Among the protective relays to be impacted by the loss of VT signals are.

• Function 21: Distance relay, used for backup to system and generator zone
phase faults.
• Function 32: Reverse power relay, used for anti-motoring function, sequential
tripping and inadvertent energization functions.
• Function 40: Loss-of-field protection.
• Function 51V: Voltage-restrained time-overcurrent relay.
• Function Inadvertent (accidental) off-line energizing ( in case of a 50/27 relay )
• Function 21: Undervoltage protection
• Function 24: Generator and generator transformer over-excitation functions
• Funtion of 64G (100%)
• Function 78: out-of-step protection
• Functions 59N or 59THD: neutral overvoltage and earth fault protection based on
3Uo calculation)
• P, Q and S measurements
• Inter turn protection based on voltage comparison
Normally these functions should be blocked if a condition of fuse failure is detected.
In order to prevent a "run-away" of the regulator, control should be transferred to a manual

operation, to another regulator or VT, if the VT upon which the loss of fuse is detected
supplies a regulator.
14.2 DETECTION OF VT FUSE FAILURE
It is a common practice for large generators to use two sets of voltage transformers for
protection, voltage regulation and measurement.
Therefore, the most common practice for loss of VT signals detection is to use a voltage
balance relay as shown in Fig.14.1 on each pair of secondary phase voltages.
When a fuse blows, the voltage relationship becomes unbalanced and the relay operates.
The voltage balance relay can detect which VT has the fuse failure so that the appropriate
action can be taken. Typically, the voltage unbalance will be at around 15%. Corrosion or
poor contacts in the VT stabs can result in a voltage drop too small for detection by the relay
but high enough to cause a regulator runaway.
139
TO 60
TO
PROTECTIVE VOLTAGE
RELAYS REGULATORS
Figure 14.1 Example of voltage balance relay
14.3 USE OF SYMMETRICAL COMPONENTS
The advent of digital relays has allowed the use of sophisticated algorithms based on
symmetrical components to detect the loss of VT signals.
When a situation of loss of one or more of the VT signals occurs, the following conditions
develop:
1. There will be a drop in the positive sequence voltage accompanied by an increase in

the negative sequence voltage. The magnitude of this drop will depend upon the
number of phases impacted by a fuse failure.
2. In case of a loss of VT voltage and contrary to a fault condition, there should not
be any change in the current magnitudes and phases. Therefore the negative
and zero sequence currents should remain below a tolerance value.
3. A fault condition can be distinguished from a loss of VT signal by monitoring the
changes in the positive and negative current levels. In case a loss of VT signal,
these changes should remain below a small tolerance level.
All the above conditions can be incorporated into a complex logic scheme to determine
if indeed there has been a loss of VT signal or a fault. Figure 14.2 represents the logic
implementation of a single and double fuse failure detection based on symmetrical
components.
A fuse failure condition is established if the following conditions are met within a time delay
longer than T1:
• The positive sequence voltage is below a voltage set-value SET_1
• The negative sequence voltage is above a voltage set-value SET_2
• There exists a small value of currents such that the positive sequence current I1
is above a small set-value SET_4 and the negative and zero sequence currents
I2 and 10 do not exceed a small set-value SET_3
Monitoring the voltage level of each phase and comparing it to a set-value SET_5 can
detect fuse failure of a specific phase. As soon as the positive sequence voltage returns
to a value greater than the set-value SET_1 and the negative sequence voltage disappear
then the fuse failure condition will reset.
140
V2 > SET_2
&
V1 < SET_1
&
FUSE
1 FAILURE
I0 > SET_3 T1 0
1 &
I2 > SET_3
PHASE A
I1 > SET_4 & FAILURE
VA < SET_5
PHASE B
& FAILURE
VB < SET_5
PHASE C
& FAILURE
VC < SET_5
Figure 14.2 Symmetrical component implementation of fuse failure detection
14.4 VT FERRORESONANCE
Ferroresonance is a complex phenomenon likely to occur in a power network where a

capacitance resonates with the nonlinear magnetizing (exciting) reactance of a transformer.
In case of generator VTs, the possibility of occurrence exists if the VTs are ungrounded and
the system turns out to be ungrounded. This will happen as shown in figure 14.2 where the
generator is disconnected and the grounded VTs are energized through the delta connected
step-up transformer.
Under this condition the voltage appearing on the VTs could be strongly distorted with
subharmonic components and the VTs could be operated well above the knee-point
of the saturation curve. As a consequence, the magnitude of the VT’s exciting
current could become substantial and the VTs driven into a thermal runaway in a
very short period of time.
Using line-to-line rated VTs connected line to ground can minimize the possibility of
ferroresonnance. In order to completely suppress ferroresonance, it may be necessary to
apply resistance loading across each phase of the secondary windings sufficient to produce
a loading equal to the thermal capability of the VT. During normal operation, these resistive
loads should be removed.
141
Voltage
Transformers
Unit
Transformer
OPEN
Back
Unit Energized
Aux.
Fig. 14.3 Configuration for Potential occurrence of VTs ferroresonance
14.5 IMPACT OF DIGITAL PROTECTION
The most obvious impact of digital protection in the area of loss of VT signal detection
is the use of symmetrical components. This function can therefore be implemented
even for small and medium size generators where there is no duplication of VT units.
Detection of blown fuses by the use of symmetrical components allows identification of

the faulted phase. Symmetrical components can be supplemented by monitoring of the
voltage change in the absence of a current change to enhance the scheme and further
detect three-phase VT loss. Therefore, complex logic schemes can be implemented
where only the protection function impacted by the phase will be blocked. As an
example, if a 51V is implemented on all three phases independently, it is only
necessary to block the function on the phase which has a fuse failure.
It should be borne in mind that with multi-function relays, the loss of VT signal
detection function would be implemented on the same system together with the
protective functions it is supposed to block.
Also, the computation of voltage and current sequence quantities benefit from the
inherent filtering properties provided by algorithms such as the full-cycle discreet
Fourier transform or Cosine filtering, with the following properties:
1. Immunity to any DC component and good suppression of exponentially

decreasing offset;
2. Immunity to harmonics;
3. Nominal response time of one cycle.
Finally, a number of techniques have been devised to make the measurement of phasor
magnitudes independent of frequency and can therefore achieve stable sensitivities over
large frequency excursions.
142
15. INADVERTENT GENERATOR ENERGIZING
15.1 INTRODUCTION
Inadvertent or accidental energization of turbine generators when it is off-line has occurred

frequently enough within the industry in recent years to warrant concern. When a generator
is energized while off-line on turning gear, or coasting to a stop, it behaves as an induction
motor and can be damaged within a few seconds. Turbine damage can also occur. A
significant number of machines have been severely damaged and in some cases,
completely destroyed. The cost to the utilities for such an occurrence is not only the cost of
repair or replacement of the damaged machine, but the substantial cost of purchasing
replacement power during the period when the unit is out of service. Operating errors,
breaker head flashovers, control circuit malfunctions or a combination of these causes have
resulted in generators becoming accidentally energized while off-line. Most major generator
manufacturers are now recommending some form of dedicated generator inadvertent
energizing protection.
15.1.1 Operational Errors
Despite extensive interlocks, operating errors have increased within the industry as high
voltage generating stations have become more complex with the use of breaker and-a-half
and ring bus configurations. These high voltage generating station configurations are widely
used in a number of countries around the world.
a) Typical breaker-and-half station b) Double busbar station
Figure 15.1: Typical plant schemes
(MT: main transformer – step-up transformer, AT: auxiliary transformer)
These station designs provide sufficient flexibility to allow a single high voltage generator
breaker (CB A or CB B) to be taken out-of-service without also requiring the unit to be
removed from service. Breaker disconnect switches (not shown) are available to isolate the
breaker for repair. When the unit is off-line, however, generator breakers (CB A and CB B)
143
are generally returned to service as bus breakers to complete a row in a breaker-and-a-half
station or to complete a ring bus. This result in the generator being isolated from the system
through only an open high voltage disconnect switch (S1). Removing generator straps or
other sectionalizing devices in the generator isophase bus can provide additional isolation
from the power system. Generally, these isophase bus devices are opened to provide safety
clearances or isolation for extended unit outages. There are many instances in which the
high voltage disconnect switch (S1) provides the only isolation between the machine and the
system. Even with extensive interlocks between the generator breakers (A and B) and the
disconnect switch (S1) to prevent accidental switch closure, a significant number of cases
have been recorded of units being accidentally energized through this disconnect switch
while off-line. Compounding this problem is the possibility that some or all generator
protection, for one reason or another may be disabled during this period.
Operating errors, which accidentally energize generators, also occur on less complicated
generator substation configurations where the generation is isolated from the system
through a dedicated circuit breaker. Another path for inadvertent energizing of a generator is
through the unit auxiliary system by accidental closure of unit auxiliary transformer breakers
(CB C). Because of the higher impedance in this path, the currents and resulting damage
are much lower than those experienced by the generator when it is energized from the
power system.
15.1.2 Breaker Head Flashover
The extreme dielectric stress associated with HV and EHV breakers and the small contact
gap spacing associated with their high-speed interrupting requirement can lead to contact
flashover. This flashover of contacts (generally one or two poles) is another method by which
generators have been inadvertently energized. The risk of a flashover is greatest just prior to
synchronizing or just after the unit is removed from service. During this period, the voltage
across the open generator breaker can be twice normal as the generator slips angularly with
respect to the system. A loss of pressure in some types of HV and EHV breakers during this
period can result in the flashover of a breaker pole(s), energizing the generator and causing
a significant flow of damaging unbalanced current in the generator windings. Tripping the
generator breaker will not arrest the situation since the breaker is mechanically open.
Activation of generator breaker failure scheme is required. This unique breaker failure
condition must be quickly detected and isolated to prevent major generator damage.
15.2 GENERATOR RESPONSE TO INADVERTENT ENERGIZING
15.2.1 Generator Response to Three-Phase Energizing
When a generator is accidentally energized with three-phase system voltage while on

