Protection & Automation Application Guide

can arise where such a fault does occur. Conventional Stator

generator protection systems would be blind to an interturn A
fault, but the extra cost and complication of providing B
detection of a purely interturn fault is not usually justified. In
C
this case, an interturn fault must develop into an earth fault
before it can be cleared. An exception may be where a
machine has an abnormally complicated or multiple winding
arrangement, where the probability of an interturn fault might
be increased.
Id > Id > Id >
17.4 STATOR WINDING PROTECTION
To respond quickly to a phase fault with damaging heavy Figure 17.4: Stator differential protection
current, sensitive, high-speed differential protection is normally
applied to generators rated in excess of 1MVA. For large 17.5.1 Biased Differential Protection
generating units, fast fault clearance will also maintain stability The relay connections for this form of protection are shown in
of the main power system. The zone of differential protection Figure 17.5(a) and a typical bias characteristic is shown in
can be extended to include an associated step-up transformer. Figure 17.5(b). The differential current threshold setting Is1
For smaller generators, IDMT/instantaneous overcurrent can be set as low as 5% of rated generator current, to provide
protection is usually the only phase fault protection applied. protection for as much of the winding as possible. The bias
Sections 17.5 to 17.8 detail the various methods that are slope break-point threshold setting Is2 would typically be set to
available for stator winding protection. a value above generator rated current, say 120%, to achieve
external fault stability in the event of transient asymmetric CT
17.5 DIFFERENTIAL PROTECTION OF DIRECT saturation. Bias slope K2 setting would typically be set at
CONNECTED GENERATORS 150%.
The theory of circulating current differential protection is I1 I2
discussed fully in Section 10.4.
High-speed phase fault protection is provided, by use of the ∼
connections shown in Figure 17.4. This depicts the derivation
of differential current through CT secondary circuit
connections. This protection may also offer earth fault IA2
protection for some moderate impedance-earthed applications. IB2
Either biased differential or high impedance differential IC2
IA1
techniques can be applied. A subtle difference with modern, IB1
biased, numerical generator protection relays is that they IC1
usually derive the differential currents and biasing currents by
algorithmic calculation, after measurement of the individual CT (a) Relay connections for biased differential protection

secondary currents. In such relay designs, there is full galvanic

separation of the neutral-tail and terminal CT secondary I diff = I1 − I 2
circuits, as indicated in Figure 17.5(a). This is not the case for
the application of high-impedance differential protection. This Operate
difference can impose some special relay design requirements K2
to achieve stability for biased differential protection in some
applications.
K1 Restrain
IS1

IS2 I1 + I 2
I bias =
2
(b) Biased differential operating characteristic
Figure 17.5: Typical generator biased differential protection

17-4
 Chapter 17 ⋅ Generator and Generator Transformer Protection

17.5.2 High Impedance Differential Protection

This differs from biased differential protection by the manner in
∼
which relay stability is achieved for external faults and by the
Rst = Stabilising
fact that the differential current must be attained through the resistor
electrical connections of CT secondary circuits. If the NLR = Non-linear
resistance (Metrosil)
impedance of each relay in Figure 17.4 is high, the event of
one CT becoming saturated by the through fault current NLR NLR NLR
Rst Rst Rst
(leading to a relatively low CT impedance), will allow the V V V
IA2
current from the unsaturated CT to flow mainly through the IB2
IC2
saturated CT rather than through the relay. This provides the
required protection stability where a tuned relay element is Figure 17.7: Relay connections for high impedance differential
protection
employed. In practice, external resistance is added to the relay
circuit to provide the necessary high impedance. The principle To calculate the primary operating current, the following
of high-impedance protection application is illustrated in expression is used:
Figure 17.6, together with a summary of the calculations
I op = N × (I s1 + nI e )
required to determine the value of external stabilising
resistance. where:
Healthy CT Saturated CT Iop = primary operating current
Protected zone
Zm N = CT ratio

RCT1 RCT2 Is1 = relay setting

RL1 If RL3 n = number of CTs in parallel with relay element

Ie = CT magnetising current at VS
Rst
Vr
Is1 is typically set to 5% of generator rated secondary current.
Id >
RL2 RL4
It can be seen from the above that the calculations for the
Voltage across relay circuit
application of high impedance differential protection are more
=Vr I f ( RCT + 2 RL )
complex than for biased differential protection. However, the
protection scheme is actually quite simple and it offers a high
and Vs = KVr
level of stability for through faults and external switching
where 1 . 0 < K ≤ 1 . 5
events. With the advent of multi-function numerical relays
Stabilising resistor, Rst, limits spill current to <Is (relay setting)
Vs and with a desire to dispense with external components, high
R= − RR
st
Is impedance differential protection is not as popular as biased
where RR= relay burden differential protection in modern relaying practice.
Figure 17.6: Principle of high impedance differential protection

In some applications, protection may be required to limit 17.5.3 CT Requirements

voltages across the CT secondary circuits when the differential The CT requirements for differential protection will vary
secondary current for an internal phase fault flows through the according to the relay used. Modern numerical relays may not
high impedance relay circuit(s), but this is not commonly a require CTs specifically designed for differential protection to
requirement for generator differential applications unless very IEC 60044-1 class PX (or BS 3938 class X). However,
high impedance relays are applied. Where necessary, shunt– requirements in respect of CT knee-point voltage will still have
connected, non-linear resistors, should be deployed, as shown to be checked for the specific relays used. High impedance
in Figure 17.7. differential protection may be more onerous in this respect
than biased differential protection. Many factors affect this,
including the other protection functions fed by the CTs and the
knee-point requirements of the particular relay concerned.
Relay manufacturers are able to provide detailed guidance on
this matter.

17-5
Protection & Automation Application Guide

17.6 DIFFERENTIAL PROTECTION OF usually applied to balance the generator differential protection
GENERATOR-TRANSFORMERS and prevent the unit transformer through current being seen
as differential current. An exception might be where the unit
A common connection arrangement for large generators is to
transformer rating is extremely low in relation to the generator
operate the generator and associated step-up transformer as a
rating, e.g. for some hydro applications. The location of the
unit without any intervening circuit breaker. The unit
third set of current transformers is normally on the primary
transformer supplying the generator auxiliaries is tapped off
side of the unit transformer. If located on secondary side of
the connection between generator and step-up transformer.
the unit transformer, they would have to be of an exceptionally
Differential protection can be arranged as follows.
high ratio, or exceptionally high ratio interposing CTs would
have to be used. Thus, the use of secondary side CTs is not to
17.6.1 Generator/Step-up Transformer Differential
be recommended. One advantage is that Unit Transformer
Protection faults would be within the zone of protection of the generator.
The generator stator and step-up transformer can be protected However, the sensitivity of the generator protection to unit
by a single zone of overall differential protection (Figure 17.8). transformer phase faults would be considered inadequate, due
This will be in addition to differential protection applied to the to the relatively low rating of the transformer in relation to that
generator only. The current transformers should be located in of the generator. Thus, the unit transformer should have its
the generator neutral connections and in the transformer HV own differential protection scheme. Protection for the Unit
connections. Alternatively, CTs within the HV switchyard may Transformer is covered in Chapter 16, including methods for
be employed if the distance is not technically prohibitive. Even stabilising the protection against magnetising inrush
where there is a generator circuit breaker, overall differential conditions.
protection can still be provided if desired.
Generator Main transformer 17.7 OVERCURRENT PROTECTION

∼ Overcurrent protection of generators may take two forms.

Plain overcurrent protection may be used as the principal form
Protected zone of protection for small generators, and back-up protection for
larger ones where differential protection is used as the primary
HV busbars
Id > method of generator stator winding protection. Voltage
dependent overcurrent protection may be applied where
differential protection is not justified on larger generators, or
where problems are met in applying plain overcurrent
protection.

Figure 17.8: Overall generator-transformer differential protection

17.7.1 Plain Overcurrent Protection
The current transformers should be rated according to Section It is usual to apply time-delayed plain overcurrent protection to
16.8.2. Since a power transformer is included within the zone generators. For generators rated less than 1MVA, this will
of protection, biased transformer differential protection, with form the principal stator winding protection for phase faults.
magnetising inrush restraint should be applied, as discussed in For larger generators, overcurrent protection can be applied as
Section 16.8.5. Transient overfluxing of the generator remote back-up protection, to disconnect the unit from any
transformer may arise due to overvoltage following generator uncleared external fault. Where there is only one set of
load rejection. In some applications, this may threaten the differential main protection, for a smaller generator, the
stability of the differential protection. In such cases, overcurrent protection will also provide local back-up
consideration should be given to applying protection with protection for the protected plant, in the event that the main
transient overfluxing restraint/blocking (e.g. based on a 5th protection fails to operate. The general principles of setting
harmonic differential current threshold). Protection against overcurrent relays are given in Chapter 9.
sustained overfluxing is covered in Section 17.4.
In the case of a single generator feeding an isolated system,
17.6.2 Unit Transformer Differential Protection current transformers at the neutral end of the machine should
energise the overcurrent protection, to allow a response to
The current taken by the unit transformer must be allowed for
winding fault conditions. Relay characteristics should be
by arranging the generator differential protection as a three-
selected to take into account the fault current decrement
ended scheme. Unit transformer current transformers are
behaviour of the generator, with allowance for the

