20

Grid Protection

Power system protection should be provided for abnormal conditions—

undervoltage, overvoltage, underfrequency, and overfrequency—that
may lead to significant equipment damage and/or grid collapse. Open-
phase and other conditions that should not cause grid collapse are not
addressed in this chapter.
Grid undervoltage can be caused by inadequate or insufficient reac-
tive support (not enough shunt capacitors and synchronous generators),
by sudden load increases, by three-phase faults with delayed clearing, by
loss of multiple facilities within a short period of time, etc. Grid overvolt-
age can be caused by sudden loss of load, by inadequate reactive com-
pensation (no shunt reactors and not enough synchronous generators), by
loss of multiple facilities within a short period of time, etc. Grid under-
frequency occurs when system load exceeds available generation. Grid
underfrequency may be caused by loss of multiple transmission paths, by
loss of multiple generators, etc. Grid overfrequency occurs when available
generation exceeds system load. Grid overfrequency may be caused by
loss of multiple transmission outlets, by sudden loss of load, etc.

20.1 Voltage Issues
Whenever var load exceeds var capability, voltage will sag. If voltage sags
gradually, time-delayed undervoltage relaying can actuate to initiate load
shedding. If voltage drops suddenly, high-speed undervoltage relaying
is needed. The setpoints of undervoltage relays need to be based on the
design objective. If the objective is to protect the power grid from collapse,
relatively high dropout undervoltage trip settings should be established.
If the objective is to protect connected loads from damage, specific calcu-
lations need to be developed for the protected loads.

315
© 2009 by Taylor & Francis Group, LLC
316 Protective Relay Principles

Power Grid Protected Area

345 KV

Undervoltage Relay
138 KV Bus

13.8 KV Bus

Load that
Can Be Shed

Figure 20.1
One-line diagram showing simplified transmission system and traditional undervoltage
relay applications.

20.2 Undervoltage Protection—Grid Protection

The objective is to protect the integrity of the power grid by shedding load
when systemic, persistent undervoltage occurs. The trip setpoint (drop-
out) for undervoltage relays should be based on the minimum acceptable
grid voltage with allowances for setting drift, relay reset, and voltage drop
between the power grid and the location of the undervoltage relay. In
order to minimize the possibility of inadvertent actuation before capacitor
banks switch, or transformer load tap changers actuate, long-time delays
(measured in seconds) are utilized.
Figure  20.1 illustrates the situation where undervoltage relays would
be supplied via voltage transformers connected to the 138-KV bus. When
actuated, the undervoltage relaying scheme would trip 13.8-KV feeder cir-
cuit breakers.
The objective is to protect the grid from sustained undervoltage condi-
tions and the basis for the setting is the minimum scheduled 138-KV volt-
age with allowances.
If the range of 138-KV system operating voltage is 138 KV to 142 KV
during peak load periods and 136 KV to 140 KV during light load peri-
ods—the range of operating voltages should be determined via load flow
analysis that includes automatic transformer tap changer action, automatic

© 2009 by Taylor & Francis Group, LLC

Grid Protection 317

Minimum Reset Voltage: 118.0 Volts, 135.9 KV

Maximum Trip Value: 115.8 Volts, 133.1 KV
Desired Trip Setting: 113.5 Volts, 130.5 KV
Minimum Trip Value: 111.2 Volts, 127.9 KV

Figure 20.2
Undervoltage trip setting, range of actual trip values, and reset voltage.

capacitor switching, etc.—the undervoltage relay setpoint would be select-

ed as follows:

Minimum grid voltage: 136 KV

  VT inaccuracy: 0.5%
  Relay setpoint inaccuracy 1.0%
  Relay setpoint drift 0.5%
  Relay pickup/dropout ratio 2.0%
  Total margins 4.0%
Maximum undervoltage setting: 136 KV × (1 – .04) = 130.56 KV,
94.6 % on 138-KV base

With 138-KV, 120-V VTs, the relay would be set at 113.5 volts, which cor-
responds to 130.5 KV. The pickup/dropout ratio needs to be taken into
account as automatic controls may enable voltage recovery and negate the
need for load shedding. Allowance for errors is application specific.
The desired trip setting, the maximum undervoltage trip value if all
errors combine to increase the actual trip setting, the minimum under-
voltage trip value if all errors combine to decrease the actual trip value,
and the minimum recovery voltage that is required to allow the under-
voltage relay to reset are illustrated in Figure 20.2.