turning gear, it behaves as an induction motor. During three-phase energization at standstill,
a rotating flux at synchronous frequency is induced in the generator rotor. The resulting rotor
current is forced into sub-transient paths in the rotor body and damper windings (if they
exist) similar to those rotor current paths for negative sequence stator currents during
generator single phasing. The machine impedance during this high slip interval is equivalent
to its negative sequence impedance jX2G.
If the generator is connected to a strong system, the machines initial stator currents will be in
the range of three to four times rating and the terminal voltage will be in the range of 50-70%
of rated, for typical values of generator and step-up transformer impedances. If the generator
is connected to a weak system, machine stator current may only be one to two times rating
and the terminal voltage only 20-40% of rated. When the generator is inadvertently
144
energized from its auxiliary transformer, stator current will be in the range of 0.1 to 0.2 times
rating because of the high impedance in this path.
15.2.2 Generator Response Due to Single Phase Energizing
Single phase energizing of a generator from its high voltage system while at a standstill
subjects the generator to a significant unbalanced current. The unbalanced current causes
negative sequence current flow and thermal rotor heating similar to that caused by three-
phase energizing. There will be no significant accelerating torque if the voltage applied to the
generator is single-phase and the unit is essentially at standstill. Both positive and negative
sequence currents will flow in the stator and each will induce approximately fundamental
frequency (50 or 60 Hz) currents in the rotor. This produces magnetic fields in opposite
directions essentially producing no net accelerating torque. Breaker head flashover is the
most frequent cause of single-phase inadvertent energizing.
15.3 GENERATOR DAMAGE DUE TO INADVERTENT ENERGIZING
The initial effect of inadvertent energizing of a generator from standstill or on turning gear is
rapid heating in iron paths near the rotor surface due to stator-induced current. These paths
primarily consist of the wedges, rotor iron, and retaining rings. The heating begins to melt the
metals, and may cause wedges to be weakened to the point of immediate or eventual
failure, depending upon the tripping time to clear the inadvertent energization incident.
In the case of a cross-compound unit, sufficient field is applied at a very low speed to keep
the generators in synchronism as they come up to speed. Inadvertent application of three-
phase voltage will attempt to start both generators as induction motors. The thermal hazard
to the rotor is the same as when no field is applied and is aggravated by the presence of
current in the rotor field winding.
Hydro-generators are salient-pole machines and are usually provided with damper windings
on each pole. These damper windings may or may not be connected together. Inadvertent
energizing may create sufficient torque in the rotor to produce some rotation. More
importantly the thermal capacity of the damper winding is rapidly exceeded. The heating of
the connecting points, combined with the lack of proper ventilation, will create damage
quickly.
15.4 RESPONSE OF CONVENTIONAL GENERATOR PROTECTION TO INADVERTENT ENERGIZING
The ability of relays, used in a typical generator protection scheme, to detect inadvertent
energizing is severely limited due to their marginal operation, lack of sensitivity or disabled
state when the generator they are protecting is off-line. Inadvertent energizing protection
needs to be in-service when the generator is out-of-service. This is the opposite of normal
protection. Frequently, utilities disable generator protection when the unit is off-line to
prevent undesirable tripping of generator breakers that have been returned to service as bus
breakers in breaker-and-a-half and ring bus substations. It is also a common operating
practice to remove generator voltage transformer (VT) fuses as a safety practice when the
generator is removed from service. This disables voltage dependent relays from providing
inadvertent energizing protection. Many utilities use auxiliary contacts (52/a) of the generator
high voltage disconnect switch to automatically disable generator protection when the unit is
off-line which can prevent these relays from being operative for inadvertent energizing
events. In many cases, engineers who rely on the normal compliment of generator relaying
for inadvertent energizing protection fail to recognize any of these common operating/control
practices that disable protection.
145
The table below outlines the major shortcomings of the following type of normal generator
protection to detect inadvertent energizing
Table 15.1 Performance of conventional relays
Operates Marginal Disabled by Will not detect

too slowly detection of loss of VT 3∅ event
event input
Loss of Field
X X
(40)
Reverse Power (32) X X
Negative
X
Sequence (46)
System Backup
X X X
(21)
Breaker Failure
X*
(50 BF)
* Discussed below
15.4.1 Generator Breaker Failure Protection
Generator breaker failure must be initiated to isolate a generator for an inadvertent

energizing condition due to breaker head flashover. A functional diagram of a typical
generator breaker failure scheme is shown in Figure 15.2.
Figure 15.2: Generator Breaker Failure Logic
(52a, 52b: circuit breaker auxiliary contacts)
When the generator protective relays detect an internal fault or an abnormal condition, they
will send a signal to trip the generator breakers and at the same time initiate the breaker
failure timer. If the breaker(s) do not open within a specified interval of time, the timer (see
figure 15.2) will trip the necessary backup breakers to remove the generator from the
system. The current detector is used to detect whether the breaker has successfully opened.
The breaker auxiliary contact (52a) must also be used in this case since there are faults
and/or abnormal generator conditions which will not produce sufficient current to operate the
current detector. If one or two poles of a breaker flash over to energize a generator, two
conditions must be satisfied to initiate breaker failure.
146
1. The flashover must be detected by a generator protective relay that would initiate the
breaker failure function/relay.
2. The breaker failure current detector must be set with sufficient sensitivity to detect
the flashover condition.
15.5 DEDICATED PROTECTION SCHEMES TO DETECT INADVERTENT ENERGIZING
Due to the severe limitation of conventional generating relaying to detect inadvertent

energizing, dedicated protection schemes have been developed and installed. Unlike
conventional protection schemes, which provide protection when equipment is in-service,
these schemes provide protection when equipment is out of service-. Thus, great care
should be taken when implementing this protection so that dc tripping power and relay input
quantities to the scheme are not removed when the unit being protected is off-line.
The judicious selection of input sources allows most of these schemes to be applied to
generators with low voltage generator breakers. Whatever scheme is used to provide
protection for accidentally energizing a generator, the protection should be connected to trip
the generator high voltage and field breakers, trip the unit auxiliary breakers, initiate
generator high voltage breaker failure backup, and be implemented so that it is not disabled
when the machine is out of service.
Five basic types of schemes are in use today within the industry to provide inadvertent
energizing protection. They are:
• Frequency supervised overcurrent relays

• Voltage supervised overcurrent relays
• Directional overcurrent relays
• Impedance relays
• Breaker auxiliary contact enabled overcurrent relays
15.5.1 Frequency Supervised Overcurrent Relays
CB A CB B
MT
AT CB C
60
81U
50
Figure 15.3: Frequency supervised overcurrent logic
147
Figure 15.3 depicts a frequency-supervised overcurrent scheme specifically designed to
detect accidental energization. The scheme utilizes an under-frequency relay to supervise
the trip output of sensitively set instantaneous overcurrent relays. The overcurrent relays are
automatically armed by the under-frequency relay as the unit is taken off-line and remain
armed while the unit is shut down. To ensure reliable high speed tripping, the overcurrent
relays (50; I>) should be set at 50% or less of the minimum current seen during accidental
energizing. The under-frequency relay (81U; f<) used to identify when the generator is off-
line should have a set point well below any emergency operating frequency. Its output
contacts should also remain closed as the applied voltage goes to zero. The voltage balance
relay (60; ∆U) prevents incorrect operations should the frequency relay lose potential under
normal operating conditions.
15.5.2 Voltage Supervised Overcurrent Relays
Figure 15.4: Voltage supervised overcurrent relays
Figure 15.4 depicts a voltage-supervised overcurrent scheme that is designed to detect

accidental energization. This scheme utilizes under-voltage relays (27-1, 27-2; U<) to
supervise the output of high-speed instantaneous phase overcurrent relays (I>) to provide
inadvertent energizing protection. The overcurrent relays are automatically armed when the
unit is taken off-line and remain armed while the unit is shut down. They are automatically
removed from service when the unit is brought on-line. The overcurrent units are set to
respond to current of 50% or less of minimum current seen during an accidental
energization. Undervoltage relays (27-1, 27-2) enable and disable the overcurrent detectors
(50; I>) through time delay relays. Two undervoltage (U<) relays are supplied from separate
voltage transformers to prevent misoperations that could result from loss of one potential
supply.
Some multifunction numerical generator protection relays use this scheme. Figure 15.5
shows an example. The current and voltage is measured on the generator side. The
undervoltage pickup is approximately 50% of the rated voltage and overcurrent element
picks up at 30%. The settings of the timer are shown in figure 15.5b. The fuse failure monitor
helps avoid a misoperation of the undervoltage relay during running conditions. The behavior
of the implemented logic is discussed next for three different situations.
148
a) Generator at standstill
The undervoltage function (27; U<) picks up and on point A is a “logical 1”. B is “logical 0”. If
a circuit breaker (for example CB A + isolator S) is closed the 27-function drops off, but point
A will be “logical 1” during the drop off delay of the timer (T2 = 1s). In parallel, the
overcurrent element I> picks up and point B becomes “logical 1”. Both inputs at the AND
gate are now at “logical 1” and a trip signal is send to the circuit breakers.
b) Normal start of generator and operation under load
After the generator is started, the undervoltage function U< drops off. After the drop off
delay, “A” becomes “logical 0” and blocks any trip signal. If the generator is on load “B”
becomes “logical 1”, but the AND gate is blocked by the “logical 0” at point A.
c) External short circuit
If there is for example a short circuit in the main transformer (MT) or on the busbar/line the
undervoltage functions picks up. Now the pickup delay of T1 becomes active. Point A will be
“logical 0” for the pickup delay (T1 = 5s). Within this interval of time, the fault must cleared by
the other protection functions. Point B is always at “logical 0”. T1 must greater than the
longest time delay of the other protection functions (line protection included).
a) Connection b) Logic
Figure 15.5: Implementation in a numerical relay
149
15.5.3 Directional Overcurrent Relays
Figure 15.6: Directional overcurrent logic
The scheme depicted in figure 15.6 employs three directional (inverse time)-overcurrent
relays. Voltage and current sensing are obtained from the generator terminals. Two different
methods are used. Method 1 uses a relay having maximum sensitivity when the current
applied to the relay leads the voltage by 30o. To assure that the underexcited loading
capability of the machine is not impaired appreciably, the 60o connection (IA-IB and Uac) is
used.
15.5.4 Impedance Relays
Figure 15.7: Impedance relay logic
There have been a number of schemes developed which use impedance relays located in
the high voltage switchyard which are polarized to 'look into" the machine as shown in figure
15.7. The impedance relay is set to detect the sum of the reactance of the unit step-up
transformer and the machine negative sequence reactance (X1T + X2G) with appropriate
margin. In some cases, the impedance relay is supervised by an instantaneous overcurrent
relay to prevent false operation on loss of potential. Some utilities connect the impedance
relay to trip the high voltage generator breakers and initiate unit shutdown without delay
150
whether the unit is on or off-line. The impedance relay will generally operate for unstable
power swings and requires a thorough stability analysis to ensure that the scheme will not
trip on stable swings. Other utilities choose to enable the scheme for high speed tripping
only when the unit is offline and add time delay to provide security when the unit is online.
15.5.5 Auxiliary Contact Enabled Overcurrent Relays
Figure 15.8 Auxiliary contact enabled overcurrent relay logic
The scheme shown in figure 15.8 uses the generator field breaker auxiliary contact to enable
and disable an overcurrent relay to detect inadvertent energizing when the unit is off-line.
This scheme consists of three non-directional instantaneous current fault detectors which
are armed to trip whenever the generator field breaker is open or racked out. Either of these
conditions will energize a time-delay pickup and drop-out auxiliary relay (62) enabling the
scheme. The overcurrent relays are set at 50% less of the minimum current seen during an
accidental energization.
15.6 DEDICATED PROTECTION SCHEMES TO DETECT GENERATOR BREAKER HEAD