17-6
 Chapter 17 ⋅ Generator and Generator Transformer Protection

performance of the excitation system and its field-forcing are discussed in the following sections. The choice depends
capability. Without the provision of fault current compounding upon the power system characteristics and level of protection
from generator CTs, an excitation system that is powered from to be provided. Voltage-dependent overcurrent relays are
an excitation transformer at the generator terminals will often found applied to generators used on industrial systems,
exhibit a pronounced fault current decrement for a terminal as an alternative to full differential protection.
fault. With failure to consider this effect, the potential exists
for the initial high fault current to decay to a value below the 17.7.2.1 Voltage Controlled Overcurrent Protection
overcurrent protection pick-up setting before a relay element Voltage controlled overcurrent protection has two time/current
can operate, unless a low current setting and/or time setting is characteristics which are selected according to the status of a
applied. The protection would then fail to trip the generator. generator terminal voltage measuring element. The voltage
The settings chosen must be the best compromise between threshold setting for the switching element is chosen
assured operation in the foregoing circumstances and according to the following criteria.
discrimination with the system protection and passage of
normal load current, but this can be impossible with plain • during overloads, when the system voltage is sustained
overcurrent protection. near normal, the overcurrent protection should have a
current setting above full load current and an operating
In the more usual case of a generator that operates in parallel time characteristic that will prevent the generating
with others and which forms part of an extensive plant from passing current to a remote external fault for
interconnected system, back-up phase fault protection for a a period in excess of the plant short-time withstand
generator and its transformer will be provided by HV limits
overcurrent protection. This will respond to the higher-level
• under close-up fault conditions, the busbar voltage
backfeed from the power system to a unit fault. Other
must fall below the voltage threshold so that the second
generators in parallel would supply this current and, being
protection characteristic will be selected. This
stabilised by the system impedance, it will not suffer a major
characteristic should be set to allow relay operation
decrement. This protection is usually a requirement of the
with fault current decrement for a close-up fault at the
power system operator. Settings must be chosen to prevent
generator terminals or at the HV busbars. The
operation for external faults fed by the generator. It is
protection should also time-grade with external circuit
common for the HV overcurrent protection relay to provide
protection. There may be additional infeeds to an
both time-delayed and instantaneous high-set elements. The
external circuit fault that will assist with grading
time-delayed elements should be set to ensure that the
protected items of plant cannot pass levels of through fault Typical characteristics are shown in Figure 17.9.
current in excess of their short-time withstand limits. The
Current
instantaneous elements should be set above the maximum
pick-up
possible fault current that the generator can supply, but less level
than the system-supplied fault current in the event of a
generator winding fault. This back-up protection will minimise
plant damage in the event of main protection failure for a I>
generating plant fault and instantaneous tripping for an HV-
side fault will aid the recovery of the power system and parallel
generation. KI>

17.7.2 Voltage-Dependent Overcurrent Protection

The plain overcurrent protection setting difficulty referred to in
Vs Voltage
the previous section arises because allowance has to be made
level
both for the decrement of the generator fault current with time
Figure 17.9: Voltage controlled relay characteristics
and for the passage of full load current. To overcome the
difficulty of discrimination, the generator terminal voltage can 17.7.2.2 Voltage Restrained Overcurrent Protection
be measured and used to dynamically modify the basic relay
The alternative technique is to continuously vary the relay
current/time overcurrent characteristic for faults close to the
element pickup setting with generator voltage variation
generating plant. There are two basic alternatives for the
between upper and lower limits. The voltage is said to restrain
application of voltage-dependent overcurrent protection, which
the operation of the current element.

17-7
Protection & Automation Application Guide

The effect is to provide a dynamic I.D.M.T. protection considerations.

characteristic, according to the voltage at the machine
terminals. Alternatively, the relay element may be regarded as 17.8.1.2 Sensitive Earth Fault Protection
an impedance type with a long dependent time delay. In This method is used in the following situations:
consequence, for a given fault condition, the relay continues to
a. direct-connected generators operating in parallel
operate more or less independently of current decrement in the
machine. A typical characteristic is shown in Figure 17.10. b. generators with high-impedance neutral earthing,
the earth fault current being limited to a few tens of
Current amps
pick-up
c. installations where the resistance of the ground
level
fault path is very high, due to the nature of the
ground
I>
In these cases, conventional earth fault protection as described
in Section 17.8.1.1 is of little use.

KI > The principles of sensitive earth fault protection are described

in Sections 9.17.1, 9.18 and 9.19. The earth fault (residual)
current can be obtained from residual connection of line CTs, a
line-connected CBCT, or a CT in the generator neutral. The
Vs2 Vs1 Voltage latter is not possible if directional protection is used. The
level polarising voltage is usually the neutral voltage displacement
Figure 17.10: Voltage restrained relay characteristics input to the relay, or the residual of the three phase voltages,
so a suitable VT must be used. For Petersen Coil earthing, a
17.8 STATOR EARTH FAULT PROTECTION wattmetric technique (Section 9.19) can also be used.
Earth fault protection must be applied where impedance
For direct connected generators operating in parallel,
earthing is employed that limits the earth fault current to less
directional sensitive earth fault protection may be necessary.
than the pick-up threshold of the overcurrent and/or
This is to ensure that a faulted generator will be tripped before
differential protection for a fault located down to the bottom
there is any possibility of the neutral overcurrent protection
5% of the stator winding from the star-point. The type of
tripping a parallel healthy generator. When being driven by
protection required will depend on the method of earthing and
residually-connected phase CTs, the protection must be
connection of the generator to the power system.
stabilised against incorrect tripping with transient spill current
in the event of asymmetric CT saturation when phase fault or
17.8.1 Direct-Connected Generators
magnetising inrush current is being passed. Stabilising
A single direct-connected generator operating on an isolated techniques include the addition of relay circuit impedance
system will normally be directly earthed. However, if several and/or the application of a time delay. Where the required
direct-connected generators are operated in parallel, only one setting of the protection is very low in comparison to the rated
generator is normally earthed at a time. For the unearthed current of the phase CTs, it would be necessary to employ a
generators, a simple measurement of the neutral current is not single CBCT for the earth fault protection to ensure transient
possible, and other methods of protection must be used. The stability.
following sections describe the methods available.
Since any generator in the paralleled group may be earthed, all
17.8.1.1 Neutral Overcurrent Protection generators will require to be fitted with both neutral
overcurrent protection and sensitive directional earth fault
With this form of protection, a current transformer in the
protection. The setting of the sensitive directional earth fault
neutral-earth connection energises an overcurrent relay
protection is chosen to co-ordinate with generator differential
element. This provides unrestricted earth-fault protection and
protection and/or neutral voltage displacement protection to
so it must be graded with feeder protection. The relay element
ensure that 95% of the stator winding is protected. Figure
will therefore have a time-delayed operating characteristic.
17.11 illustrates the complete scheme, including optional
Grading must be carried out in accordance with the principles
blocking signals where difficulties in co-ordinating the
detailed in Chapter 9. The setting should not be more than
generator and downstream feeder earth-fault protection occur.
33% of the maximum earth fault current of the generator, and
a lower setting would be preferable, depending on grading

17-8
 Chapter 17 ⋅ Generator and Generator Transformer Protection

Feeder The relay element must be insensitive to third harmonic

voltages that may be present in the system voltage waveforms,
I > I >>
as these will sum residually.
As the protection is still unrestricted, the voltage setting of the
relay must be greater than the effective setting of any
downstream earth fault protection. It must also be time-
delayed to co-ordinate with such protection. Sometimes, a
second high-set element with short time delay is used to
Ursd > Ursd >
provide fast-acting protection against major winding earth
I > I >
Block* Block* faults. Figure 17.12 illustrates the possible connections that
I > I >
may be used.
~ ~

Open ∼
Re Re

2
* Optional interlocked
Vph-n earth-fault protection
Minimum earth fault level = IF = if grading problems exist
Re
3 1
Figure 17.11: Comprehensive earth-fault protection scheme for direct-
connected generators operating in parallel Va
Vb
For cases (b) and (c) above, it is not necessary to use a Vc
directional facility. Care must be taken to use the correct RCA Vn
setting – for instance if the earthing impedance is mainly 1 Derived from phase neutral voltages
resistive, this should be 0o. On insulated or very high 2 Measured from earth impedance
impedance earthed systems, an RCA of –90o would be used, 3 Measured from broken delta VT
as the earth fault current is predominately capacitive. Figure 17.12: Neutral voltage displacement protection
Directional sensitive earth-fault protection can also be used for
detecting winding earth faults. In this case, the relay element
17.8.2 Indirectly-Connected Generators
is applied to the terminals of the generator and is set to As noted in Section 17.2, a directly-earthed generator-
respond to faults only within the machine windings. Hence transformer unit cannot interchange zero-sequence current
earth faults on the external system do not result in relay with the remainder of the network, and hence an earth fault
operation. However, current flowing from the system into a protection grading problem does not exist. The following
winding earth-fault causes relay operation. It will not operate sections detail the protection methods for the various forms of
on the earthed machine, so that other types of earth-fault impedance earthing of generators.
protection must also be applied. All generators must be so
fitted, since any can be operated as the earthed machine. 17.8.2.1 High Resistance Earthing – Neutral Overcurrent
Protection
17.8.1.3 Neutral Voltage Displacement Protection A current transformer mounted on the neutral-earth conductor
In a balanced network, the addition of the three phase-earth can drive an instantaneous and/or time delayed overcurrent
voltages produces a nominally zero residual voltage, since relay element, as shown in Figure 17.13. It is impossible to
there would be little zero sequence voltage present. Any earth provide protection for the whole of the winding, and Figure
fault will set up a zero sequence system voltage, which will 17.13 also details how the percentage of winding covered can
give rise to a non-zero residual voltage. This can be measured be calculated. For a relay element with an instantaneous
by a suitable relay element. The voltage signal must be derived setting, protection is typically limited to 90% of the winding.
from a VT that is suitable – i.e. it must be capable of This is to ensure that the protection will not maloperate with
transforming zero-sequence voltage, so 3-limb types and those zero sequence current during operation of a primary fuse for a
without a primary earth connection are not suitable. This VT earth fault or with any transient surge currents that could
unbalance voltage provides a means of detecting earth faults. flow through the interwinding capacitance of the step-up

17-9
Protection & Automation Application Guide

transformer for an HV system earth fault.