20.3 Undervoltage Protection—Load Protection

The objective is to protect equipment by tripping connected loads when
persistent undervoltage occurs. The trip setpoint for undervoltage relays
should be based on the minimum acceptable load voltage with allowances
for setting drift, and voltage drop between the load and the location of the
undervoltage relay. When this type of protective scheme is applied, trans-
formers equipped with automatic load tap changers cannot be located
between the load and the protective relay. In order to minimize the pos-
sibility of inadvertent actuation before capacitor banks can switch or load
tap changers can actuate, long time delays (measured in seconds) should
be utilized.

© 2009 by Taylor & Francis Group, LLC

318 Protective Relay Principles

Power Grid
345 KV

138 KV Bus

Undervoltage Relay
13.8 KV Bus

Low Voltage Bus

Low Voltage Protected Area

Loads

Figure 20.3
One-line diagram showing multiple voltage transformations.

Figure 20.3 illustrates the situation where the undervoltage relays would

be supplied via voltage transformers connected to the 13.8-KV bus. When
actuated, the undervoltage relaying scheme would trip 13.8-KV feeder cir-
cuit breakers.
The objective is to protect the load from sustained undervoltage and basis for
the setting would be the minimum acceptable load voltage with allowances.
If the low voltage system is a 480-volt system with 460-volt motors that
have a minimum voltage requirement of 414 volts (90% of 460 volts), the
undervoltage relay setpoint would be selected as follows:

VT inaccuracy: 0.5%
Relay setpoint inaccuracy 1.0%
Relay setpoint drift 0.5%
Relay pickup/dropout ratio 2.0%
Voltage drop in motor cables 3.0% (calculated, application specific)
Voltage drop in 13.8-KV, 480-volt transformer 2.0% (calculated, application specific)

© 2009 by Taylor & Francis Group, LLC

Grid Protection 319

Minimum motor voltage, 480-volt base: 414 volts ÷ 480 volts = 86.2%
Minimum 480-V bus voltage: 86.2% + 3% = 89.2%
Minimum transformer 13.8-KV voltage: 89.2% + 2% = 91.2%
Minimum undervoltage setting: 13.8 KV × (91.2% + 4%) = 13.14 KV,
95.3% on 13.8-KV base

With 13.8-KV, 120-V VTs, the undervoltage relay would be set at 109.5
volts, which corresponds to 13.14 KV.
The load protection criteria would be used to determine setpoints when
undervoltage relays are used to initiate load transfers to standby sources.
When undervoltage settings for grid protection are lower than under-
voltage settings for load protection, the settings developed for grid pro-
tection should be applied. If the settings developed for load protection
are applied, inadvertent scheme actuation may occur when grid voltage is
low, but acceptable.

20.4 Overvoltage Protection—Grid Protection

The objective is to protect the integrity of the grid by tripping capacitor
banks and/or energizing shunt reactors when systemic, persistent over-
voltage occurs. The trip setpoint for overvoltage relays should be based
on the maximum acceptable grid voltage with allowances for setting drift
and the location of the overvoltage relay. When this type of protective
scheme is applied, transformers equipped with automatic load tap chang-
ers cannot be located between the load and the protective relay. In order to
minimize the possibility of inadvertent actuation before capacitor banks
can switch, load tap changers can actuate, etc., long time delays (measured
in seconds) should be utilized.
Figure 20.4 illustrates the situation where the overvoltage relays would
be supplied via voltage transformers connected to the 138-KV bus. When
actuated, the overvoltage relaying scheme would trip capacitors and/or
energize shunt reactors.
The objective is to protect the grid from sustained overvoltage condi-
tions and basis for the setting would be the maximum scheduled 138-KV
voltage with allowances.
If the range of 138-KV system operating voltage is 138 KV to 142 KV dur-
ing peak load periods and 136 KV to 140 KV during light load periods, the
overvoltage relay setpoint would be selected as follows:

© 2009 by Taylor & Francis Group, LLC

320 Protective Relay Principles

Power Grid Protected Area

345 KV

Overvoltage Relay
138 KV Bus

13.8 KV Bus

Figure 20.4
One-line diagram showing two voltage transformations and two switched capacitors.