FLASHOVER
For the flashover of a generator high voltage breaker pole, re-tripping of the breaker will not
de-energize the machine. The initiation of breaker failure relaying is required to trip
additional local and possibly remote breakers to de-energize the generator. Some of the
schemes discussed in the preceding section can be set to detect breaker head flashovers
and provide protection in conjunction with generator breaker failure protection. Other
schemes are inoperative when the generator is near rated speed and voltage prior to
synchronizing and must be supplemented by additional protection.
Unbalance currents associated with breaker head flashover will generally cause the
generator negative sequence relay to operate. Breaker failure will be initiated, if the breaker
failure current detectors are set with sufficient sensitivity to detect the situation. Specifically
designed schemes to speed up the detection and isolation of this unique form of breaker
failure are described below.
151
15.6.1 Modified Breaker Failure Scheme
Figure 15.9: Modified breaker failure logic
One approach used to speed up the detection of a breaker flashover is to modify the breaker
failure scheme shown in Figure 15.9. An instantaneous overcurrent relay (50N; IN>) is
connected in the neutral of the generator step-up transformer and is set to respond to an
EHV breaker pole flashover. The relay output is supervised by the generator breaker “52b”
contact providing an additional start to the breaker failure scheme. When the generator
breaker is open and one or two poles of the breaker flash over, the resulting transformer
neutral current is detected by the 50N relay without the delay that would be associated with
negative sequence or some of the previously described inadvertent energizing schemes.
The current detectors associated with the generator breaker failure scheme must be set with
sufficient sensitivity to detect this flashover condition.
15.7 CONCLUSIONS
Inadvertent energizing of synchronous generators has become a significant problem within

the industry in recent years as generating stations have become more complex. The
widespread use of breaker-and-a-half and ring bus schemes has added significant operating
flexibility to high voltage generating stations. These configurations have also increased
complexity and the risk of the generator being inadvertently energized while off-line.
Operating errors, breaker head flashovers, control circuit malfunctions or a combination of
these causes have resulted in generators becoming accidentally energized. Because the
damage to the machine can occur within a few seconds, it must be detected and isolated by
automatic relay action. Although there are relays used as part of the normal complement of
generator protection, their ability to detect inadvertent generator energizing is generally
marginal. These relays are generally disabled at the time when the machine is inadvertently
energized, or operate too slowly to prevent damage to the generator and/or turbine. For
these reasons, major turbine-generator manufacturers have recommended, and many
utilities are installing, dedicated inadvertent energizing protection schemes. The major
schemes in service have been described in this chapter. These schemes vary because the
operating practices and protection philosophies of utilities using them are different.
Protection engineers must asses the risks and determine the impact on protection of their
own company's operating practices prior to deciding which scheme best suits their particular
needs.
152
15.8 REFERENCES
[1] IEEE Power System Relay Working Group Report No. 88SM527-4 “Inadvertent
Energizing Protection of Synchronous Generators" IEEE Transactions on Power
Delivery, Vol. 4, No. 2 April 1989.
[2] IEEE "Guide for A.C. Generator Protection," ANSI/IEEE C37.102-1988
[3] M. Meisinger, G. Rockefeller, L. Schulze, “RAGUA: Protection Against Accidental

Energization of Synchronous Machines," presented to the Pennsylvania Electric
Association Relay Committee, September 14, 1983.
153
16. GENERATOR TRIPPING PRACTICE
16.1 INTRODUCTION
This section deals with the practice of removing a generating unit from service, once an
abnormality or a fault has been detected. A proper choice of the tripping actions to be
performed and of their correct sequence requires a broad knowledge of the generator and its
behaviour during the normal and abnormal running conditions.
To establish a generator protection tripping practice, the following requirements need to be

considered:
16.1.1 Security requirement
As soon as a protection relay detects a fault or an abnormal service condition of the

generating unit, it is mandatory to remove the faulted section from service, in order to
minimise damage.
Tripping circuits must be designed to overcome the failure of a single component. In other
words, a single failure must not prevent the protective system from operating in case of
major danger. This can be achieved by redundancy, i.e., by duplicating the tripping circuits
and devices. Redundancy must be accompanied by continuous supervision.
16.1.2 Selectivity requirement
The second but equally important objective is to restrict damage and disturbance to the
faulted area only, preventing the effect of the fault on the sound part of the plant.
When isolating the unit from the power sources, it is important to evaluate in which manner
the disconnection takes place. In particular the disconnection must be limited to the faulted
parts of the plant, keeping in service, if possible, those components not affected by the fault
which could help the unit to be guided to a soft and controlled stop. This makes service
restoration after fault removal easier.
16.2 TYPE OF TRIPPING SCHEMES
The design of the tripping schemes to be used derives from the requirements listed above.
The protective relays concerning the same part of the plant and activated by faults or
abnormalities requiring the same protective action(s) are grouped together to activate the
same circuit.
In case of electromechanical relay outputs, grouping is made by connecting make contacts

in parallel for tripping devices which must be energised in order to operate, and by
connecting break contacts in series for tripping devices which must be de-energised. For
analogue relay outputs grouping can be made by a pin matrix. For multi-function numerical
relays, outputs of the protection functions are grouped by internal programmable scheme
logic settings.
In each tripping circuit a tripping auxiliary relay must be inserted. It has the function of
multiplying trip contacts and, if it is provided, of latching the tripping action. Instead of a
single grouping for each tripping circuit, several groupings can be used. Trip relays or
contact groupings are provided as many as final tripping devices are.
154
The most common tripping sequences are described below.
16.2.1 Simultaneous tripping
This method, also called total tripping as well as general tripping, involves simultaneous
generator and field breaker tripping, prime mover shut-down and auxiliaries transfer. When a
low voltage (LV) generator breaker is provided, it must also be tripped.
Simultaneous tripping is used for all internal faults which require a quick and unconditional
disconnection of the unit from all power sources. Due to the urgency of the protective action
required, no time delay is allowed, thus the risk of overspeed must be accepted.
16.2.2 Generator tripping
This mode of isolation is equal to simultaneous tripping except for the prime mover, which is
not shut down. It can be used for power system disturbances, rather than for internal faults
or abnormalities, and permits a quick reconnection of the unit to the grid if a quick correction
of the abnormality is possible. If there is a potential of significant overspeed condition (steam
turbines, Francis turbine) or if a complete runback operation following load rejection is
impossible, this mode of tripping must be avoided.
In the units consisting of a LV generator circuit breaker with full interrupting capability and
auxiliary transformer(s) connected between this breaker and the step-up transformer,
generator tripping is used instead of total tripping for faults and abnormalities restricted to
the generator. Isolation is accomplished by tripping the generator LV and the field breakers
and also by shutting down the prime mover. In this case a quick service restoration following
an internal fault/abnormality is not expected. With this practice the high voltage (HV) breaker
remains closed. Thus all the transformers and auxiliary services remain in service, fed by the
HV network.
16.2.3 Unit separation
In this tripping mode only the HV circuit breaker is tripped and the unit auxiliary services
remain connected to the generator. The method is used in case of external faults and
disturbances and is recommended whenever a runback following a load rejection is possible,
in order to have an independent power source for the internal loads and a quick
reconnection of the unit to the system after an abnormality removal. In case of hydro-plants,
and in case of thermal/nuclear plants, the boiler/reactor of which is capable to remain at the
minimum load with no time limits, this tripping method might be very important for a possible
network re-energising in case of black-out.
16.2.4 Sequential tripping
In power plants provided with steam or Francis turbines, when high-speed de-energising of
the generating unit is not required by the protection system, it is recommended to trip first of
all the prime mover, with the generator left excited and connected to the system. This is to
prevent overspeed after the prime mover’s disconnection if steam/water valves are not
completely closed. To ensure that steam/water flow has been reduced below the amount
causing overspeed after generator breaker tripping, a reverse power relay is used. The
tripping logic causes simultaneous tripping following reverse power relay operation and
checking that the valves are closed.
155
To overcome possible relay or valve limit switch failures which can cause no trip output to be
emitted, a second reverse power relay is provided. This can give a signal to the operator and
can provide a backup tripping with a long time delay. The setting value can be either the
same or a bit higher than the first reverse power relay. However, it must be lower, with
adequate margin, than the minimum turbine motoring power.
Instead of using a reverse power relay in the above discussion, a low forward power relay
can also be used to ensure that the operation of protection will not be inhibited by a small
steam leakage through the stop valves of an amount insufficient to produce a dangerous
speed rise.
16.3 SUGGESTED TRIP LOGIC FOR SPECIFIC RELAY FUNCTIONS
The trip logic to be adopted for specific protective functions depends on the plant and outer
system scheme, the type and the behaviour of the prime mover, the fault type and location
and the time foreseen for its removal. In the following Table, possible trip logic is presented.
Notes to Table I:
(1) Alarm only from a low-set element.