A time-delayed relay is more secure in this respect, and it may
have a setting to cover 95% of the stator winding. Since the
generating units under consideration are usually large,
instantaneous and time delayed relay elements are often Loading
applied, with settings of 10% and 5% of maximum earth fault resistor
current respectively; this is the optimum compromise in
performance. The portion of the winding left unprotected for
an earth fault is at the neutral end. Since the voltage to earth I>
at this end of the winding is low, the probability of an earth
fault occurring is also low. Hence additional protection is often
not applied.
(a) Protection using a current element
V

a
If Loading
V>
Is resistor

aV
R If =
R
IsR
a min = (b) Protection using a voltage element
V
% covered = (1 − a min) × 100% Figure 17.14: Generator winding earth-fault protection - distribution
transformer earthing

Figure 17.13: Earth fault protection of high-resistance earthed As discussed in Section 17.8.2.1, the protection should be time
generator stator winding using a current element delayed when a sensitive setting is applied, in order to prevent
maloperation under transient conditions. It also must grade
17.8.2.2 Distribution Transformer Earthing Using a with generator VT primary protection (for a VT primary earth
Current Element fault). An operation time in the range 0.5s-3s is usual. Less
In this arrangement, shown in Figure 17.14(a), the generator sensitive instantaneous protection can also be applied to
is earthed via the primary winding of a distribution provide fast tripping for a heavier earth fault condition.
transformer. The secondary winding is fitted with a loading
resistor to limit the earth fault current. An overcurrent relay 17.8.2.3 Distribution Transformer Earthing Using a
element energised from a current transformer connected in the Voltage Element
resistor circuit is used to measure secondary earth fault Earth fault protection can also be provided using a voltage-
current. The relay should have an effective setting equivalent measuring element in the secondary circuit instead. The
to 5% of the maximum earth fault current at rated generator setting considerations would be similar to those for the current
voltage, in order to protect 95% of the stator winding. The operated protection, but transposed to voltage. The circuit
relay element response to third harmonic current should be diagram is shown in Figure 17.14(b).
limited to prevent incorrect operation when a sensitive setting
is applied. Application of both voltage and current operated elements to a
generator with distribution transformer earthing provides some
advantages. The current operated function will continue to
operate in the event of a short-circuited loading resistor and
the voltage protection still functions in the event of an open-
circuited resistor. However, neither scheme will operate in the
event of a flashover on the primary terminals of the
transformer or of the neutral cable between the generator and

17-10
 Chapter 17 ⋅ Generator and Generator Transformer Protection

the transformer during an earth fault. A CT could be added in The initial bias slope is commonly set to 0% to provide
the neutral connection close to the generator, to energise a maximum sensitivity, and applied up to the rated current of
high-set overcurrent element to detect such a fault, but the the generator. It may be increased to counter the effects of CT
fault current would probably be high enough to operate the mismatch. The bias slope above generator rated current is
phase differential protection. typically set to 150% of rated value. The initial current setting
is typically 5% of the minimum earth fault current for a fault at
17.8.2.4 Neutral Voltage Displacement Protection the machine terminals.
This can be applied in the same manner as for direct-
connected generators (Section 17.8.1.3). The only difference 17.8.3.2 High Impedance REF Protection
is that the are no grading problems as the protection is The principle of high impedance differential protection is given
inherently restricted. A sensitive setting can therefore be used, in Chapter 10 and also described further in Section 17.5.2.
enabling cover of up to 95% of the stator winding to be The same technique can be used for earth-fault protection of a
achieved. generator, using three residually connected phase CTs
balanced against a similar single CT in the neutral connection.
17.8.3 Restricted Earth Fault Protection Settings of the order of 5% of maximum earth fault current at
This technique can be used on small generators not fitted with the generator terminals are typical. The usual requirements in
differential protection to provide fast acting earth fault respect of stabilising resistor and non-linear resistor to guard
protection within a defined zone that encompasses the against excessive voltage across the relay must be taken,
generator. It is cheaper than full differential protection but where necessary.
only provides protection against earth faults. The principle is
that used for transformer REF protection, as detailed in 17.8.4 Earth Fault Protection for the Entire Stator
Section 16.7. However, in contrast to transformer REF Winding
protection, both biased low-impedance and high-impedance All of the methods for earth fault protection detailed so far
techniques can be used. leave part of the winding unprotected. In most cases, this is of
no consequence as the probability of a fault occurring in the 5%
17.8.3.1 Low-Impedance Biased REF Protection of the winding nearest the neutral connection is very low, due
This is shown in Figure 17.15. The main advantage is that the to the reduced phase to earth voltage. However, a fault can
neutral CT can also be used in a modern relay to provide occur anywhere along the stator windings in the event of
conventional earth-fault protection and no external resistors insulation failure due to localised heating from a core fault. In
are used. Relay bias is required, as described in Section cases where protection for the entire winding is required,
10.4.2, but the formula for calculating the bias is slightly perhaps for alarm only, there are various methods available.
different and also shown in Figure 17.15.
17.8.4.1 Measurement of Third Harmonic Voltage
Phase CT ratio 1000/1
One method is to measure the internally generated third
∼
Phase A
Phase B
Phase C
harmonic voltage that appears across the earthing impedance
due to the flow of third harmonic currents through the shunt
Neutral CT ratio capacitance of the stator windings etc. When a fault occurs in
200/1
the part of the stator winding nearest the neutral end, the third
harmonic voltage drops to near zero, and hence a relay
element that responds to third harmonic voltage can be used
IA to detect the condition. As the fault location moves
IB progressively away from the neutral end, the drop in third
IC
harmonic voltage from healthy conditions becomes less, so
IN that at around 20-30% of the winding distance, it no longer
becomes possible to discriminate between a healthy and a
I BIAS =
(highest of I A , I B , I C ) + (IN × scaling factor ) faulty winding. Hence, a conventional earth-fault scheme
2
neutral CTratio 200
should be used in conjunction with a third harmonic scheme,
where scaling factor = = = 0.2
phase CTratio 1000 to provide overlapping cover of the entire stator winding. The
IDIFF = I A + I B + I C + ( scaling factor × IN ) measurement of third harmonic voltage can be taken either
Figure 17.15: Low impedance biased REF protection of a generator from a star-point VT or the generator line VT. In the latter

17-11
Protection & Automation Application Guide

case, the VT must be capable of carrying residual flux, and this may cause a sudden rise in terminal voltage due to the
prevents the use of 3-limb types. If the third harmonic voltage trapped field flux and/or overspeed
is measured at the generator star point, an undervoltage
Sudden loss of load should only cause a transient overvoltage
characteristic is used. An overvoltage characteristic is used if
while the voltage regulator and governor act to correct the
the measurement is taken from the generator line VT. For
situation. A maladjusted voltage regulator may trip to manual,
effective application of this form of protection, there should be
maintaining excitation at the value prior to load loss while the
at least 1% third harmonic voltage across the generator neutral
generator supplies little or no load. The terminal voltage will
earthing impedance under all operating conditions.
increase substantially, and in severe cases it would be limited
A problem encountered is that the level of third harmonic only by the saturation characteristic of the generator. A rise in
voltage generated is related to the output of the generator. speed simply compounds the problem. If load that is sensitive
The voltage is low when generator output is low. In order to to overvoltages remains connected, the consequences in terms
avoid maloperation when operating at low power output, the of equipment damage and lost revenue can be severe.
relay element can be inhibited using an overcurrent or power Prolonged overvoltages may also occur on isolated networks,
elements (kW, kVAr or kVA) and internal programmable logic. or ones with weak interconnections, due to the fault conditions
listed earlier.
17.8.4.2 Use of Low-Frequency Voltage Injection
For these reasons, it is prudent to provide power frequency
Another method for protecting the entire stator winding of a overvoltage protection, in the form of a time-delayed element,
generator is to deploy signal injection equipment to inject a low either IDMT or definite time. The time delay should be long
frequency voltage between the stator star point and earth. An enough to prevent operation during normal regulator action,
earth fault at any winding location will result in the flow of a and therefore should take account of the type of AVR fitted and
measurable injection current to cause protection operation. its transient response. Sometimes a high-set element is
This form of protection can provide earth fault protection when provided as well, with a very short definite-time delay or
the generator is at standstill, prior to run-up. It is also an instantaneous setting to provide a rapid trip in extreme
appropriate method to apply to variable speed synchronous circumstances. The usefulness of this is questionable for
machines. Such machines may be employed for variable generators fitted with an excitation system other than a static
speed motoring in pumped-storage generation schemes or for type, because the excitation will decay in accordance with the
starting a large gas turbine prime mover. open-circuit time constant of the field winding. This decay can
last several seconds. The relay element is arranged to trip both
17.9 OVERVOLTAGE PROTECTION the main circuit breaker (if not already open) and the
Overvoltages on a generator may occur due to transient surges excitation; tripping the main circuit breaker alone is not
on the network, or prolonged power frequency overvoltages sufficient.
may arise from a variety of conditions. Surge arrestors may be
required to protect against transient overvoltages, but relay 17.10 UNDERVOLTAGE PROTECTION
protection may be used to protect against power frequency Undervoltage protection was historically rarely fitted to
overvoltages. generators. It is sometimes used as an interlock element for
A sustained overvoltage condition should not occur for a another protection function or scheme, such as field failure
machine with a healthy voltage regulator, but it may be caused protection or inadvertent energisation protection, where the
by the following contingencies: abnormality to be detected leads directly or indirectly to an
undervoltage condition. A transmission system undervoltage
• defective operation of the automatic voltage regulator condition may arise when there is insufficient reactive power
when the machine is in isolated operation generation to maintain the system voltage profile and the
• operation under manual control with the voltage condition must be addressed to avoid the possible
regulator out of service. A sudden variation of the load, phenomenon of system voltage collapse. However, it should
in particular the reactive power component, will give be addressed by the deployment of ’system protection’
rise to a substantial change in voltage because of the schemes. The generation should not be tripped. The greatest
large voltage regulation inherent in a typical alternator case for undervoltage protection being required would be for a
• sudden loss of load (due to tripping of outgoing feeders, generator supplying an isolated power system or to meet the
leaving the set isolated or feeding a very small load) public Utility demands for connection of embedded generation
(see Section 17.21).