Maximum grid voltage: 142 KV

  VT inaccuracy: 0.5%
  Relay setpoint inaccuracy 1.0%
  Relay setpoint drift 0.5%
  Relay pickup/dropout ratio 2.0%
  Total margins 4.0%
Maximum overvoltage setting: 142 KV × (1 + 0.04) = 147.7 KV

With 138-KV, 120-V VTs, the relay would be set at 128.5 volts, which cor-
responds to 147.8 KV. Allowance for errors is application specific.
Automatic tripping of generators to reduce system voltage is not rec-
ommended unless load flow studies are developed to demonstrate that
system voltage will decrease when generators are tripped off-line. The
concern is that if generators that are absorbing vars are tripped off-line,
system voltage may increase rather than decrease.

20.5 Overvoltage Protection—Load Protection

The objective is to protect equipment by tripping connected loads when
persistent overvoltage occurs. The trip setpoint for overvoltage relays

© 2009 by Taylor & Francis Group, LLC

Grid Protection 321

should be based on the maximum acceptable load voltage with allowances

for setting drift. Little or no allowance should be taken for voltage drop
between the load and the location of the overvoltage relay. Transformers
equipped with automatic load tap changers cannot be located between the
load and the protective relay. In order to minimize the possibility of inad-
vertent actuation before capacitor banks can switch, or load tap changers
can actuate, long time delays (measured in seconds) should be utilized.
Figure 20.5 illustrates the situation where the overvoltage relays would
be supplied via voltage transformers connected to the 13.8-KV bus. When
actuated, the overvoltage relaying scheme would trip all 13.8-KV feeder cir-
cuit breakers as voltage would be expected to increase as load decreases.
The objective is to protect the equipment from sustained overvoltage
conditions and basis for the setting would be the maximum motor or
transformer secondary voltage.
The maximum voltage for a 460-volt motor is 506 volts (110% of 460 volts)
and the maximum voltage for a transformer with a 480-volt secondary
would be 504 volts (105% of 480 volts) at full load and 528 volts (110% of
480 volts) at no load.

Power Grid
345 KV

138 KV Bus

Overvoltage Relay
13.8 KV Bus

Low Voltage Bus

Low Voltage Protected Area

Loads

Figure 20.5
One-line diagram showing multiple voltage transformations.

© 2009 by Taylor & Francis Group, LLC

322 Protective Relay Principles

The overvoltage relay setpoint would be selected as follows:

Maximum transformer voltage: 504 volts, full load 528 volts, no load
  VT inaccuracy: 0.5% 0.5%
  Relay setpoint inaccuracy 1.0% 1.0%
  Relay setpoint drift 0.5% 0.5%
  Voltage drop –2.0% 0%
  Relay pickup/dropout ratio 2.0% 2.0%
  Total margins 2.0% 4.0%

Maximum overvoltage setting:

  Peak load: 13.8 KV × (1 + 0.05) × (1 – 0.02) = 14.20 KV
  No load: 13.8 KV × (1 + 0.1) × (1 – 0.04) = 14.57 KV

With 13.8-KV, 120-V VTs, the relay would be set at 123.5 volts, which cor-
responds to 14.2 KV. Allowance for errors and voltage drop is application
specific.

20.6 Frequency Control
Whenever there is a load-generation mismatch, system frequency can-
not be maintained at 60 Hz. The usual range of system frequency is 59.98
to 60.02 Hz. On a thirty day average, the system frequency should be
60.00 Hz so every time there is a minor underfrequency excursion (0.01
to 0.02 Hz), there will be an intentional, minor, overfrequency correction
period. Similarly every time there is a minor overfrequency excursion,
there will be an intentional, minor underfrequency correction period.
Generator governors respond to frequency changes and automatically
restore system frequency. If, however, system frequency decays substan-
tially, there may not be enough generation on the system to restore system
frequency. Underfrequency relays are then utilized to shed customer load
and restore the balance between load and generation. Likewise, if system
frequency rises substantially, the only alternative may be to trip genera-
tors off-line. Overfrequency relays are utilized as a last-ditch attempt to
shed generators when prolonged overfrequency conditions occur.
If power system frequency drops below 57 Hz, damage to blades of large
steam turbine generators (caused by mechanical resonance) should be
anticipated. For this reason, large steam turbine generators are equipped
with underfrequency relays that trip generators off-line when the power
system frequency decays to 57 Hz for several seconds. This, however, is
a last resort because when generators are tripped off-line as system fre-
quency is decaying, system collapse will occur.