(2) For reverse power protection operation, depending on the generator unit scheme, the
tripping logic may be as follows:
• Generator tripping if the LV circuit breaker is provided,
• Simultaneous tripping in case of unit connected generator.
When the protection has an anti-motoring function, tripping might be either immediate
or initiated by breaker tripping and completed by underfrequency and/or overexcitation.
(3) For voltage balance relay operation, an alarm is given and the voltage dependent
relays are blocked.
(4) For breaker failure protection, the circuit breaker is re-tripped (if duplicate trip circuit of
the same breaker is available) or the necessary back-up breakers tripped. An alarm
only is given in case of breaker tripping is not initiated by a protection relay.
(5) For field ground fault protection operation, either alarm only (small machine) or alarm
(high resistance, long delay setting) and trip (low resistance, short delay setting)
according to the manufacturer recommendation.
(6) Only the low-side circuit breaker, when provided.
(7) Only for unit-connected generators.
(8) If a LV breaker is provided, only the HV breaker is tripped.
(9) For the Negative Phase Sequence protection operation, in case both LV and HV
breakers are provided, a decision must be taken on which of them must be tripped
(maintaining auxiliary services fed either from the unit or from the network, depending
on the respective integrity and reliability). For an unbalanced load, a fault in the
network is generally assumed and only the HV breaker is opened.
(10) For a loss of synchronism (or pole-slipping) condition it can be assumed that the fault
was in the network which has now been cleared. Therefore only the LV breaker might
be opened.
(11) When a stator ground fault occurs to an unit-connected generator -
If a LV breaker is provided, another voltage relay, connected to the open delta winding of a
VT connected to the transformer-side bus, causes, with a longer delay, HV breaker tripping
and auxiliary service transfer.
156
Table I Example of a generator Protection trip logic
Device Protection Function 52G 52F Auxil. Turbine Alarm

No Trip Trip Transfer Trip Only
21-1 Generator Distance 1st zone (or 2nd (6) x (7) x
zone and C.B. open)
21-2 Generator distance 2nd zone (8)
21-HV HV distance forward direction (8)
HV distance reverse direction x x x x
24 Generator over-excitation (6) x (7) x (1)
32 Reverse Power (anti-motoring) (6) (2) (2) (2)
40 Loss of field (6) x (7) x
46 Unbalance (NPS) overcurrent (9) (1)
49 Stator thermal monitoring x
51AT Aux transformer neutral overcurrent x x x x
51GN Generator ground overcurrent (6) x (7) x
51TN Transformer neutral Overcurrent x x x x
51V Voltage dependent overcurrent (8)
59 Overvoltage (6) x (7) x (1)
59GN Stator ground fault (6) x (11) x
60 Voltage balance (3)
61 Generator interturn x x x x
62B Breaker failure (4) (4)
63 Transformer oil gas pressure x x x x
64F Field ground fault (6) x (7) x (5)
71 Transformer oil gas level x
78 Loss of synchronism (pole slipping) (10)
81 Frequency protection (8)
87AT Aux. transformer differential x x x x
87G Generator differential (6) x (7) x
87GN Stator ground fault differential (6) x (7) x
87T Unit transformer differential x x x x
87TN Transformer restricted earth fault x x x x
87U Overall differential x x x x
Inadvertent energising x x x x
157
16.4 CONTROL CIRCUIT REDUNDANCY
Redundant relays for the same protective functions require redundant tripping circuits
(paths) and tripping devices (trip coils etc). Redundancy is very important when there is a
possibility of a complete relay testing during service is provided. In this case the relay under
test has to be prevented from tripping and often all the relay of the same tripping path are
disabled. In any case at least one relay for each protective function must remain active on its
own tripping path up to the final protective device. After relay test the continuity of all tripping
paths must be restored and continuously supervised.
Some protective functions have to be disabled from tripping when the unit is out of service.
For this purpose manual disabling switches must be avoided because of the risk of forgetting
to restore the tripping circuit. Automatic disabling is performed by auxiliary contacts directly
coupled with the most significant equipments of the plant whose status is considered a
faithful image of the plant service conditions (water/steam valves, circuit breakers,
disconnect switches, etc). Auxiliary contacts are also used for changing the mode of
operation and/or setting values of some protective functions.
Particular care is required in any case for the choice and installation of these contacts. In
general, redundancy and continuous supervision are adopted, giving alarm and protection
enabling in case of discrepancy.
16.5 IMPACT OF NUMERICAL RELAYING IN THIS AREA
Numerical technology has the following features:
1. it permits protection functions, settings and their outputs (tripping, signaling,

recording etc) to change quickly
2. It performs a continuous testing of all internal circuits and functions (self-
diagnostic)
Software programming of the outputs avoids the use of single trip auxiliary relays with their
contacts, which must be combined together using trip matrix.
Continuous self-diagnosis avoids the use of complicated, unreliable and expensive

additional test apparatus. Periodic testing procedures can be restricted to the final tripping
devices. This test can be performed periodically, when the unit is put out of service for
maintenance.
158
17. POWER PLANT TRANSMISSION GRID INTERACTION
17.1 GENERAL
The power plants are integrated parts of the total power system. The incidents occurring in
the grid will influence the operation of power plants and vice versa. In this chapter the
different phenomena are described and the protection system impact on the consequences
of a variety of events is discussed.
17.2 PHENOMENA IN THE GRID INFLUENCING POWER PLANTS AND REQUIREMENTS FOR
PROTECTION SYSTEM
There are a number of abnormal situations in the grid that will influence the operation of
power plants and increase the risk for damages of the power plant equipments.
17.2.1 Short circuits in the grid
In case of short circuits in the grid, mainly on transmission lines, the voltage will decrease
during the fault period. Since the output power from generation plants is proportional to the
voltage, the active power output will be reduced during the fault. As a consequence, the
generating unit will accelerate and there will be a risk that the generator looses synchronism
with the rest of the grid (fall out of phase). Therefore it is of great importance to limit the fault
clearance time enough to insure transient stability. Consequently, it should be common
practice, at least for line faults, that both basic and back-up fault clearance shall be
performed faster than the critical fault clearance time of power plants.
Another risk is that the fault clearance is not accomplished properly, leading to unselective
tripping of healthy pieces of equipment. As an example, consider the power plant, with two
generating units, connected to the grid as shown in figure 17.1.
Figure 17.1. An example of power plant connected to the HV Substation with two busbars sections
Assume that we get a short circuit at one of the two busbar sections. With proper fault
clearance, the faulted busbar section, along with connected objects, will be tripped
instantaneously after operation of the busbar protection. This is a basic performance that the
power system normally shall be designed to withstand without resulting in a more severe
159
disturbance (black-out). For this fault type, there is a risk that the fault clearance is not
performed correctly. As an example, consider the following scenarios:
a) The busbar protection of the faulted busbar section fails to operate. If there is no
parallel instantaneous busbar protection, the back-up trip will disconnect also the
generating unit connected to the non-faulted busbar section.
b) The sectionalizing breaker between the two busbar sections fails to open. In this
case then, the back-up trip will disconnect also the generating unit connected to
the non-faulted busbar section.
c) The busbar protection of the non-faulted busbar section will give unwanted trip.
This can be the consequence of a faulted busbar protection relay or error in
application of the protection.
There are also other causes such as unselective protection co-ordination, which can cause
disturbances where more than one power plant unit is disconnected from the grid, and thus
might give a more severe disturbance case than considered in operation and planning.
Here are the following requirements for protections that should operate in short circuits
cases:
° the clearance time of all phase to phase short circuits in the grid must be shorter
than 200 ms,
° the differential busbar protection (87B) must be fast and selective. In case the
circuit breaker between two busbar sections is closed, it must be immediately
tripped by the overcurrent protection (50), to allow for busbar protection
selective operation,
° as a main protection for the unit HV line (that connects the unit transformer with
one of the busbar) a differential protection (87L) is required,
° a distance protection (21HV) looking in two directions is required on the HV unit

side as a back-up protection (Fig. 2). It should have four (or more) zones – two
looking toward the unit, and two looking toward the network direction. The 1st
zone as the back up protection for the unit HV line and about 70% of the unit
transformer should operate with a short time delay <100 ms).It should trip both
the unit (QU) and the generator (QG) circuit breaker. The 2nd zone as the unit
transformer and the generator back-up protection should operate with the time
delay. It should trip: the unit (QU), the generator (QG) and the excitation circuit
breakers. The 3rd and the 4th zones serve the purpose of back-up protections to
the network and should operate with a time delay and should trip only the unit
circuit breaker(QU) ,
° the impedance protection (21G), that measures impedance on the generator

terminals, is required as a unit back-up protection. It should operate with a time
delay in two steps. The 1st time delay step should trip the unit circuit breaker
(OU), the 2nd time delay step should trip the generator. The main task of unit
impedance protection is to protect the generator and transformer windings
against overvoltage when short circuits in the network are not eliminated.
Consequently, its time delays are long (1,00 – 1,50 s),
160
87BI
I
II
87BII 50
QU
I>>
21HV
87L Z4<
I Z3<
Z2<
Z1<
51N HV
t
Io>
87L
DI
51N
Y t2
t1
Io>
21G
QG t2
t1
Z<
Figure 17.2. The external short circuits protection system for generator-transformer unit
° the zero sequence overcurrent protection (51NHV) on the HV unit side is required
as back-up to network earth fault protection. It should trip the unit circuit breaker
with a time delay. Its setting should be coordinated with the network earth fault
protections setting,
° the zero sequence overcurrent protection (51N) in the earthed transformer

neutral point is required as a back-up transformer and the network earth fault
protection. It should operate with a time delay, in two steps. The 1st time delay
step should trip the unit circuit breaker (QU), the 2nd time delay step should trip
the generator.
As a prevention to system black-out, the unit auxiliaries should not be tripped, they should
be supplied by the generator [2]. If the system protection fails during a close-in fault, then the
time delay trip of the reverse stage trip following the coordination adjustment may be too
slow. In this situation, the so-called power station de-coupling is applied. It is located either
on the upper or lower voltage side of the unit transformer and processes the three phase
currents and the phase-earth voltages. The applied measuring modules are shown in figure
17.3.
The power station de-coupling picks up when a negative real power jump (-dP/dt) occurs,
the real power drops below a lower threshold (P<), simultaneous undervoltage (U<) and
three-phase overcurrent (I>) are present. The time delay is in the range of 100 up to 140 ms,
so that the total tripping time of will lie between 150 and 200 ms. .
161
Figure 17.3. Power station de-coupling
17.2.2 Unbalance load
The unsymmetrical short circuits, breaks in the HV circuits (i.e. the line conductor break off
or non symmetrical circuit breaker opening or closing) are the main reasons of unbalanced
generator loading. As a consequence, severe generator rotor disturbances can occur. The
generator manufacturer determines two values of permissible unbalanced load – the long-
term and the short-term unbalanced load. The magnitude level of the negative sequence
current (I2) is the criterion to identify unbalanced load by the protection. The long-term
permissible negative sequence current is defined by the following formula:
I*2G=k1 (17.1)
where:
I 2G
I *2 G = – the relative value of current negative sequence,
I nG
InG – the generator nominal current,
k1 – constant depending on the generator construction,
Overcrossing the value (1) should be signaled.
The permissible time of the generator unbalance load is defined by formula:
I *2 2G
t= (17.2)
k2
where:
k2 – constant depending on the generator construction.
The generator protection should operate with time delay defined by equation (2). It should
trip the unit circuit breaker.
For the large units the unbalance protection should be applied both in power station and in
HV substation. The latter should operate with a shorter time delay than the former.
162
17.2.3 Undamped power oscillations (small signal stability)
Different generators in a power system rotate synchronous in relation to each other. The
generators can be said to be connected to each other via an “electric shaft”. If small
disturbances occur in the power system, for example changes in load, the generators will
change their relative phase angle against each other. This can be seen as power oscillations
in the grid. In normal operation such oscillations have good damping and will not create the
risk of further larger disturbances. If the grid is weak, i.e. the impedance between generators
is large, the oscillations tend to have poor damping. Power oscillations with increasing
amplitude can result in loss of synchronization between generators, and thus split of the grid.
The loss of the synchronism between the generator and the system can also occur as a
result of the long short circuit time clearance (longer than 200 ms). It cans occur also as
result of short circuit in a weak system. Here are listed some of the following consequences
of loss of synchronism between the generator and the system:
° higher values of the generator stator windings currents ,
° higher values of torsion torques on the turbo-generator shaft. The destructive
impact of these torques have a cumulative effect and reduce the shaft live. In
extreme conditions, the torsion torque can break he shaft ,
° overheating of the generator stator outermost elements,
° higher values of the unit transformer currents,
° the motors abnormal operation as result of the voltage and frequency instability.
In weak power systems, small signal stability can be improved by additional permanent
regulation in some feedback loops in the power system:
a) Power system stabilizers connected to the AVR of critical generators. The