17-12
 Chapter 17 ⋅ Generator and Generator Transformer Protection

In the case of generators feeding an isolated system, sensitivity of better than 3% is required, a metering class CT
undervoltage may occur for several reasons, typically
Motoring Power Protection
overloading or failure of the AVR. In some cases, the Prime Mover Possible Damage
Setting
(% of rated)
performance of generator auxiliary plant fed via a unit
Fire/explosion due to
transformer from the generator terminals could be adversely unburnt fuel
affected by prolonged undervoltage. Diesel Engine 5-25
Mechanical damage to
Where undervoltage protection is required, it should comprise gearbox/shafts

an undervoltage element and an associated time delay. 10-15

Settings must be chosen to avoid maloperation during the (split shaft)
Gas Turbine Gearbox damage
inevitable voltage dips during power system fault clearance or >50%
associated with motor starting. Transient reductions in voltage (single shaft) 50% of motoring
power
down to 80% or less may be encountered during motor 0.2-2
starting. (blades out of water)
Hydro Blade and runner cavitation
>2
17.11 LOW FORWARD POWER/REVERSE (blades in water)
POWER PROTECTION Turbine blade damage
Low forward power or reverse power protection may be Steam Turbine 0.5-6 Gearbox damage on geared
required for some generators to protect the prime mover. sets
Parts of the prime mover may not be designed to experience should be employed to avoid incorrect protection behaviour
reverse torque or they may become damaged through due to CT phase angle errors when the generator supplies a
continued rotation after the prime mover has suffered some significant level of reactive power at close to zero power factor.
form of failure.
Table 17.1: Generator reverse power problems

17.11.1 Low Forward Power Protection The reverse power protection should be provided with a
definite time delay on operation to prevent spurious operation
Low forward power protection is often used as an interlocking
with transient power swings that may arise following
function to enable opening of the main circuit breaker for non-
synchronisation or in the event of a power transmission system
urgent trips – e.g. for a stator earth fault on a high-impedance
disturbance.
earthed generator, or when a normal shutdown of a set is
taking place. This is to minimise the risk of plant overspeeding
17.12 UNBALANCED LOADING
when the electrical load is removed from a high-speed
cylindrical rotor generator. The rotor of this type of generator A three-phase balanced load produces a reaction field that, to
is highly stressed mechanically and cannot tolerate much a first approximation, is constant and rotates synchronously
overspeed. While the governor should control overspeed with the rotor field system. Any unbalanced condition can be
conditions, it is not good practice to open the main circuit resolved into positive, negative and zero sequence
breaker simultaneously with tripping of the prime mover for components. The positive sequence component is similar to
non-urgent trips. For a steam turbine, for example, there is a the normal balanced load. The zero sequence component
risk of overspeeding due to energy storage in the trapped produces no main armature reaction.
steam, after steam valve tripping, or in the event that the
steam valve(s) do not fully close for some reason. For urgent 17.12.1 Effect of Negative Sequence Current
trip conditions, such as stator differential protection operation, The negative sequence component is similar to the positive
the risk involved in simultaneous prime mover and generator sequence system, except that the resulting reaction field
breaker tripping must be accepted. rotates in the opposite direction to the d.c. field system.
Hence, a flux is produced which cuts the rotor at twice the
17.11.2 Reverse Power Protection rotational velocity, thereby inducing double frequency currents
Reverse power protection is applied to prevent damage to in the field system and in the rotor body. The resulting eddy-
mechanical plant items in the event of failure of the prime currents are very large and cause severe heating of the rotor.
mover. Table 17.1 gives details of the potential problems for So severe is this effect that a single-phase load equal to the
various prime mover types and the typical settings for reverse normal three-phase rated current can quickly heat the rotor
power protection. For applications where a protection slot wedges to the softening point. They may then be extruded
under centrifugal force until they stand above the rotor surface,

17-13
Protection & Automation Application Guide

when it is possible that they may strike the stator core. 10000

A generator is assigned a continuous negative sequence rating.

For turbo-generators this rating is low; standard values of 10%
1000
and 15% of the generator continuous rating have been
adopted. The lower rating applies when the more intensive
cooling techniques are applied, for example hydrogen-cooling
100
with gas ducts in the rotor to facilitate direct cooling of the
winding.
Short time heating is of interest during system fault conditions 10

Time (sec)
and it is usual in determining the generator negative sequence
withstand capability to assume that the heat dissipation during
such periods is negligible. Using this approximation it is 1

possible to express the heating by the law:

I 22t = K 0.1

where:
I2 = negative sequence component (per unit of machine rated 0.01

current) 0.01 0.1 1 10

Negative sequence current (p.u.)
t = time (seconds)
Indirectly cooled (air) Using I2t model
K = constant proportional to the thermal capacity of the Indirectly cooled (H2) Using true thermal model
generator rotor 350MW direct cooled
660MW direct cooled
For heating over a period of more than a few seconds, it is 1000MW direct cooled
necessary to allow for the heat dissipated. From a
Figure 17.16: Typical negative phase sequence current withstand of
combination of the continuous and short time ratings, the cylindrical rotor generators
overall heating characteristic can be deduced to be:
17.12.2 Protection for Negative Sequence Current
I 1
M= 2 = ( )
This protection is applied to prevent overheating due to
I2R − I 2 R 2t
negative sequence currents. Small salient-pole generators
1− e K
have a proportionately larger negative sequence capacity and
where: may not require protection. Modern numerical relays derive the
negative sequence current level by calculation, with no need
I2R = negative phase sequence continuous rating in per unit of
for special circuits to extract the negative sequence
per unit of machine rated current
component. A true thermal replica approach is often followed,
The heating characteristics of various designs of generator are to allow for:
shown in Figure 17.16.
• standing levels of negative sequence current below the
continuous withstand capability. This has the effect of
shortening the time to reach the critical temperature
after an increase in negative sequence current above
the continuous withstand capability
• cooling effects when negative sequence current levels
are below the continuous withstand capability
The advantage of this approach is that cooling effects are
modelled more accurately, but the disadvantage is that the
tripping characteristic may not follow the withstand
characteristic specified by the manufacturer accurately.
The typical relay element characteristic takes the form of

17-14
 Chapter 17 ⋅ Generator and Generator Transformer Protection

 I  element is used, and gated with a three-phase undervoltage

K
t=− log e 1 −  2 set  element (fed from a VT on the generator side of the circuit
I 22set   I2  breaker) to provide the protection. The overcurrent element
Equation 17.1 can have a low setting, as operation is blocked when the
generator is operating normally. The voltage setting should be
where:
low enough to ensure that operation cannot occur for transient
t = time to trip faults. A setting of about 50% of rated voltage is typical. VT
2
failure can cause maloperation of the protection, so the
 I flc  element should be inhibited under these conditions.
K = Kg ×  
I 
 p  17.14 UNDER/OVERFREQUENCY/
 I flc  OVERFLUXING PROTECTION
I 2 set = I 2 cmr ×  × I
I  n These conditions are grouped together because these problems
 p  often occur due to a departure from synchronous speed.
Kg = negative sequence withstand coefficient (Figure 17.16)
17.14.1 Overfluxing
I2cmr = generator maximum continuous I2 withstand
Overfluxing occurs when the ratio of voltage to frequency is
Iflc = generator rated primary current too high. The iron saturates owing to the high flux density and
Ip = CT primary rated current results in stray flux occurring in components not designed to
carry it. Overheating can then occur, resulting in damage.
In = Relay rated current The problem affects both direct-and indirectly-connected
Figure 17.16 also shows the thermal replica time generators. Either excessive voltage, or low frequency, or a
characteristic described by Equation 17.1, from which it will be combination of both can result in overfluxing, a voltage to
seen that a significant gain in capability is achieved at low frequency ratio in excess of 1.05p.u. normally being indicative
levels of negative sequence current. Such a protection element of this condition. Excessive flux can arise transiently, which is
will also respond to phase-earth and phase-phase faults where not a problem for the generator. For example, a generator can
sufficient negative sequence current arises. Grading with be subjected to a transiently high power frequency voltage, at
downstream power system protection relays is therefore nominal frequency, immediately after full load rejection. Since
required. A definite minimum time setting must be applied to the condition would not be sustained, it only presents a
the negative sequence relay element to ensure correct grading. problem for the stability of the transformer differential
A maximum trip time setting may also be used to ensure protection schemes applied at the power station (see Chapter
correct tripping when the negative sequence current level is 16 for transformer protection). Sustained overfluxing can arise
only slightly in excess of the continuous withstand capability during run up, if excitation is applied too early with the AVR in
and hence the trip time from the thermal model may depart service, or if the generator is run down, with the excitation still
significantly from the rotor withstand limits. applied. Other overfluxing instances have occurred from loss
of the AVR voltage feedback signal, due to a reference VT
17.13 PROTECTION AGAINST INADVERTENT problem. Such sustained conditions must be detected by a
ENERGISATION dedicated overfluxing protection function that will raise an
alarm and possibly force an immediate reduction in excitation.
Accidental energisation of a generator when it is not running
may cause severe damage to it. With the generator at Most AVRs have an overfluxing protection facility included.
standstill, closing the circuit breaker results in the generator This may only be operative when the generator is on open
acting as an induction motor; the field winding (if closed) and circuit, and hence fail to detect overfluxing conditions due to
the rotor solid iron/damper circuits acting as rotor circuits. abnormally low system frequency. However, this facility is not
Very high currents are induced in these rotor components, and engineered to protection relay standards, and should not be
also occur in the stator, with resultant rapid overheating and solely relied upon to provide overfluxing protection. A separate
damage. Protection against this condition is therefore relay element is therefore desirable and provided in most
desirable. modern relays.
A combination of stator undervoltage and overcurrent can be It is usual to provide a definite time-delayed alarm setting and
used to detect this condition. An instantaneous overcurrent an instantaneous or inverse time-delayed trip setting, to match