© 2009 by Taylor & Francis Group, LLC

Grid Protection 323

Similarly if the power system frequency rises above 63 Hz, damage

caused by overspeed should be anticipated. For this reason, some genera-
tors may be equipped with overfrequency relays that trip generators off-
line when the system frequency rises above 60 Hz. Overfrequency events
should, however, be a recoverable transient, that is, system collapse should
not occur, if the “proper” generator trips are initiated. If system frequency
is increasing and generators are tripped off-line, the system should stabi-
lize at a frequency close to normal.
Diesel generators and other types of generation are not as sensitive to abnor-
mal frequency operation as large steam turbine generators, but in the United
States large steam turbines are the workhorses of the industry. Therefore, the
system must be designed to protect the large steam turbine generators.
Modeling power system frequency transients is an inexact science. The
reason for this is that the exact sequence of events that will cause underfre-
quency or overfrequency excursions cannot be predicted. If the sequence
could be predicted, corrective action would be initiated.
Underfrequency events are caused by multiple contingencies, such
as line and generator outages, occurring within a short period of time.
Underfrequency conditions may be caused by a shortfall of generation,
that is, customer load exceeds available generation when several large
generators trip within a short period of time, or by the inability to transfer
sufficient power across the power system to meet system demand. The
solution to the problem is to reduce customer load. Therefore, under­
frequency relays are installed in distribution substations. These relays
trip and lockout preselected feeders at each substation. Underfrequency
relays that are located at generating stations trip generators off-line and
ensure power system shutdown.
Overfrequency conditions are caused by a sudden drop in load, that
is, customer load is significantly less than available generation, or by the
inability to transfer generator output power across the power system from
generation centers to load centers. The solution to this problem is to trip
some generators off-line. Therefore, overfrequency relays are installed in
generating stations.
Indicators of power system collapse are voltage and frequency. If trans-
mission system voltage is dropping and system frequency is rising (to as
little as 60.05 Hz), then there is too much generation (watts) and too little
vars on the system. The solution to this problem is to either start generation
at load centers or to shed customer load at load centers. If transmission sys-
tem voltage is dropping and system frequency is also dropping, then there
is insufficient generation on the system. The solution to this problem is to
start additional generation, shed customer load, or initiate brownouts.
Power systems are designed so that they remain stable for transient and
steady-state conditions with numerous contingencies, such as line faults,
line faults with other lines out of service, line faults with stuck break-
ers, bus faults, etc. Power systems are operated so that they remain stable

© 2009 by Taylor & Francis Group, LLC

324 Protective Relay Principles

for transient and steady-state conditions in the event the next worse-case,
single contingency occurs. When equipment failures occur, a grace period
(10–15 minutes) is allowed before the generation mix needs to be adjusted
to cover the next worse-case, single contingency.
Considering the above, it is easy to see that underfrequency and overfre-
quency relay setpoints cannot be determined as exactly as other types of
protective relays. The first hurdle is that the initiating event is undefined.
The second hurdle is that pre-event system conditions are undefined. The
third hurdle is that postevent system conditions are undefined. In special
cases, such as the state of Florida and the city of San Francisco, pre-event
and postevent conditions are better understood, but the initiating event
remains undefined.