excitation of the generator will be controlled by the addition of a signal in phase
opposition to the power oscillation, thus giving improved damping.
b) Power system stabilizers connected to the voltage control of SVC units. SVC will
be controlled by the addition of a signal in phase opposition to the power
oscillation, thus giving improved damping.
c) Power system stabilizers connected to the power control of HVDC links. The
HVDC link will be controlled by the addition of a signal in phase opposition to the
power oscillation, thus giving improved damping.
To protect the generator and transformer against the loss of synchronism, out-of-step
protection is required. The criterion of out-of-step protection is the dynamic change of the
impedance measured on the generator terminals. The Figure 17.4 illustrates the trajectory
of the impedance measured on the generator terminals for three-phase short circuit in the
network with clearance time t=500 ms`[1].
163
0,4
0,2
0
-0,6 -0,4 -0,2 0 0,2 0,4 0,6 0,8 1 1,2 1,4
-0,2
Serie1
-0,4
jX
Serie2
Serie3
-0,6
-0,8
-1
-1,2
R
Figure 17.4. The trajectory of the impedance measured on the generator terminals.
Generator nominal power P=200 MW. The clearance time of the three phase short
circuit in the network is t=500 ms.
The out-of-step protection should trip only the unit circuit breaker.
17.2.4 Voltage instability
A limiting factor for active power transfer in many transmission systems is the risk of voltage
instability. The reactive power balance has a significant influence on the maximum active
power transfer to be allowed due to the risk of voltage collapse. The phenomenon is often
described with the so called “nose curve”, showing the voltage at one end of a transmission
line (or parallel lines) as a function of power transfer.
Consider the power system pictured in Figure 17.5.
164
Figure 17.5. An example of power net with generating units (G1, G3)
The voltages on the receiving ends of lines depends on the total active power transfer
through the bottle-neck (Figure 17.6)
Voltage level
Power Transfer
Figure 17.6. The line voltage level as function of power transfer
If any of the transmission lines shown on Figure 17.5 is tripped, the power transfer capability
will be reduced as shown on Figure 17.7.
Voltage level
After line trip
Power Transfer
Figure 17.7. The line voltage level after the line tripping
The reactive power generation of the generators in the receiving ends of the lines is normally
assumed to have unlimited reactive power generation capability. This is however not always
165
true during disturbances where the excitation limiters might be activated and thus change
the curve. If a generator trips the consequences will be further limitation of the maximum
possible power transfer.
Voltage level
Reactive power
generator limitation
After line trip
Power Transfer
Figure 17.8. Influence of the excitation limiters on the voltage level
For the generator protection system the following points are of great importance:
a) The undervoltage protection should not trip “too early”, i.e. should have sufficient
delay so that other actions (for example voltage controlled load shedding) can be
activated during severe situations. The recommended settings of the
undervoltage protection are the following: Upick-up= (80÷85) % of nominal voltage,
t=(7÷10) s. The undervoltage protection should only trip the unit circuit breaker.
As prevention of system black-out, the unit auxiliaries should be not tripped they
should be supplied by the generator.
b) The overexcitation protection should be coordinated to the overexcitation
limitation of the excitation system.
It is also important that the generators be operated with the automatic excitation system in
service so that there are reactive power reserves available to handle grid disturbances.
17. 2.5 Sub-synchronous oscillations
In power systems having series compensated lines and/or HVDC transmission connections,
there is a risk of occurrence of sub-synchronous resonance (SSR). This means that we get
resonance oscillations involving the electrical grid and the generator-turbine shaft of a power
unit. In worst-case scenario, the SSR oscillations will have increasing amplitude that can,
after some time, reach magnitudes that could cause the breaking of the generator shaft.
17.3 PHENOMENA IN POWER PLANTS INFLUENCING THE GRID
17.3.1 Excitation system or Automatic Voltage Regulator malfunction
As discussed above the excitation system of a generator is of paramount importance for

voltage stability in the grid.
Consider the system of figure 17.9 representing a generator connected to an infinite bus
through impedance Ze. Assume that the infinite bus has unity voltage, the impedance is 15
percent and the generator voltage takes the three values 09.5, 1.0, and 1.05 p. u. The circle
diagram of Fig. 17.10 represents the relationship between the real and reactive power at the
166
generator for all three cases. For each case, the circle has the next coordinates for the
centre and next radius value:
Center =(0, E 2t/Z e )

E tEs (17.3)
Radius =
Ze
Fig.17.9. Generator supplying an infinite bus
Examining Fig. 10, one can see that when the generator voltage is above the system
voltage, the generator produces positive VARs. When the generator voltage is equal to the
system voltage, the generator has a power factor close to one (it does not produce any
VARs at all). Finally, when the generator voltage is below the system voltage, the generator
will absorb negative VARs.
This observation indicates that when a generator goes underexcited to the point that the
absorbed negative VARs might get below the generator capability limit, the solution is to
increase the generator output voltage until the absorbed VARs get above the limit. This is
precisely what the AVR will do by way of its associated UEL function (see Chapter 1,
paragraph 1.12.3). The UEL function will produce a positive error signal that will be supplied
to the AVR summing point when it requires the generator AVR to increase the output
voltage. In doing so, the AVR will prevent the generator to be driven below the steady-state
stability limit where it could simply become unstable and lose synchronism with respect to
the grid [3].
8
Et=1.05
C=7.35, R=7
Et=1.00
6 C=6.66, R=6.66
Et=0.95
C=6.017, R=6.33
4
Per Unit Q
GTMC
-2
-1 0 1 2 3 4
Per Unit P
Fig. 17.10. Control of the generator reactive power by its output voltage
When a malfunction exists in the excitation system or when the AVR does not operate
normally, a partial or total loss-of-field condition will develop. The generator will then go
under-excited with the consequence that it could become unstable and at the same time be
the cause of a voltage collapse in the surrounding grid [4].
167
17.4 SHORT CIRCUIT POWER
The function of many protections in the grid is dependent of the short circuit power (source
impedance) in the grid. Both the positive and zero sequence source impedance are
dependent of the generator dispatch, i.e. generator in/out of service. As the power demand
varies from one hour to another and from one season to another the short circuit power will
vary. This is often most pronounced in power systems having a large proportion of hydro
power plants, often situated at long distance from load areas. The following protections are
influenced by the short circuit power:
a) Phase overcurrent protections: The reach is directly dependent on the short

circuit power. Traditionally the settings are chosen to handle both high and low
short circuit power conditions. Modern numerical protections often have different
pre-defined setting groups that can be activated according to the generator
dispatch.
b) Residual overcurrent protections: The reach is directly dependent on the short
circuit power, both positive and zero sequence. Therefore the switching of the
unit transformer, when a generator is taken out of service, is important.
Traditionally the settings are chosen to handle both high and low short circuit
power conditions. Modern numerical protections often have different pre-defined
setting groups that can be activated according to the generator dispatch.
c) Distance protection: Faults at the protected line itself is not significantly
influenced of changes in the short circuit power. In case of faults at other objects
the reach can be significantly changed, thus influencing the remote back-up
protection possibilities.
17.6 REFERENCES
[1] S. Wróblewska, Funkcje automatyki zabezpieczeniowej bloku w stanach awaryjnych

systemu elektroenergetycznego, Energetyka, Zeszyt tematyczny nr 11, Black-out, a
krajowy system elektroenergetyczny, 04/2004
[2] H-J. Herrmann , H. Föhring, H. Kühn, German Practice of Transmission System

Protection, CIGRE B5 Colloquium, 2007 Madrid
[3] Gabriel Benmouyal, “The Impact of Synchronous Generators Excitation Supply on

Protection and Relays”, 34th Annual Western Protective Relay Conference, Spokane,
WA, October 16–18, 2007.
[4] Ran Xu, Gary Kobet and Vaithianathan “Mani” Venkatasubramanian, “Loss of Field
Protection and its Impact on Power System stability”, 35th Annual Western Protective
Relay Conference, Spokane, WA, October 16–18, 2008.
168
18. OSCILLOGRAPHIC MONITORING AND SEQUENCE OF EVENTS
18.1 INTRODUCTION
The main purpose of protection relays installed in an electrical power system is to detect
faults or defects on the grid, and isolate the affected zone as fast as possible so as to keep
the damages to the minimum possible level.
Another important function is the event report. It consists first of the oscillography or the
recording of electrical waveforms before and during the fault and second, the sequence of
events or the chronological recording of all binary signals changes related to the event.
This event report associated with a synchronous generator protection scheme helps in the
following post-mortem analysis:
1. Determine if the protection system has operated properly and evaluate its
security and reliability,
2. Confirm the adequate coordination between the electrical network protection
schemes and the generator protection system,
3. Determine the nature of the fault and its location,
4. Determine the origin and the cause of the event.
An oscillographic record supplies, therefore, very important information about the evolution
of the analog waveforms before and during the event, and about logical signals from control
and protection systems.
18.2 OSCILLOGRAPHIC RECORDER CONFIGURATION
18.2.1 Analog waveforms
The oscillographic function should acquire all the voltage and current waveforms supplied to
the protective system to detect the existence of a fault. Additionally and when the analog
channels number is not limited, waveforms originating from the high voltage side of the main
transformer and auxiliary transformer should also be recorded.
When the oscillographic recorder is a function installed in a digital relay, the recorded analog
waveforms correspond to the waveforms processed by the protective functions. In case of
independent oscillograph equipment, the number and the configuration of the analog
channels depend on the equipment specifications.
18.2.1.1 Current wavefroms
All three generator phase currents should be recorded. The three currents should be
acquired from the CT´s installed in the three phases close to the neutral point. In doing so,
the recording of the phase currents will continue even after the generator tripping. It is well
established that even if the filed breaker has been opened, the fault current could continue to
be supplied, for a few seconds, through the remnant energy in the field.
It is recommended to record currents from the CT´s mounted on the high voltage side of the
main transformer.
169
It is recommended to record the homopolar current from the CT mounted in the HV side of
neutral leg of the usually wye-connected unit transformer.
It is recommended to record currents from the CT´s mounted on the auxiliary transformer (s).
In general, current waveforms should be supplied to the oscillographic equipment at the end
of the secondary current circuits. This allows insuring that the current recorded by the
oscillograph is the same current supplied to the protective relays, even if a loss of insulation
occurs that would short-circuit the secondary circuit.
As shown in figure 1, currents recorded in the middle of the secondary current circuit would
not correspond to the current entering the relays located after the point where the insulation
defect took place. Figure 2 shows the proper way to bring the currents to the oscillograph
after they have been supplied to all the relays existing in the protective scheme.
Figure 18.1 Example of currents improperly recorded in the middle of the secondary circuit
Figure 18.2 Example of currents properly recorded at the end of the secondary circuit
Every current channel should have provision to be selected as an event report trigger with
an independently adjustable threshold level.
18.2.1.2 Voltage waveforms
The generator three phase voltages or phase-to-phase voltages should be recorded.
170
It is recommended to record the homopolar voltage from the VT connected to the loading
resistor on the secondary of the voltage transformer mounted between the neutral of the
stator windings and earth.
Every voltage channel should have provision to be selected as an event report trigger with
an independently adjustable threshold level.
18.2.2 Binary input signals
Oscillographs can usually process a large number of binary inputs. Furthermore, the binary
input channels can accommodate a wide range of DC voltage.
It is recommended to record the following signals:
1. Position of HV electrical equipments and signals from control systems:
• Main breaker position,
• Field breaker position,
• Prime mover state,
• Pumping mode (in case of generator/pump hydro-machines).