17-15
Protection & Automation Application Guide

the withstand characteristics of the protected generator and B2 A

transformer. It is very important that the VT reference for F=
8π
overfluxing protection is not the same as that used for the
AVR. where:
A = area
17.14.2 Under/Overfrequency
B = flux density
The governor fitted to the prime mover normally provides
protection against overfrequency. Underfrequency may occur It will be seen from Figure 17.17 that the flux is concentrated
as a result of overload of generators operating on an isolated on one pole but widely dispersed over the other and
system, or a serious fault on the power system that results in a intervening surfaces. The attracting force is in consequence
deficit of generation compared to load. This may occur if a large on one pole but very weak on the opposite one, while flux
grid system suffers a major fault on transmission lines linking on the quadrature axis will produce a balancing force on this
two parts of the system, and the system then splits into two. axis. The result is an unbalanced force that in a large machine
It is likely that one part will have an excess of generation over may be of the order of 50-100 tons. A violent vibration is set
load, and the other will have a corresponding deficit. up that may damage bearing surfaces or even displace the
Frequency will fall fairly rapidly in the latter part, and the rotor by an amount sufficient to cause it to foul the stator.
normal response is load shedding, either by load shedding Field Winding Short Circuit
relays or operator action. However, prime movers may have to
be protected against excessively low frequency by tripping of
the generators concerned.
With some prime movers, operation in narrow frequency
bands that lie close to normal running speed (either above or
below) may only be permitted for short periods, together with
a cumulative lifetime duration of operation in such frequency
bands. This typically occurs due to the presence of rotor
torsional frequencies in such frequency bands. In such cases,
monitoring of the period of time spent in these frequency
bands is required. A special relay is fitted in such cases,
arranged to provide alarm and trip facilities if either an
individual or cumulative period exceeds a set time.

17.15 ROTOR FAULTS Figure 17.17: Flux distribution on rotor with partial winding short
circuit
The field circuit of a generator, comprising the field winding of
the generator and the armature of the exciter, together with
17.15.1 Rotor Earth-Fault Protection
any associated field circuit breaker if it exists, is an isolated d.c.
circuit which is not normally earthed. If an earth fault occurs, Two methods are available to detect this type of fault. The first
there will be no steady-state fault current and the need for method is suitable for generators that incorporate brushes in
action will not be evident. the main generator field winding. The second method requires
at least a slip-ring connection to the field circuit:
Danger arises if a second earth fault occurs at a separate point
in the field system, to cause the high field current to be • Potentiometer method
diverted, in part at least, from the intervening turns. Serious • A.C. injection method
damage to the conductors and possibly the rotor can occur
very rapidly under these conditions. 17.15.1.1 Potentiometer Method
More damage may be caused mechanically. If a large portion This is a scheme that was fitted to older generators, and it is
of the winding is short-circuited, the flux may adopt a pattern illustrated in Figure 17.18. An earth fault on the field winding
such as that shown in Figure 17.17. The attracting force at would produce a voltage across the relay, the maximum
the surface of the rotor is given by: voltage occurring for faults at the ends of the winding.
A ‘blind spot' would exist at the centre of the field winding. To
avoid a fault at this location remaining undetected, the tapping

17-16
 Chapter 17 ⋅ Generator and Generator Transformer Protection

point on the potentiometer could be varied by a pushbutton or

switch. The relay setting is typically about 5% of the exciter
Generator
voltage. field Exciter
winding

Field
winding I > Exciter L.F. injection
supply ∼
∼
∼ V>

Figure 17.18: Earth fault protection of field circuit by potentiometer

method
(a) Low frequency a.c. voltage injection - current measurement
17.15.1.2 Injection Methods
Two methods are in common use. The first is based on low
frequency signal injection, with series filtering, as shown in
Generator
Figure 17.19(a). It comprises an injection source that is field Exciter
connected between earth and one side of the field circuit, winding

through capacitive coupling and the measurement circuit. The

field circuit is subjected to an alternating potential at
substantially the same level throughout. An earth fault
anywhere in the field system will give rise to a current that is Injection
supply ∼
detected as an equivalent voltage across the adjustable resistor Z<
by the relay. The capacitive coupling blocks the normal d.c.
field voltage, preventing the discharge of a large direct current
through the protection scheme. The combination of series
capacitor and reactor forms a low-pass tuned circuit, the
intention being to filter higher frequency rotor currents that (b) Power frequency a.c. voltage injection - impedance measurement
may occur for a variety of reasons. Figure 17.19: Earth fault protection of field circuit by a.c. injection

Other schemes are based on power frequency signal injection.

17.15.2 Rotor Earth Fault Protection for Brushless
An impedance relay element is used, a field winding earth fault
reducing the impedance seen by the relay. These suffer the Generators
draw back of being susceptible to static excitation system A brushless generator has an excitation system consisting of:
harmonic currents when there is significant field winding and
• a main exciter with rotating armature and stationary
excitation system shunt capacitance. Greater immunity for
field windings
such systems is offered by capacitively coupling the protection
scheme to both ends of the field winding, where brush or slip • a rotating rectifier assembly, carried on the main shaft
ring access is possible (Figure 17.19(b)). line out
• a controlled rectifier producing the d.c. field voltage for
The low–frequency injection scheme is also advantageous in
the main exciter field from an a.c. source (often a small
that the current flow through the field winding shunt
‘pilot’ exciter)
capacitance will be lower than for a power frequency scheme.
Such current would flow through the machine bearings to Hence, no brushes are required in the generator field circuit.
cause erosion of the bearing surface. For power frequency All control is carried out in the field circuit of the main exciter.
schemes, a solution is to insulate the bearings and provide an Detection of a rotor circuit earth fault is still necessary, but this
earthing brush for the shaft. must be based on a dedicated rotor-mounted system that has
a telemetry link to provide an alarm/data.

17.15.3 Rotor Shorted Turn Protection

As detailed in Section 17.15, a shorted section of field winding
will result in an unsymmetrical rotor flux pattern and in

17-17
Protection & Automation Application Guide

potentially damaging rotor vibration. Detection of such an instantaneously as it flows in a highly inductive circuit.
electrical fault is possible using a probe consisting of a coil Consequently, the flux energy must be dissipated to prevent an
placed in the airgap. The flux pattern of the positive and excessive inductive voltage rise in the field circuit. For
negative poles is measured and any significant difference in machines of moderate size, it is satisfactory to open the field
flux pattern between the poles is indicative of a shorted turn or circuit with an air-break circuit breaker without arc blow-out
turns. Automated waveform comparison techniques can be coils. Such a breaker permits only a moderate arc voltage,
used to provide a protection scheme, or the waveform can be which is nevertheless high enough to suppress the field current
inspected visually at regular intervals. An immediate fairly rapidly. The inductive energy is dissipated partly in the
shutdown is not normally required unless the effects of the arc and partly in eddy-currents in the rotor core and damper
fault are severe. The fault can be kept under observation until windings.
a suitable shutdown for repair can be arranged. Repair will
With generators above about 5MVA rating, it is better to
take some time, since it means unthreading the rotor and
provide a more definite means of absorbing the energy without
dismantling the winding.
incurring damage. Connecting a ‘field discharge resistor’ in
Since short-circuited turns on the rotor may cause damaging parallel with the rotor winding before opening the field circuit
vibration and the detection of field faults for all degrees of breaker will achieve this objective. The resistor, which may
abnormality is difficult, the provision of a vibration a detection have a resistance value of approximately five times the rotor
scheme is desirable – this forms part of the mechanical winding resistance, is connected by an auxiliary contact on the
protection of the generator. field circuit breaker. The breaker duty is thereby reduced to
that of opening a circuit with a low L/R ratio. After the breaker
17.15.4 Protection Against Diode Failure has opened, the field current flows through the discharge
A short-circuited diode will produce an a.c. ripple in the exciter resistance and dies down harmlessly. The use of a fairly high
field circuit. This can be detected by a relay monitoring the value of discharge resistance reduces the field time constant to
current in the exciter field circuit, however such systems have an acceptably low value, though it may still be more than one
proved to be unreliable. The relay would need to be time second. Alternatively, generators fitted with static excitation
delayed to prevent an alarm being issued with normal field systems may temporarily invert the applied field voltage to
forcing during a power system fault. A delay of 5-10 seconds reduce excitation current rapidly to zero before the excitation
may be necessary. system is tripped.