20.7 Underfrequency Relaying
Given all of the above listed uncertainties, plus knowledge of past events,
such as the Great Northeast Blackout of 1965, how can underfrequency
setpoints be determined?
One method is to define protective needs based on the limited informa-
tion that is available. For example, the expected drop in system frequency
for loss of generation can be calculated. Assume that the frequency of a
system that has 50,000 MW of generation operating and 50,000 MW of cus-
tomer load is expected to drop 0.2 Hz for the loss of 2200 MW of genera-
tion (loss of two large 1100-MW generators within minutes of each other).
This indicates that the maximum underfrequency setpoint at a distri-
bution substation should be 59.8 Hz minus margins for relay accuracy. If
the underfrequency relay accuracy is ±1%, the maximum underfrequency
relay setpoint should be 59.2 Hz.
If the underfrequency setpoint at generating stations is 57 Hz, then the
minimum underfrequency setpoint at a distribution substation should be
57 Hz plus 1% margin for relay accuracy at generating stations plus 1%
margin for relay accuracy at distribution substations or 58.2 Hz.
If three-step underfrequency relaying is installed, the underfrequency
setpoints at distribution substations could be 59.2 Hz, 58.7 Hz, and 58.2 Hz.
If six-step underfrequency relaying is installed, the underfrequency
setpoints at distribution substations could be 59.2 Hz, 59.0 Hz, 58.8 Hz,
58.6 Hz, 58.4 Hz, and 58.2 Hz.
The next consideration is how much load shedding is necessary. Is 10%
or 5000 MW the proper amount? Would 20% or 10,000 MW be better?
How about 30% or 15,000 MW? There is no absolute answer to this ques-
tion. When setting underfrequency relays, the assumption is that there

© 2009 by Taylor & Francis Group, LLC

Grid Protection 325

61
Frequency, Hertz

60

59
Time, Variable
Design Transient, No Load Shedding

Figure 20.6
Power system frequency, recoverable underfrequency transient, no load shedding.

is insufficient generation (spinning reserve) to automatically correct the

problem via generator governor actuations.
With three-step underfrequency relaying, a good practice seems to be to
shed 10% of the system load for “minor” disturbances of 59.2 Hz; an addi-
tional 10% of system load for “greater” disturbances of 58.7 Hz; and an
additional 10% of system load if system collapse is imminent as evidenced
by a system frequency of 58.2 Hz. Figure  20.6 illustrates a theoretical
underfrequency event and system response for a design basis event with
no load shedding. The frequency dip is not low enough to cause underfre-
quency load shedding. The duration of the recovery transient is a function
of the thermal (load-following) capability of generating units.
Figure  20.7 illustrates a theoretical underfrequency event and system
response for an event with 10% load shedding. Power system frequency
suddenly decreases to 59.0 hertz. After 10% load shedding, power system
frequency recovers to 59.5 hertz. The duration of the recovery transient is
a function of the thermal (load-following) capability of generating units.
Figure  20.8 illustrates a theoretical underfrequency event and system
response for an event with 20% load shedding. Power system frequency
suddenly decreases to 58.0 hertz. After 20% load shedding (10% at 59.2 Hz
and 10% at 58.7 Hz), power system frequency recovers to 59.0 hertz. The
duration of the recovery transient is a function of the thermal (load-fol-
lowing) capability of generating units.
Figure  20.9 illustrates a theoretical underfrequency event and system
response for an event with 30% load shedding. Power system frequency
suddenly decreases to 57.0 hertz. After 30% load shedding (10% at 59.2 Hz,
10% at 58.7 Hz, and 10% at 58.2 Hz), power system frequency recovers to
58.7 hertz. The duration of the recovery transient is a function of the ther-
mal (load-following) capability of generating units.
The consensus among experts is that if the load exceeds generation
by more than 30%, automatic recovery via load shedding will not occur.

© 2009 by Taylor & Francis Group, LLC

326 Protective Relay Principles

60.5

60
Frequency, Hertz

59.5

59

58.5
Time, Variable
Level 1 Transient, 10% Load Shed at 59.2 Hertz

Figure 20.7
Power system frequency, 10% load shed.

60.5

60

59.5
Frequency, Hertz

59

58.5

58

57.5
Time, Variable
Level 2 Transient, 10% Load Shed at 58.7 Hz, 10% at 59.2 Hz

Figure 20.8
Power system frequency, 20% load shed.

Figure  20.10 illustrates a theoretical underfrequency event and system

response for an unrecoverable event with 30% load shedding. Power sys-
tem frequency suddenly decreases to 54.0 hertz. After 30% load shedding
(10% at 59.2 Hz, 10% at 58.7 Hz, and 10% at 58.2 Hz), power system fre-
quency recovers to 57 hertz, but underfrequency relays associated with
some major generating units actuate due to setpoint tolerances. When one

© 2009 by Taylor & Francis Group, LLC

Grid Protection 327

60

59
Frequency

58

57

56
Time, Variable
Level 3 Transient, 10% Load Shed at 58.2 Hz, 58.7 Hz, and 59.2 Hz

Figure 20.9
Power system frequency, 30% load shed.