2. Tripping and closing signals from generator protection system:
3. Tripping signals from network protection system. This information will help verify
the coordination between the generator and network protection systems.
Every binary input should have provision to be programmed to trigger the event report from
either transition (changing to TRUE or FALSE state).
18.2.3 Triggering inputs
Oscillograph recorder should have an independent triggering input. It could be supplied with
the “ORED” outputs of the relays that detect any fault in the grid. As an example, it could be:
• Low phase voltage (27) measured in HV side of main transformer.
• Overcurrent in neutral (51NT) at wye-connected HV side of the main transformer.
18.2.4 Output signals
The oscillograph recorder should have a number of output signals that should be monitored
by another recorder or protection relay. As an example, we could have:
• Internal fault detector or watch-dog
• Starting of a record.
18.2.5 Auxiliary supply
Oscillograph recorder should be specified to accept a wide enough range of DC or AC power

supply input voltage.
171
Oscillograph recorder should be specified to incorporate an autonomous source of energy
like a battery to insure the following capabilities in case the auxiliary supply is lost:
• keep settings and parameters,
• keep the records in memory,
• keep running the internal clock.
18.2.6 Synchronization
Time reference should be specified to be taken from an accurate timing source like IRIG-B
or GPS signals in order to institute synchronization between oscillograph recorders and grid
protection system equipments.
IRIB-B signal allows synchronisation to the nearest millisecond when GPS signal allows
synchronization to a few microseconds.
18.2.7 Recording
When the oscilograph receives a triggering signal, the recorded file must include pre-fault
and faulted waveforms so as to allow a complete analysis of the protection system response.
Provision should exist within the oscillograph equipment to select the desired response to a
succession of triggering signals. The next possibilities should be selectable:
• ignore a second trigger signal within a specified interval of time
• continue the first record and start a second record immediately after the first one
is over,
• stop the first record and start a new record.
18.3 RECORD MANAGEMENT
18.3.1 Triggering
Event reports should be triggered by:
• exceeding an adjustable threshold in every analog channel,
• change of state with an settable direction (going from FALSE to TRUE state or
from TRUE to FALSE state) for every binary input,
• activation of an independent triggering input.

Oscillograph recorder should have provision to specify the course of action when the
recording memory is full. The next two possibilities should be selectable:
• erase the oldest record and write the new one
• stop recording and generate an alarm displayed on the control panel of the
power plant.
172
18.3.2 Communication and monitoring
Many utilities have a central protection department where network events and disturbances
are analyzed. A remote communication capability with the oscillograph recorder is
recommended in order to speed up the information capture process.
18.3.3 Record information
Every generated record file will have some additional information:
• A power plant identification label for remote identification,
• An oscillograph recorder identification label for remote identification,
• A sequence record number related to date and time of the event ,
• Identification labels of every channel,
• Transformer ratio of CT´s and VT´s in every analogical channel,

The event report should have the capability to be translated to an IEEE/COMTRADE format
compatible with the standard dynamic relay test kits so that the captured waveforms could
be re-played into the protective relays for the purpose of analyzing their performance during
the incidents.
18.4 INDEPENDENT RECORDING EQUIPMENT VS. OSCILLOGRAPH FUNCTION IN DIGITAL

RELAY
Oscillograph recorders were developed during the 50´s for electromecanical and
subsequently solid-state relays when each protective function was imbedded in an
independent and different module. Recording equipment and protective relays were by
definition independent and separate.
Digital technology has changed the nature of event reporting in the sense that the event
report function could be performed by the relay itself.
A new issue was then raised as to whether one should have a separate oscillographic
equipment or use the event report supplied by the relay itself. The next analysis tries to bring
an answer to this question.
18.4.1 Oscillograph function in digital relays

18.4.1.1 Advantages
• Same time reference for all the protection functions performed in the digital relay.
• Recorded waveforms and binary inputs are the same that have been processed by
the protective functions performed by relay.
• Each relay in the protection scheme will provide a redundant event report. The loss
of one relay does not compromise the oscillograph function. .
18.4.1.2 Disadvantages
• Loss of independence between the protection functions and the oscillogarph function.
173
• Oscillograph function inside a relay has very often a limited memory capacity.
• Repetitive triggering during a fault could lose the first records replaced by the last
one.
• Lack of synchronization to GPS in some relays.
• Sampling frequency of analog channels limited by the relay sampling frequency
18.4.2 Independent recorder equipment

18.4.2.1 Advantages
• Independence between protection functions and the recorder. Any damage in a relay
does not compromise to the oscillographic function.
• Technical specifications are usually superior in independent recorder equipment than

in a relay oscillographic function: higher number of analog and digital channels, more
triggering options, higher sampling frequency.
18.4.2.2 Disadvantages
• Additional space needed in the control and protection panel.
• Higher cost and more complicated power station design.
174
Annex A: EXAMPLE OF PROTECTION SETTINGS
CALCULATION
1. TECHNICAL DATA OF UNIT EQUIPMENT
1.1 Generator
Rated power SNG=270,6 MW

Nominal power factor cosϕ=0,85
Nominal voltage UNG=15,75 kV
Rated current ING=9919 A
Reactances:
− Direct-axis reactance xd = 209,5 %
− Direct-axis transient reactance x’d = 34,2 %
− Direct-axis sub transient reactance x”d = 18,7 %
− Negative sequence reactance x2 = 27,6 %
− Zero sequence reactance x0 = 9,8 %
Rated speed n=3000
RMT, 50Hz
Stator winding – earth capacity CG = 0,315 µF/3ph
Long-term permissible negative sequence current I2 = 8%IN
Short-term permissible negative sequence current I22 t ≤ 8s
Rated excitation voltage UfN=353 V
Rated excitation current IfN=2534 A
Short-term permissible rotor overload:
If/IfN 1,1 1,2 1,5 2,0
t[s] 600 240 60 20
1.2 Unit transformer
Rated power SN=270 MV.A

Rated voltage - HV UHV(UT)=126,5 kV
- LV ULV(UT)= 15,75 kV
Vector group Yd11
Short circuit voltage usc(UT)=12,67%
Capacity between HV and LV windings CT = 14,8 nF/3f
Capacity between earth and LV winding CTLV = 65,4 nF/3f
1.3 Generator circuit breaker
Capacity between earth and each c. b. pole: CCB=130 µF
1.4 Auxiliary transformer
Rated power SN=25/12,5/12,5MVA

Rated voltage - HV UHV=15,75±8,7% kV
- LV ULV1=ULV2=ULV(AT)=6,3kV
Short circuit voltage usc(AT)%=12,06/11,88/26,86%
Vector group YyOyO
1-A
1.5 Excitation transformer
Rated power SN=1150 MVA

Rated voltage - HV UHV=15,75 kV
- LV ULV=0,48 kV
Short circuit voltage usc(ET)%=8%
Vector group Yd11
1.6 Phase short circuit power on the 110 kV bus bars

Sscmax= 4 000 MVA
Sscmin= 2400 MVA
2. THE SHORT-CIRCUIT ELECTRICAL QUANTITIES
On figure 4 the short circuit points for which calculations have been made are shown.
They are marked by letters A, B, C, D, E and F.
The values of short-circuit currents and voltages are provided in Table I.1.
The following convention is applicable for the short-circuit identifier subscripts: the
short circuit point comes first, then the unit state and then the kind of short circuit. For
example in the notation IA03:
“A” means the short circuit point,
“O” means the unit state i. e. the unit is disconnected from the system (unit circuit
breaker is open) the generator feeds the auxiliaries,
“3” means the 3 - phase short circuit.
2-A
Figure I.1. Unit generator- transformers; short circuit points
3-A
TABLE I.1. The short circuit currents and voltages
Item The s. c. The s. c. The unit Electrical quantity

point kind work state value Measuring
point
1. Generator 3 – phase The unit is Subtransient s. c. current: 1
terminals discon- 110
nected from I AO 3 = " I NG = 58,35kA
"
(A) the system x d%
The s. c. current after 0,1s time: 1

I A03( 0,1s ) = kI NG = 39,67 kA
k=4,0 – from Is.c.(t) curve for 0,1s and
x”d=18,7%
2. 2 – phase Subtransient s. c. current: 1

3 ⋅110
I " AO 2 = " I NG = 40,77 kA
( x d + x2 )%
The s. c. current after 0,1s time: 1