Fuses to disconnect the faulty diode after failure may be fitted. 17.16 LOSS OF EXCITATION PROTECTION
The fuses are of the indicating type, and an inspection window
Loss of excitation may occur for a variety of reasons. If the
can be fitted over the diode wheel to enable diode health to be
generator was initially operating at only 20%-30% of rated
monitored manually.
power, it may settle to run super-synchronously as an
A diode that fails open-circuit occurs less often. If there is induction generator, at a low level of slip. In doing so, it will
more than one diode in parallel for each arm of the diode draw reactive current from the power system for rotor
bridge, the only impact is to restrict the maximum continuous excitation. This form of response is particularly true of salient
excitation possible. If only a single diode per bridge arm is pole generators. In these circumstances, the generator may be
fitted, some ripple will be present on the main field supply but able to run for several minutes without requiring to be tripped.
the inductance of the circuit will smooth this to a degree and There may be sufficient time for remedial action to restore the
again the main effect is to restrict the maximum continuous excitation, but the reactive power demand of the machine
excitation. The set can be kept running until a convenient during the failure may severely depress the power system
shutdown can be arranged. voltage to an unacceptable level. For operation at high initial
power output, the rotor speed may rise to approximately 105%
17.15.5 Field Suppression of rated speed, where there would be low power output and
The need to rapidly suppress the field of a machine in which a where a high reactive current of up to 2.0p.u. may be drawn
fault has developed should be obvious, because as long as the from the supply. Rapid automatic disconnection is then
excitation is maintained, the machine will feed its own fault required to protect the stator windings from excessive current
even though isolated from the power system. Any delay in the and to protect the rotor from damage caused by induced slip
decay of rotor flux will extend the fault damage. Braking the frequency currents.
rotor is no solution, because of its large kinetic energy.
The field winding current cannot be interrupted

17-18
 Chapter 17 ⋅ Generator and Generator Transformer Protection

17.16.1 Protection against Loss of Excitation XG XT ZS

EG ES
The protection used varies according to the size of generator
being protected. A

+jX
17.16.1.1 Small Generators ( ≤ 5MVA) D

On the smaller machines, protection against asynchronous ZS

XG+XT+ZS
running has tended to be optional, but it may now be available
by default, where the functionality is available within a modern
numerical generator protection package. If fitted, it is EG
=1 XT
arranged either to provide an alarm or to trip the generator. If ES
the generator field current can be measured, a relay element θ
ZR
can be arranged to operate when this drops below a preset -R +R
A
value. However, depending on the generator design and size
relative to the system, it may well be that the machine would
XG
be required to operate synchronously with little or no excitation
under certain system conditions.
The field undercurrent relay must have a setting below the C
minimum exciting current, which may be 8% of that
corresponding to the MCR of the machine. Time delay relays
are used to stabilise the protection against maloperation in -jX
response to transient conditions and to ensure that field Figure 17.20: Basic interconnected system
current fluctuations due to pole slipping do not cause the
A relay to detect loss of synchronism can be located at point A.
protection to reset.
It can be shown that the impedance presented to the relay
If the generator field current is not measurable, then the under loss of synchronism conditions (phase swinging or pole
technique detailed in the following section is utilised. slipping) is given by:

17.16.1.2 Large Generators (>5MVA) ( X G + X T + Z S )n(n − cos θ − j sin θ )

ZR = − XG
For generators above about 5MVA rating, protection against (n − cos θ )2 + sin 2 θ
loss of excitation and pole slipping conditions is normally Equation 17.2
applied.
where:
Consider a generator connected to network, as shown in
Figure 17.20. On loss of excitation, the terminal voltage will EG generated
n= = voltage
begin to decrease and the stator current will increase, resulting ES system
in a decrease of impedance viewed from the generator
Ө = angle by which EG leads ES
terminals and also a change in power factor.
If the generator and system voltages are equal the above
expression becomes:

 j cot θ 
(X G + X T + Z S )1 − 
 2 −X
ZR = G
2
The general case can be represented by a system of circles
with centres on the line CD; see Figure 17.21. Also shown is a
typical machine terminal impedance locus during loss of
excitation conditions.

17-19
Protection & Automation Application Guide

+jX EG the direct-axis value, the ratio of Xd/XG being known as the
=15
.
ES saliency factor. This factor varies with the slip speed. The
1.8
effect of this factor during asynchronous operation is to cause
2.0 XG to vary at slip speed. In consequence, the loss of excitation
Load point impedance locus does not settle at a single point, but it
2.5
continues to describe a small orbit about a mean point.
Loss of field
5.0
locus A protection scheme for loss of excitation must operate
D
decisively for this condition, but its characteristic must not
inhibit stable operation of the generator. One limit of operation
corresponds to the maximum practicable rotor angle, taken to
-R +R
be 120°. The locus of operation can be represented as a circle
EG
=1 on the impedance plane, as shown in Figure 17.22, stable
ES
C operation conditions lying outside the circle.
Locus of
+jX constant MVA

0.5
ZS
0.6
0.7 -R XT +R
0.75X'd
-jX
2X'd
Figure 17.21: Swing curves and loss of synchronism locus

The special cases of EG=ES and EG=0 result in a straight-line

locus that is the right-angled bisector of CD and in a circular
locus that is shrunk to point C, respectively. Xd

Limiting
When excitation is removed from a generator operating generation
synchronously the flux dies away slowly, during which period point

the ratio of EG/ES is decreasing, and the rotor angle of the

machine is increasing. The operating condition plotted on an
impedance diagram therefore travels along a locus that crosses Relay
characteristic
the power swing circles. At the same time, it progresses in the Locus of constant Diameter =Xd /2
load angle
direction of increasing rotor angle. After passing the anti- Trip within circle

phase position, the locus bends round as the internal e.m.f. -jX

collapses, condensing on an impedance value equal to the Figure 17.22: Locus of limiting operating conditions of synchronous
machine reactance. The locus is illustrated in Figure 17.21. machine

The relay location is displaced from point C by the generator On the same diagram the full load impedance locus for one per
reactance XG. One problem in determining the position of these unit power can be drawn. Part of this circle represents a
loci relative to the relay location is that the value of machine condition that is not feasible, but the point of intersection with
the maximum rotor angle curve can be taken as a limiting
impedance varies with the rate of slip. At zero slip XG is equal
operating condition for setting impedance-based loss of
to Xd, the synchronous reactance, and at 100% slip XG is equal
excitation protection.
to X’’d, the sub-transient reactance. The impedance in a
typical case has been shown to be equal to X’d, the transient
17.16.2 Impedance-Based Protection Characteristics
reactance, at 50% slip, and to 2X’d with a slip of 0.33%. The
slip likely to be experienced with asynchronous running is low, Figure 17.21 alludes to the possibility that a protection scheme
perhaps 1%, so that for the purpose of assessing the power for loss of excitation could be based on impedance
measurement. The impedance characteristic must be
swing locus it is sufficient to take the value XG=2Xd’.
appropriately set or shaped to ensure decisive operation for
This consideration has assumed a single value for XG. loss of excitation whilst permitting stable generator operation
However, the reactance Xq on the quadrature axis differs from within allowable limits. One or two offset mho under

17-20
 Chapter 17 ⋅ Generator and Generator Transformer Protection

impedance elements (see Chapter 11 for the principles of • time delay on drop-off, tdo1 = 0s
operation) are ideally suited for providing loss of excitation
If a fast excitation system is employed, allowing load angles of
protection as long as a generator operating at low power
up to 120o to be used, the impedance diameter must be
output (20-30%Pn) does not settle down to operate as an
reduced to take account of the reduced generator impedance
induction generator. The characteristics of a typical two-stage
seen under such conditions. The offset also needs revising. In
loss of excitation protection scheme are illustrated in Figure
these circumstances, typical settings would be:
17.23. The first stage, consisting of settings Xa1 and Xb1 can
be applied to provide detection of loss of excitation even where • impedance element diameter Xb1 = 0.5Xd
a generator initially operating at low power output (20-30%Pn) • impedance element offset Xa1 = -0.75X’d
might settle down to operate as an induction generator.
• time delay on pick-up, td1 = 0.5s – 10s
X • time delay on drop-off, tdo1 = 0s
Normal machine
operating impedance The typical impedance settings for the second element, if used,
R are:
-Xa2 -Xa1 Alarm
angle kV 2
X b2 =
MVA
Xb2 Xb1 kV 2
X a 2 = −0.5 X d'
MVA
The time delay settings td2 and tdo2 are set to zero to give
instantaneous operation and reset.

17.17 POLE SLIPPING PROTECTION

A generator may pole-slip, or fall out of synchronism with the
Figure 17.23: Loss of excitation protection characteristics
power system for a number of reasons. The principal causes
Pick-up and drop-off time delays td1 and tdo1 are associated are prolonged clearance of a heavy fault on the power system,
with this impedance element. Timer td1 is used to prevent when the generator is operating at a high load angle close to
operation during stable power swings that may cause the the stability limit, or partial or complete loss of excitation.
impedance locus of the generator to transiently enter the locus Weak transmission links between the generator and the bulk
of operation set by Xb1. However, the value must short of the power system aggravate the situation. It can also occur
enough to prevent damage as a result of loss of excitation with embedded generators running in parallel with a strong
occurring. If pole-slipping protection is not required (see Utility network if the time for a fault clearance on the Utility
Section 17.17.2), timer tdo1 can be set to give instantaneous network is slow, perhaps because only IDMT relays are
reset. The second field failure element, comprising settings provided. Pole slipping is characterised by large and rapid
Xb1, Xb2, and associated timers td2 and tdo2 can be used to give oscillations in active and reactive power. Rapid disconnection
instantaneous tripping following loss of excitation under full of the generator from the network is required to ensure that
load conditions. damage to the generator is avoided and that loads supplied by
the network are not affected for very long.
17.16.3 Protection Settings Protection can be provided using several methods. The choice
The typical setting values for the two elements vary according of method will depend on the probability of pole slipping
to the excitation system and operating regime of the generator occurring and on the consequences should it occur.
concerned, since these affect the generator impedance seen by
the relay under normal and abnormal conditions. For a 17.17.1 Protection Using Reverse Power Element
generator that is never operated at leading power factor, or at
During pole-slipping, there will be periods where the direction
load angles in excess of 90o the typical settings are:
of active power flow will be in the reverse direction, so a
• impedance element diameter Xb1 = Xd reverse power relay element can be used to detect this, if not
• impedance element offset Xa1 = -0.5X’d used for other purposes. However, since the reverse power
conditions are cyclical, the element will reset during the
• time delay on pick-up, td1 = 0.5s – 10s

17-21
Protection & Automation Application Guide

forward power part of the cycle unless either a very short pick- 17.17.3.1 Pole Slipping Protection by Impedance
up time delay and/or a suitable drop-off time delay is used to Measurement
eliminate resetting. Although a mho type element for detecting the change in
The main advantage of this method is that a reverse power impedance during pole-slipping can be used in some
element is often already present, so no additional relay applications, but with performance limits, a straight line ‘ohm’
elements are required. The main disadvantages are the time characteristic is more suitable. The protection principle is that
taken for tripping and the inability to control the system angle of detecting the passage of the generator impedance through a
at which the generator breaker trip command would be issued, zone defined by two such impedance characteristics, as shown
if it is a requirement to limit the breaker current interruption in Figure 17.24. The characteristic is divided into three zones,
duty. There is also the difficulty of determining suitable A, B, and C. Normal operation of the generator lies in zone A.
settings. Determination of settings in the field, from a When a pole-slip occurs, the impedance traverses zones B and
deliberate pole-slipping test is not possible and analytical C, and tripping occurs when the impedance characteristic
studies may not discover all conditions under which pole- enters zone C.
slipping will occur.
+jX