60

59

58
Frequency, Hertz

57

56

55

54

53

52
Time, Variable
System Collapse

Figure 20.10
Power system frequency, system collapse.

or more major generating units trip off during recovery, recovery will be
unsuccessful and system collapse will occur.
Three-step underfrequency relaying with 10% load shedding is one
alternative. Six-step underfrequency relaying (59.2 Hz, 59.0 Hz, 58.8 Hz,
58.6 Hz, 58.4 Hz, and 58.2 Hz) with 5% load shedding at each frequency
and twelve-step underfrequency relaying (0.1 Hz increments with 2.5%
load shedding at each step) are other alternatives.

© 2009 by Taylor & Francis Group, LLC

328 Protective Relay Principles

Generally underfrequency relays that trip radial feeders are allowed to

operate instantaneously, underfrequency relays that trip networked feed-
ers are delayed 1 or 2 seconds, and underfrequency relays that trip gen-
erators are delayed several seconds. Setpoints for underfrequency relays
associated with distributed generation must be selected using the same
philosophy as for major generating stations.
If too much load is shed when underfrequency relays actuate, system
frequency will rise above 60 Hz and the pendulum can swing to an over-
frequency concern. One alternative is to set the 59.2-Hz relays so that they
trip connected loads in six cycles, set 58.7-Hz relays so that they trip con-
nected loads in 15 cycles, and set 58.2-Hz relays so that they trip connected
loads in six cycles. The 15-cycle delay allows for underfrequency relays
set at 59.2 Hz to actuate, for three-cycle circuit breakers to open, and for
frequency to begin to recover before underfrequency relays set at 58.7 Hz
actuate.
Underfrequency schemes are installed at numerous substations
through­out the power system. They are applied only to lines with non-
critical, nonemergency loads. Only one trip value, that is, either 59.2 Hz,
58.7 Hz, or 58.2 Hz, is used at any one substation. Redundant relays are
used to minimize the possibility of inadvertent scheme actuation due to
setting inaccuracies and other unknowns.
Underfrequency relaying schemes should be used only to trip radial
distribution lines or distribution networks where all network feeds can be
tripped simultaneously. Underfrequency relaying schemes should not be
applied where a significant portion of connected loads are equipped with
automatic load transfer switches that transfer loads to other distribution
lines.
In some power system configurations, motors, acting as generators dur-
ing coastdown, can be a concern. Usually this is only a concern when the
motors are large and the ratio of motor to nonmotor load is high. Two things
can be done to alleviate this concern. First, outputs from two under­
frequency relays (with the same underfrequency trip setting), connected
to separate power sources, can be connected in series so that two under-
frequency relays must actuate simultaneously for underfrequency load
shedding to be initiated at any substation. Second, underfrequency relays
can be supervised by an overcurrent element. This second scheme is
designed so that the underfrequency load shedding scheme actuates only
if load current is above a preset threshold.
Tables 20.1 and 20.2 show the underfrequency set points and time delays
for three-step and six-step underfrequency load shedding schemes. It is
important to remember that these settings are based on concepts rather
than calculations. It is also important to remember that time delays
are arbitrary. The time delays used in these examples were six cycles
for underfrequency relays with no intentional time delay and 15 cycles for
underfrequency relays with intentional time delays.

© 2009 by Taylor & Francis Group, LLC

Grid Protection 329

Table 20.1
Three-Step Underfrequency Load Shedding
Underfrequency Underfrequency
Desired Result Relay Setpoints Scheme Time Delay
Shed 10% customer load 59.2 Hz   6 cycles
Shed 10% customer load 58.7 Hz 15 cycles
Shed 10% customer load 58.2 Hz   6 cycles
Shutdown system 57.6 Hz   5 seconds

Table 20.2
Six-Step Underfrequency Load Shedding
Underfrequency Underfrequency
Desired Result Relay Setpoints Scheme Time Delay
Shed 5% customer load 59.2 Hz   6 cycles
Shed 5% customer load 59.0 Hz 15 cycles
Shed 5% customer load 58.8 Hz   6 cycles
Shed 5% customer load 58.6 Hz 15 cycles
Shed 5% customer load 58.4 Hz   6 cycles
Shed 5% customer load 58.2 Hz 15 cycles
Shutdown system 57.6 Hz   5 seconds