I " A02
I A02( 0,1s ) = I A03(0,1s ) = 30,34kA
I " A03
3. Unit 3 – phase The unit is Subtransient s. c. current: 1, 2

transformer discon- 110
HV nected from I "B 03 = S NG
I NG
terminals the system ( x "
d + u scUT )%
S NUT
(B) I "B 03 = 34,75kA
The s. c. current after 0,1s time: 1, 2
I BO 3(0,1s ) = kI GN = 25,79kA
S NG
x = ( x" d + u scUT )%
S NUT
4. Unit 2 – phase Subtransient s. c. current: 1,2

transformer
3 ⋅110
HV I "B 02 = I NG
S
terminals ( x"d + x2 + 2 NG u scUT )%
S NUT
(B) I "B 02 = 26,36kA
The s. c. current after 0,1s time: 1,2
I "BO 2
I BO 2(0,1s ) = I BO 3( 0,1s ) = 16,55kA
I "BO 3
4-A
point
5. Unit 1 – phase The unit is The voltage zero sequence on the 1
transformer connected generator side by the s. c. in the unit HV
HV to the circuits (the unit c. b. is closed):
terminals system U0GB01(min) =
UNHV
(B) k⋅ ⋅ CT
= 3 =
1
(CT + CT 2 + 2CCB + CG ) + 2 22
9ω R
= 454V
k=0,6 – zero sequence voltage coefficient
The voltage zero sequence on the unit LV 2

circuits by the s. c. in the unit HV circuits
(the unit c. b. is open)
U0GB01(max) =
UNHV
k⋅ ⋅ CT
= 3 =
1
(CT + CT 2 + CCB) + 2 22
9ω R
= 525V
6. Unit 1 – phase The unit is The s. c. current 5

transformer discon- 3 ⋅ 100 ⋅ I NG
HV nected from I B 01 = = 1337 A
terminals the system ( x1 + x2 + x0 )% ⋅ ϑUT
where:
(B) x1=xd+xUT=134,1 Ω
x2=x2+xUT=23,8 Ω
x0=0,8xUT=6 Ω
ϑUT – unit transformer ratio
u % ⋅ U 2 LV (UT )
xUT = scUT = 7,5Ω
100 ⋅ S NUT
7. Auxiliary 3 – phase The unit is Subtransient s. c. current: 4

transformer discon- 110 ⋅ I NG
LV nected from I "CO 3 = S NG
= 7408 A
terminals the system (x d +
"
x AT )%
S NAT
(C)
The s. c. current after time t = 0,1s:
I CO 3(0,1s ) = kI NG = 6447 A
S NG
x = ( x"d + x AT )%
S NAT
5-A
point
8. 2 – phase The unit is Subtransient s. c. current: 4
discon- 3 ⋅ 110 ⋅ I NG
nected from I "CO 2 = =
S NG
the system ( x d + x2 + 2
"
u scAT )%
S NAT
= 6219 A
The s. c. current after time t = 0,1s:
I "CO 2
I CO 2(0,1s ) = I CO 3( 0,1s ) = 5412 A
I "CO 3
Steady state s. c. current:
3 ⋅100 ⋅ I NG
I C 02 = = 3471A
S
( xd + x2 + 2 NG x AT )%
S NAT
9. 3 - phase The unit is Subtransient s. c. current:: 4

connected 110
to the I "C 3 = I NG = 7900 A
system
X C 3 %
where:
X C3 =
S NG
( xs + xUT ) x"d
S NUT S
= + x AT NG = 138%
S NG S NAT
xs + xUT + x"d
S NUT
10. Unit 3 – phase The unit is Subtransient s. c. current: 3

transformer connected 110 ⋅ U LV (UT )
LV to the I "D 3 = = 6997 A
terminals system 3 ( xs + xUT )ϑUT
(D) where:
U
xs = NG = 0,062Ω - network reactance
S sc max
11. 110 kV 3 – phase The unit is The impedance on generator terminals: 2

Bus bars connected Z = x = 0,116Ω
UT
to the
(E) system
13. Excitation 2 – phase The unit is Steady state s. c. current: 6
discon-
transformer 3 ⋅100 ⋅ I NG
LV nected from I F 02 = = 429 A
S NG
terminals the system ( xd + x2 + 2 xET )%
S NET
(F)
6-A
point
14. 3 - phase The unit is Subtransient s. c. current:: 6
connected 110 ⋅ I NG
to the I "F 3 = = 576 A
system
X F 3 %
where:
( xs + xET ) x"d
X F3 = + xET = 1892%
xs + xET + x"d
7-A
3. PROTECTION SETTINGS
The generator-transformer unit protection system is shown in Figure I.2 (see Figure 4,
Chapter 2).
The setting criteria and the setting values for all protections are given in the Table I.2
for the group A and Table I.3 for the group B. The identification of electrical quantities
correspond to the convention of paragraph 1 and Table I.1.
8-A
Figure I.2 Protection system for unit generator-transformers (P≥200MW)
9-A
TABLE I. 2 Unit protections settings, group A
Item Protection ANSI Protection Setting

number settings criterion Measuring unit Delay
[s]
1. Generator 87G Sensitivity factor (ks) for the 2 –
differential phase s. c. on the generator Idiff.0 = 1,25 A 0
terminals, for the disconnected kr = 0,25
unit from the system, after 0,1 s,
should be ≥ 2
I A02 (0,1s )
ks = (1 − kr ) ≥ 2
I dif . foϑCT
From calculations:
kc = 9,1
I diff .
kr = – restraint coefficient
IN
ϑCT = 2000 – CT’s ratio
(measuring point 1)
2. Unit 87UT Sensitivity coefficient (ks) for the Idiff.o =2,5 A 0

transformer 2 – phase s. c. on the HV kr=0,5
differential transformer terminals for the
disconnected unit from the Other settings depend
system, after 0,1 s, should be ≥ 2 on the relay
construction solution.
The should be taken
I B 02(0,1s ) kr into consideration
ks = (1 − )≥2
I diff .oϑCT 2 following values:
− restraint coefficient
for the 2‘nd current
From calculations: harmonic,
ks = 4,9 − the highest value of
the restraint current,
ϑCT = 2000 – CT’s ratio − the highest value of
differential current,
(measuring point 2) − currents
compensations
coefficients,
− transformer
connection group
compensation.
3. Stator earth 64S Decreasing of stator winding Settings depend on 5,0
fault insulation resistance the relay
(100%)
construction
solution.
10-A
[s]
4. Stator earth 59NG Zero sequence voltage Upick-up=6V 0,3
fault occurrences.
The pick-up voltage:
U OGB 01(min)
U pick −up ≥ kb
ϑu
kb = 1,3 – anti pick-up factor
ϑu = 100 – VT’s ratio
(measuring point 0)
Protection scope: 94% stator

windings
5. Reverse 32R The pick-up reverse active power Ppick-up = 0,005PN 1’step (the
power should be higher than the power turbine fast
losses turbine set (ca. 0.5% PNG) valve is
closed)
t1=2,0
2’nd step
t2=120 – with
the unit that
integrates
reverse
power relay
energizing
6. Unit 21.1 Decreasing of impedance value Zpick-up = 1,8 Ω t1=1,0
impedance measured on the generator terminals. should trip
Impedance relay with operation only the unit
characteristic shown on the Fig. 3.a): circuit
The sensitivity factor ks for the 2 – breaker
phase s.c. on the bars 110 kV should
be ≥ 1,2
21.2 k s Z ≤ Z pick −up ≤ Z ' t2=1,5

should trip
Sensitivity factor : the unit
ks = 1,72
xUT
Z =
ϑz
U min
Z '
= – minimal value of
I max
generator impedance in normal state
Umin=0,95UN
Imax=1,05IN
ϑVT
ϑz = = 0,078 – impedance
ϑCT
ratio,
ϑVT=157,5 – VT’s ratio
(measuring point 2’)
ϑCT=2000 – CT’s ratio
(measuring point 1)
11-A
[s]
6 Unit 21.2 Impedance relay with operation X2=1,2 Ω t2=0,1
Impedance characteristic shown on the R2=1,2 Ω
Fig.3.b):
0,8 ⋅ xUT
X1 =
ϑz
21.1 X1=1,8Ω t1=1,2
1,2 ⋅ xUT R1=1,8 Ω
X2 =
ϑz
7. Generator 46.s Alarm unit should operate when Ipick-up = 0,08ING 10
unbalance the long-term permissible
load negative sequence current is
defined by the following formula
is overcrossed:
I 2G
I *2 G = = k1
I NG
where:
ING – the generator nominal current,
k1=0,08 – constant depending on the
generator construction,
8. Generator 46.1 Tripping unit – 1’st step: t1=f(I2)
unbalance should operate when the k2=8 s
load permissible time of the generator
unbalance load defined by
following formula is overdressed:
I2 2
( ) ⋅ t = k2
I NG
where:
k2 – the constant depending on
the generator construction.
9. 46.2 Tripping unit – 2’nd step: t2 = t1+ 0,6
12-A
[s]
10. Pole slip 78 Dynamic change of impedance XS=1,72 Ω
(out-of-
step)
measured on the generator XUT=1,5 Ω
terminals
X’d=4,74 Ω
The setting values are defined by
following parameters:
Xs – the second value of system
equivalent reactance:
2
U NG
Xs = k
S sc (min)ϑz
k=1,3 – anti pic-up factor
Szw(110kV)=2400 MVA
ϑz = 0,078 – impedance ratio
(see pos. 6)
XUT – the second value of unit
transformer equivalent reactance:
uscUT % ⋅ U NUT ( LV )
2
X UT =
100 ⋅ S NUT ⋅ ϑz
X’d – the second value of
generator direct-axis sub transient
reactance:
x'd % ⋅ U NG
2
X 'd =
100 ⋅ S NG ⋅ ϑz
11. Stator 51G It should operate when the long- Ipick-up = 5,3 A
overload term permissible stator current is 10
overcrossed:
kI NG
I pick −up =
krϑCT
k = 1,05 – stator overload factor
kr = 0,98 – re-setting ratio
(measuring point 1)
12. Under- 81< It should disconnect the unit from Fpick-up= 47,5 Hz 10
frequency the system when the frequency is (57 Hz)
lower than 0,95 of nominal
frequency
13-A
[s]
13. Under- 40 Generator steady state limit For impedance 5,0
excitation criterion. rely:
(Loss of
field)
For impedance relay (operation Xd=−24,5Ω
characteristic – see Fig. 4a) the
X 'd
setting values are defined by = −2,0Ω
following parameters: 2
Xd – the second value of
generator direct-axis reactance:
xd % ⋅ U NG
2
Xd =
100 ⋅ S NG ⋅ ϑz
ϑz = 0,078 – impedance ratio
ϑVT
ϑz =
ϑCT
ϑVT=157,5 – VT’s ratio
ϑCT=2000 – CT’s ratio
(measuring point 2)
X’d – the second value of For admittance
generator direct-axis sub transient relay:
reactance:
10
X 'd x'd % ⋅ U NG
2
λ1=0,48; α1=80o
= λ2=0,43; α2=90o
2 100 ⋅ S NG ⋅ ϑz 0,1
For admittance relay (operation λ3=0,96; α=110o
characteristic – see Fig. 4b) the
setting values are defined by
following parameters:
100
λ1 = ; α1=70o.....90o
xd %
λ2 = (0,8 ÷ 0,9)λ1 ; α2=90o
λ3 = 2λ1 ; α3=110o
14. Over- 59 It should operate when the Upick-up = 122 V 1,8