17.17.2 Protection Using an Under Impedance

Element ZS
Relaying point
With reference to Figure 17.21, a loss of excitation under
impedance characteristic may also be capable of detecting loss XT
of synchronism, in applications where the electrical centre of
XG Slip locus
the power system and the generator lies ‘behind’ the relaying EG=ES
point. This would typically be the case for a relatively small
generator that is connected to a power transmission system
(XG >> (XT + XS)). With reference to Figure 17.23; if C B A
pole-slipping protection response is required, the drop-off -R +R
timer tdo1 of the larger diameter impedance measuring element
should be set to prevent its reset of in each slip cycle, until the
td1 trip time delay has expired. As with reverse power
protection, this would be an elementary form of pole-slipping -jX
protection. It may not be suitable for large machines where Ohm relay 2

rapid tripping is required during the first slip cycle and where Ohm relay 1
some control is required for the system angle at which the Figure 17.24: Pole slipping detection by ohm relays
generator circuit breaker trip command is given. Where Tripping only occurs if all zones are traversed sequentially.
protection against pole-slipping must be guaranteed, a more Power system faults should result in the zones not being fully
sophisticated method of protection should be used. A typical traversed so that tripping will not be initiated. The security of
reset timer delay for pole-slipping protection might be 0.6s. this type of protection scheme is normally enhanced by the
For generator transformer units, the additional impedance in addition of a plain under impedance control element (circle
front of the relaying point may take the system impedance about the origin of the impedance diagram) that is set to
outside the under impedance relay characteristic required for prevent tripping for impedance trajectories for remote power
loss of excitation protection. Therefore, the acceptability of this system faults. Setting of the ohm elements is such that they
pole-slipping protection scheme will be dependent on the lie parallel to the total system impedance vector, and enclose
application. it, as shown in Figure 17.24.

17.17.3 Dedicated Pole-Slipping Protection 17.17.3.2 Use of Lenticular Characteristic

Large generator-transformer units directly connected to grid A more sophisticated approach is to measure the impedance of
systems often require a dedicated pole-slipping protection the generator and use a lenticular impedance characteristic to
scheme to ensure rapid tripping and with system angle control. determine if a pole-slipping condition exists. The lenticular
Historically, dedicated protection schemes have usually been characteristic is shown in Figure 17.25. The characteristic is
based on ohm-type impedance measurement characteristic. divided into two halves by a straight line, called the blinder.

17-22
 Chapter 17 ⋅ Generator and Generator Transformer Protection

The inclination, θ, of the lens and blinder is determined by the X

ZA
angle of the total system impedance. The impedance of the Left-lens Right-lens
A
system and generator-transformer determines the forward B

reach of the lens, ZA, and the transient reactance of the P

ZS
Power Swing In System
ZO
generator determines the reverse reach ZB.
P'
R4 R3 S R2 R1
Blinder XT
a
M
Ang
X T2
ZA T1 Pole Slipping

P Characteristic

Stable Power Swing

XS

ZB
P' Blinder

α θ
R
Figure 17.26: Definition of zones for lenticular characteristic

If the impedance locus lies above line PP’, the swing lies far
Lens out in the power system – i.e. one part of the power system,
including the protected generator, is swinging against the rest
ZB of the network. Tripping may still occur, but only if swinging is
prolonged – meaning that the power system is in danger of
complete break-up. Further confidence checks are introduced
Figure 17.25: Pole-slipping protection using lenticular characteristic
by requiring that the impedance locus spends a minimum time
and blinder within each zone for the pole-slipping condition to be valid.
The trip signal may also be delayed for a number of slip cycles
The width of the lens is set by the angle α and the line PP’,
even if a generator pole-slip occurs – this is to both provide
perpendicular to the axis of the lens, is used to determine if the
confirmation of a pole-slipping condition and allow time for
centre of the impedance swing during a transient is located in
other relays to operate if the cause of the pole slip lies
the generator or power system.
somewhere in the power system. Should the impedance locus
Operation in the case of a generator is as follows. The traverse the zones in any other sequence, tripping is blocked.
characteristic is divided into 4 zones and 2 regions, as shown in
Figure 17.26. Normal operation is with the measured 17.18 STATOR OVERHEATING
impedance in zone R1. If a pole slip develops, the impedance Overheating of the stator may result from:
locus will traverse though zones R2, R3, and R4. When entering
zone R4, a trip signal is issued, provided the impedance lies • overload
below reactance line PP’ and hence the locus of swing lies • failure of the cooling system
within or close to the generator – i.e. the generator is pole • overfluxing
slipping with respect to the rest of the system.
• core faults
Accidental overloading might occur through the combination
of full active load current component, governed by the prime
mover output and an abnormally high reactive current
component, governed by the level of rotor excitation and/or
step-up transformer tap. With a modern protection relay, it is
relatively simple to provide a current-operated thermal replica
protection element to estimate the thermal state of the stator
windings and to issue an alarm or trip to prevent damage.
Although current-operated thermal replica protection cannot
take into account the effects of ambient temperature or

17-23
Protection & Automation Application Guide

uneven heat distribution, it is often applied as a back-up to To minimise overspeed on load rejection and hence the
direct stator temperature measuring devices to prevent mechanical stresses on the rotor, the following sequence is
overheating due to high stator current. With some relays, the used whenever electrical tripping is not urgently required:
thermal replica temperature estimate can be made more
• trip prime mover or gradually reduce power input to zero
accurate through the integration of direct measuring resistance
temperature devices. • allow generated power to decay towards zero
• trip generator circuit breaker only when generated
Irrespective of whether current-operated thermal replica
power is close to zero or when the power flow starts to
protection is applied or not, it is a requirement to monitor the
reverse, to drive the idle turbine
stator temperature of a large generator in order to detect
overheating from whatever cause.
17.19.3 Loss of Vacuum
Temperature sensitive elements, usually of the resistance type, A failure of the condenser vacuum in a steam turbine driven
are embedded in the stator winding at hot-spot locations generator results in heating of the tubes. This then produces
envisaged by the manufacturer, the number used being strain in the tubes, and a rise in temperature of the low-
sufficient to cover all variations. The elements are connected to pressure end of the turbine. Vacuum pressure devices initiate
a temperature sensing relay element arranged to provide alarm progressive unloading of the set and, if eventually necessary,
and trip outputs. The settings will depend on the type of stator tripping of the turbine valves followed by the high voltage
winding insulation and on its permitted temperature rise. circuit breaker. The set must not be allowed to motor in the
event of loss of vacuum, as this would cause rapid overheating
17.19 MECHANICAL FAULTS of the low-pressure turbine blades.
Various faults may occur on the mechanical side of a
generating set. The following sections detail the more 17.20 COMPLETE GENERATOR PROTECTION
important ones from an electrical point of view. SCHEMES
From the preceding sections, it is obvious that the protection
17.19.1 Failure of the Prime Mover
scheme for a generator has to take account of many possible
When a generator operating in parallel with others loses its faults and plant design variations. Determination of the types
power input, it remains in synchronism with the system and of protection used for a particular generator will depend on the
continues to run as a synchronous motor, drawing sufficient nature of the plant and upon economic considerations, which
power to drive the prime mover. This condition may not appear in turn is affected by set size. Fortunately, modern, multi-
to be dangerous and in some circumstances will not be so. function, numerical relays are sufficiently versatile to include all
However, there is a danger of further damage being caused. of the commonly required protection functions in a single
Table 17.1 lists some typical problems that may occur. package, thus simplifying the decisions to be made. The
Protection is provided by a low forward power/reverse power following sections provide illustrations of typical protection
relay, as detailed in Section 17.11. schemes for generators connected to a grid network, but not
all possibilities are illustrated, due to the wide variation in
17.19.2 Overspeed generator sizes and types.
The speed of a turbo-generator set rises when the steam input
is in excess of that required to drive the load at nominal
17.20.1 Direct-Connected Generator
frequency. The speed governor can normally control the A typical protection scheme for a direct-connected generator is
speed, and, in any case, a set running in parallel with others in shown in Figure 17.27. It comprises the following protection
an interconnected system cannot accelerate much functions:
independently even if synchronism is lost. However, if load is • stator differential protection
suddenly lost when the HV circuit breaker is tripped, the set
• overcurrent protection – conventional or voltage
will begin to accelerate rapidly. The speed governor is
dependent
designed to prevent a dangerous speed rise even with a 100%
load rejection, but nevertheless an additional centrifugal • stator earth fault protection
overspeed trip device is provided to initiate an emergency • overvoltage protection
mechanical shutdown if the overspeed exceeds 10%. • undervoltage protection

17-24
 Chapter 17 ⋅ Generator and Generator Transformer Protection

• overload/low forward power/reverse power protection differential protection) may utilise a separate relay. Figure
(according to prime mover type) 17.28 shows a typical overall scheme.
• unbalanced loading Electrical trip of governor Governor
trip

• overheating
Emergency push button

• pole slipping
Stator differential (biased/high
• loss of excitation impedance)
Stator E/F (or neutral voltage
• underfrequency displacement)
Back-up overcurrent (or voltage
dependent O/C)
• inadvertent energisation Lubricating oil failure
Mechanical faults (urgent)