20.8 Overfrequency Relaying
Overfrequency relay settings cannot be determined as easily as underfre-
quency relay settings. The reason for this is that large power systems may
have less than 50 large generators (500 MW and larger) and numerous
small generators. The large generators are base load units and the small
generators are peaking units, cogenerators, independent power produc-
ers, waste recovery generators, etc. Loss of a single large generator can
stress the power system. Loss of two or more large generators will stress
the power system unless the generator tripping can be matched geograph-
ically to system load.
Small generators, while more numerous, may not be on-line when an
overfrequency transient occurs. Nuclear powered generating units are
exempt from overfrequency tripping unless the overfrequency trip was
initiated by the reactor protection system.
With only 50 large generating stations, it is difficult to ensure that over-
frequency tripping of generators will not cause oscillations between over-
frequency and underfrequency conditions. The recommended practice
for setting overfrequency relays is as follows:

© 2009 by Taylor & Francis Group, LLC

330 Protective Relay Principles

1. Determine overspeed limitations of turbine generators and other

equipment sensitive to overspeed in generating stations.
2. Determine overfrequency setpoints of reactor protection systems
at nuclear power plants.
3. Develop regional studies, within an ISO control area or a NERC
region, that include load flows showing the effect of several simul-
taneous generator trips on power system operation. These studies
should demonstrate the continued viability of the power system
after preselected generators trip off-line.

The selected setpoints should ensure that overfrequency limits are not
exceeded, nuclear units do not trip off-line, and the system returns to a
stable configuration.
Many utilities rely on mechanical overspeed devices to trip generators off-
line when the turbine-generator speed rises to unacceptably high levels.
When overfrequency relaying schemes are installed, the preferred alter-
native is to delay overfrequency scheme tripping as long as possible so
that generator governors can act to restore the load-generation balance.
Table  20.3 lists three-step, overfrequency relays setpoints that could
be used to trip 15% of generators off-line in four groups. Each generat-
ing station that participates in the overfrequency trip program would be
equipped with two overfrequency relays set to trip at 61.4 Hz, 62.2 Hz, or
63.0 Hz. Two overfrequency relays are connected in series so that a single
relay misoperation will not cause a generator to trip off-line.
Figure 20.11 shows a design basis overfrequency transient. Figure 20.12
shows a severe overfrequency transient that would likely result in genera-
tor tripping. A severe overfrequency transient would most likely be caused
by loss of transmission lines between remote generators and load centers
that are importing a large portion of the power they are consuming. Most
likely, the load center will experience underfrequency load shedding at
the same time that remote generators are experiencing overfrequency.
Table 20.3
Three-Step Overfrequency Generator Tripping
Overfrequency Group 1 Group 2 Group 3 Group 4
Desired Relay Time Time Time Time
Result Setpoints Delay Delay Delay Delay
Shed 5% 61.4 Hz 15 seconds 30 seconds 45 seconds 60 seconds
generation
Shed 5% 62.2 Hz 5 seconds 10 seconds 15 seconds 20 seconds
generation
Shed 5% 63.0 Hz 1.0 seconds 1.5 seconds 2.0 seconds 2.5 seconds
generation

© 2009 by Taylor & Francis Group, LLC

Grid Protection 331

61

60.5
Frequency, Hertz

60

59.5

59
Time, Variable

Design Transient

Figure 20.11
Power system frequency, recoverable overfrequency transient.

65

64

63
Frequency, Hertz

62

61

60

59
Time, Variable
Severe Transient

Figure 20.12
Power system frequency, system collapse

© 2009 by Taylor & Francis Group, LLC

332 Protective Relay Principles

Figure  20.12 differs from Figures  20.8 and 20.9 in that overfrequency
relay actuations are delayed whereas underfrequency relay actuations are
instantaneous. The reason for the time delays used with overfrequency
relaying is to ensure that the recovery is unlikely without overfrequency
relay operation. If power system frequency rises to 66 Hz, mechanical
overspeed trip devices will actuate instantaneously and may trip numer-
ous generators simultaneously.

© 2009 by Taylor & Francis Group, LLC