voltage generator limit voltage value is
overcrossed:
kU NG
U pick −up =
kr ⋅ ϑVT
k = 1,2
ϑVT = 157,5 x – VT’s ratio
15. Unit 24 The criterion of U/f value as a koex = 1,3 3,0
transformer transformer over excitation
Over- factor:
fexcitation
BS U ⋅ f N
(over- koex = =
fluxing) BN U N ⋅ f
14-A
[s]
16. HV unit 51N1 It should operate in the earth fault Ipick-up= 14 A 2,4
circuits case in the unit HV circuits when
earth fault the unit is disconnected from the Should be
system. The sensitivity factor (ks) coordinated
should be ≥ 1,2 with 110 kV
net earth
fault
I B 01
I pick −up ≤ protection
k sϑCT setting
ϑCT=60 – CT’s ratio (measuring
17.
point 5)
51N2
kc =1,5
3,0
18 Auxiliary 50AT It should not operate in 3 – phase Ipick-up=35,0 A 0
transformer s. c. case on the transformer LV
overcurrent terminals.
instantane- The pick-up current:
ous
kI "C 3
I pick −up ≥
ϑCT
k=1,3 – anti-pick up factor
(measuring point 4)
19. Auxiliary 51AT It should operate in 2 phase s. c. Ipick--up = 9,0A 1,0
transformer on the LV transformer terminals
overcurrent case. It should not operate in
motors start –up case.
kr I NAT I
≤ I pick −up ≤ C 02
krϑCT k sϑCT
INAT=920 A – transformer rated
current
kr=2,8 – anti-pick up factor in
motors start-up case
ks=1,2 – sensitivity factor
ϑCT = 300
20. Auxiliary 51ATo It should operate when the long- Ipick-up = 3,6 A 10
transformer term permissible transformer
overcurrent current is overcrossed:
1,05 I AT max
I pick −up =
krϑCT
IATOmax=1,087INAT=1000 A (for
min tap changer position)
ϑCT = 300
15-A
[s]
21. Excitation 50ET It should not operate in 3 – phase Ipick-up=38 A 0
transformer s. c. case on the transformer LV
overcurrent terminals.
instantane- The pick-up current:
ous
kI "F 3
I pick −up ≥ = 38 A
ϑCT
(measuring point 6)
22. Excitation 51ET It should operate in 2 phase s. c. Ipick--up = 9,0A 1,0
transformer on the LV transformer terminals
overcurrent case. It should not operate in the
normal state.
kb I NET I
≤ I pick −up ≤ F 02
krϑCT k sϑCT
INET=42,2 A – transformer rated
current
kb=1,2 – anti-pick up factor in
motors start-up case
ks≥1,2 – sensitivity factor
ϑCT = 20
23. Auxiliary 51ETo It should operate when the long- Ipick-up= 2,3 A 10
transformer term permissible transformer
overcurrent current is overcrossed:
1,05 I NET
I pick −up =
krϑCT
kr=0,98
ϑCT = 20
24. Rotor earth 64R Decreasing of rotor winding Settings depend on 10

construction
solution.
16-A
TABLE I.3. Unit protections settings, group B

[s]
1. Unit 87U Sensitivity coefficient (ks) for Idiff.o =2,5 A 0
differential the 2 – phase s. c. on the HV kr=0,5
transformer terminals for the
system, after 0,1 s, should be on the relay
≥2 construction solution.
The should be taken
into consideration
I B 02(0,1s ) kr
ks = (1 − )≥2 following values:
I diff .oϑCT 2 − restraint coefficient
− the highest value of
ϑCT = 2000 – CT’s ratio differential current,
compensations
coefficients,
− transformer
connection group
compensation.
2. Auxiliary 87AT Sensitivity coefficient (ks) for Idiff.o =2,5 A 0

transformer the 2 – phase s. c. on the LV kr=0,5
differential
transformer terminals for the
system, after 0,1 s, should be on the relay
≥2 construction solution.
The should be taken
into consideration
I C 02(0,1s ) kr
ks = (1 − )≥2 following values:
I diff .oϑCT 2 − restraint coefficient
− the highest value of
ϑCT = 300 – CT’s ratio differential current,
compensations
coefficients,
− transformer
connection group
compensation.
17-A
[s]
3. Unit 50UT It should not operate in 3 – Ir =7,6 A 0
transformer phase
overcurrent
s. c. case on the transformer
LV terminals.
The pick-up current:
kI "D 3
I pick −up ≥
ϑCT
(measuring point 3)
4. Reverse 32R See Table 2 pos. 5
power
5. Unbalance 46.1 See Table 2 pos.8
6. load 46.2
7. Under- 40 See Table 1 pos.13
excitation
(Loss of
field)
8. Under- 27 It should disconnect the unit Upick-up=70 V 10
voltage from the system when the
voltage is lower than 0,7 of
rated voltage The generator
should feed the unit auxiliary
motors so long as possible
9. Inadverted 51 Stator current occurrence Ipick-up = 1,5 A
Engine- when the generator is not
ering
27 excited i.e. the generator Urpick-up =60 V Delay: 2
voltage was lower than 0,6 of Reset: 0,5
rated voltage 0,5 s before this
occurrence
10. Stator earth 59N.1 The pick-up voltage: Upick-up=5V t1= 0,3
fault
U OGB 01(max)
U pick −up ≥ kb
ϑu
kb = 1,3 – anti pick-up factor
59N2 ϑu = 278 – VT’s ratio t2=0,6
(measuring point 2”)

Protection operating scope:
95% stator windings
11. Over- 59.1 See Table 2 pos. 14 Upick-up = 122 V t1 =1,8
voltage
59.2 t2 =2,1
12. Under- 81< It should disconnect the unit Fpick-up= 47,5 Hz (57 10
frequency from the system when the Hz)
frequency is lower than 0,95
of rated frequency. The
generator should feed the unit
auxiliary so long as possible
18-A
[s]
13. Unit 24 See Table 2 pos. 15 koex = 1,3 3,0
transformer
Over-
fexcitation
(over-
fluxing)
14. HV unit 51N1 See Table 2 pos. 16 Ipick-up= 14 A 2,4
circuits
earth fault Should be
coordinated
with 110 kV
net earth
fault
protection
setting
51N2 3,0
15. Excitation 50ET See Table 2 pos. 21 Ipick-up=38 A 0
transformer
overcurrent
instantane-
ous
16. Excitation 51ET See Table 2 – pos. 22 Ipick--up = 9,0A 1,0
transformer
overcurrent
17. Rotor 49R It should be realized characteristic given by generator producer
overload t=f(I)
If/If 1,1 1,2 1,5 2,0
N
t[s] 600 240 60 20
18. Rotor earth 64R Decreasing of rotor winding Settings depend on 10
construction
solution.
19-A
Figure I.3. Examples of unit impedance protection operations characteristics
a) Circle characteristic, b) Rectangular characteristic
Figure I.4. Examples of pole-slip protection operating characteristics

a) Impedance relay, b) Admittance relay
20-A

CIGRE - International Guide On The Protection of Synchronous Generators

Hochgeladen von

Dokumentinformationen

Originaltitel

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

CIGRE - International Guide On The Protection of Synchronous Generators

Hochgeladen von

Copyright:

Verfügbare Formate

479

ISBN: 978- 2- 85873- 170-1

1. Synchronous generators fundamentals

Annex A Example of generator protection settings calculation………………………….. 1-A

1. Synchronous generators fundamentals

1. Synchronous generators fundamentals

This chapter is an introduction to the basic physics principles of synchronous

2. Current state of generator protection

The purpose of this section is to present the state-of-the-art in generator protection

Note 1: Only for hydro generators

Fig.2 Solution with full-redundancy

3. Stator phase fault short circuit protection

4. Stator earth fault protection

5. Rotor earth fault protection

6. Underexcitation (loss-of-field) protection

7. Abnormal frequency operation

8. Breaker failure protection

9. Unbalanced load (negative sequence current) protection

11. Overexcitation (Volts per Hertz) and overvoltage protection

12. Out-of-step protection

13. Anti-motoring protection

14. VT Applications and loss of voltage signals

15. Inadvertent generator energizing

When a generator is energized while off-line on turning gear, or coasting to a stop, it

16. Generator tripping practice

17. Power plant transmission grid interaction

18. Oscillographic monitoring and sequence of events

1.2 TYPES OF SYNCHRONOUS GENERATORS

Figure 1.1 Round-rotor (a) and salient-rotor (b) synchronous generator

1.3 SYNCHRONOUS GENERATOR PRIME MOVERS

1.4 GENERATOR PARAMETERS

1.5 CONNECTIONS OF GENERATORS TO THE POWER NETWORK

Table 1.1 Generator Ratings

Description Symbol Turbo-generator Hydro-generator

Table 1.3 Generator Electrical time constants*

Description Symbol Turbo-generator Hydro-generator

1.6 EARTHING OF SYNCHRONOUS GENERATORS

There are basically 4 methods for earthing a generator:

• low and medium resistance earthing

1.6.1 High-resistance earthing

1.6.2 Low and medium-resistance earthing

1. Medium-resistance earth with a wye-earthed/delta transformer: With this method, a

1.6.4 Other earthing methods

1.7 SYNCHRONOUS GENERATORS DURING A SHORT-CIRCUIT

Expressing the steady state, transient, and subtransient currents as:

we finally get for equation (1.2):

I ac = I s + I ' 0 e − t / T 'd + I "0 e − t / T "d (1.4)

where: α - angle that determines voltage phase in the fault moment

Figure 1.2 Generator three-phase short-circuit phase current

1.8 SYMMETRICAL COMPONENT REPRESENTATION OF GENERATORS

Negative sequence reactance can be measured by applying negative sequence currents to

a) Positive sequence network

b) Negative sequence network

c) Zero sequence network

Figure 1.4 Synchronous generator sequence networks

1.9 GENERATOR STEADY-STATE PHASORS DIAGRAM

Kind of short Short circuit current

3 ph E01 E01 E01

2 ph 3E01 3E01 3E01

Figure 1.5 Idealized synchronous generator

The generator total power is provided:

P = 0.9, Q = 0.434, Ut = 1.0, It = 1.0, xd = 1.0, xq = 0.6

Figure 1.6 Generator steady-state phasor diagram at lagging power factor