• overfluxing Reverse/low forward power

Underfrequency

• mechanical faults Pole slipping

Overfluxing
Inadvertent energisation
Figure 17.27 illustrates which trips require an instantaneous
electrical trip and which can be time delayed until electrical Overall differential (transformer
differential)
power has been reduced to a low value. The faults that require Buchholz
Excitation
circuit

tripping of the prime mover as well as the generator circuit HV overcurrent

HV restricted E/F
breaker
Generator
breaker are also shown. Transformer winding temperature circuit
breaker
Loss of excitation Low power
Stator winding temperature interlock
Electrical trip of governor Governor Unbalanced loading
trip
Under/overvoltage

Emergency push button Mechanical faults (non-urgent)

N.B. Alarms and time delays omitted for simplicity

Stator differential (biased/high
impedance)
Figure 17.28: Typical tripping arrangements for generator-transformer
Stator E/F (or neutral voltage
displacement) unit
Back-up overcurrent
(or voltage dependent O/C)
Lubricating oil failure 17.21 EMBEDDED GENERATION
Mechanical faults (urgent)
Reverse/low forward power In recent years, through de-regulation of the electricity supply
Underfrequency
Pole slipping industry and the ensuing commercial competition, many
Overfluxing
Inadvertent energisation
electricity users connected to MV power distribution systems
have installed generating sets to operate in parallel with the
Loss of excitation public supply. The intention is either to utilise surplus energy
Stator winding temperature
Unbalanced loading
Excitation
circuit
from other sources, or to use waste heat or steam from the
Under/overvoltage Low power
breaker prime mover for other purposes. Parallel connection of
interlock
Mechanical faults (non-urgent)
Generator
circuit generators to distribution systems did occur before de-
breaker
regulation, but only where there was a net power import from
N.B. Alarms and time delays omitted for simplicity the Utility. Power export to Utility distribution systems was a
Figure 17.27: Typical protection arrangement for a direct-connected relatively new aspect. Since generation of this type can now
generator
be located within a Utility distribution system, as opposed to
being centrally dispatched generation connected to a
17.20.2 Generator-Transformer Units
transmission system, the term ‘Embedded Generation’ is often
These units are generally of higher output than direct- applied. Figure 17.2 illustrates such an arrangement.
connected generators, and hence more comprehensive Depending on size, the embedded generator(s) may be
protection is warranted. In addition, the generator transformer synchronous or asynchronous types, and they may be
also requires protection, for which the protection detailed in connected at any voltage appropriate to the size of plant being
Chapter 16 is appropriate. Overall biased generator/generator considered.
transformer differential protection is commonly applied in
addition, or instead of, differential protection for the The impact of connecting generation to a Utility distribution
transformer alone. A single protection relay may incorporate system that was originally engineered only for downward power
all of the required functions, or the protection of the distribution must be considered, particularly in the area of
transformer (including overall generator/generator transformer protection requirements. In this respect, it is not important
whether the embedded generator is normally capable of export to

17-25
Protection & Automation Application Guide

the Utility distribution system or not, since there may exist fault In addition, particular circumstances may require additional
conditions when this occurs irrespective of the design intent. protection functions:
If plant operation when disconnected from the Utility supply is • neutral voltage displacement
required, underfrequency protection (Section 17.14.2) will • reverse power
become an important feature of the in-plant power system.
• directional overcurrent
During isolated operation, it may be relatively easy to overload
the available generation, such that some form of load In practice, it can be difficult to meet the protection settings or
management system may be required. Similarly, when performance demanded by the Utility without a high risk of
running in parallel with the Utility, consideration needs to be nuisance tripping caused by lack of co-ordination with normal
given to the mode of generator operation if reactive power power system faults and disturbances that do not necessitate
import is to be controlled. The impact on the control scheme tripping of the embedded generation. This is especially true
of a sudden break in the Utility connection to the plant main when applying protection specifically to detect loss of the
busbar also requires analysis. Where the in-plant generation is Utility supply (also called ‘loss of mains’) to cater for operating
run using constant power factor or constant reactive power conditions where there would be no immediate excursion in
control, automatic reversion to voltage control when the Utility voltage or frequency to cause operation of conventional
connection is lost is essential to prevent plant loads being protection functions.
subjected to a voltage outside acceptable limits.
17.21.1 Protection Against Loss of Utility Supply
Limits may be placed by the Utility on the amount of
power/reactive power import/export. These may demand the If the normal power infeed to a distribution system, or to the
use of an in-plant Power Management System to control the part of it containing embedded generation is lost, the effects
embedded generation and plant loads accordingly. Some may be as follows:
Utilities may insist on automatic tripping of the interconnecting a. embedded generation may be overloaded, leading to
circuit breakers if there is a significant departure outside generator undervoltage/ underfrequency
permissible levels of frequency and voltage, or for other reasons.
b. embedded generation may be underloaded, leading
From a Utility standpoint, the connection of embedded to overvoltage/overfrequency
generation may cause problems with voltage control and c. little change to the absolute levels of voltage or
increased fault levels. The settings for protection relays in the frequency if there is little resulting change to the
vicinity of the plant may require adjustment with the load flow through the PCC
emergence of embedded generation. It must also be ensured
The first two effects are covered by conventional voltage and
that the safety, security and quality of supply of the Utility
frequency protection. However, if condition (c) occurs,
distribution system is not compromised. The embedded
conventional protection may not detect the loss of Utility
generation must not be permitted to supply any Utility
supply condition or it may be too slow to do so within the
customers in isolation, since the Utility supply is normally the
shortest possible auto-reclose dead-times that may be applied
means of regulating the system voltage and frequency within
in association with Utility overhead line protection. Detection
the permitted limits. It also normally provides the only system
of condition (c) must be achieved if the requirements of the
earth connection(s), to ensure the correct performance of
Utility are to be met. Many possible methods have been
system protection in response to earth faults. If the Utility
suggested, but the one most often used is the Rate of Change
power infeed fails, it is also important to disconnect the
of Frequency (ROCOF) relay. Its application is based on the
embedded generation before there is any risk of the Utility
fact that the rate of change of small changes in absolute
power supply returning on to unsynchronised machines. In
frequency, in response to inevitable small load changes, will be
practice this generally requires the following protection
faster with the generation isolated than when the generation is
functions to be applied at the ‘Point of Common Coupling’
in parallel with the public, interconnected power system.
(PCC) to trip the coupling circuit breaker:
However, problems with nuisance tripping in response to
• overvoltage national power system events, where the system is subject to
• undervoltage significant frequency excursions following the loss of a large
generator or a major power interconnector, have occurred.
• overfrequency
This is particularly true for geographically islanded power
• underfrequency systems, such as those of the British Isles. An alternative to
• loss of Utility supply ROCOF protection is a technique sometimes referred to as

17-26
 Chapter 17 ⋅ Generator and Generator Transformer Protection

‘voltage vector shift’ protection. In this technique the rate of ROCOF relay for a set time, or once the voltage phase angle
phase change between the directly measured generator bus drift exceeds the set angle, tripping occurs to open the
voltage is compared with a memorised a.c. bus voltage connection between the in-plant and Utility networks.
reference.
While it is possible to estimate the rate of change of frequency
Sources of embedded generation are not normally earthed, from knowledge of the generator set inertia and MVA rating,
which presents a potential safety hazard. In the event of an this is not an accurate method for setting a ROCOF relay
Utility system earth fault, the Utility protection should operate because the rotational inertia of the complete network being
to remove the Utility power infeed. In theory, this should also fed by the embedded generation is required. For example,
result in removal of the embedded generation, through the there may be other embedded generators to consider. As a
action of the stipulated voltage/frequency protection and result, it is invariably the case that the relay settings are
dependable ‘loss of mains’ protection. However, in view of determined at site during commissioning. This is to ensure
safety considerations (e.g. fallen overhead line conductors in that the Utility requirements are met while reducing the
public areas), an additional form of earth fault protection may possibility of a spurious trip under the various operating
also be demanded to prevent the backfeed of an earth fault by scenarios envisaged. However, it is very difficult to determine
embedded generation. The only way of detecting an earth whether a given rate of change of frequency will be due to a
fault under these conditions is to use neutral voltage ‘loss of mains’ incident or a load/frequency change on the
displacement protection. The additional requirement is only public power network, and hence spurious trips are impossible
likely to arise for embedded generation rated above 150kVA, to eliminate. Thus the provision of Loss of Utility Supply
since the risk of small embedded generators not being cleared protection to meet power distribution Utility interface
by other means is negligible. protection requirements, may actually conflict with the
interests of the national power system operator. With the
17.21.2 ROCOF Relay Description growing contribution of non-dispatched embedded generation
A ROCOF relay detects the rate of change of frequency in to the aggregate national power demand, the loss of the
excess of a defined setpoint. The signal is obtained from a embedded generation following a transmission system incident
voltage transformer connected close to the Point of Common that may already challenge the security of the system can only
Coupling (PCC). The principal method used is to measure the aggravate the problem. There have been claims that voltage
time period between successive zero-crossings to determine vector shift protection might offer better security, but it will
the average frequency for each half-cycle and hence the rate of have operation times that vary with the rate of change of
change of frequency. The result is usually averaged over a frequency. As a result, depending on the settings used,
number of cycles. operation times might not comply with Utility requirements
under all circumstances. [Reference 17.1] provides further
17.21.3 Voltage Vector Shift Relay Description details of the operation of ROCOF relays and the problems that
may be encountered.
A voltage vector shift relay detects the drift in voltage phase
angle beyond a defined setpoint as long as it takes place within Nevertheless, because such protection is a common
a set period. Again, the voltage signal is obtained from a requirement of some Utilities, the ‘loss of mains’ protection
voltage transformer connected close to the Point of Common may have to be provided and the possibility of spurious trips
Coupling (PCC). The principal method used is to measure the will have to be accepted in those cases. Site measurements
time period between successive zero-crossings to determine over a period of time of the typical rates of frequency change
the duration of each half-cycle, and then to compare the occurring may assist in negotiations of the settings with the
durations with the memorised average duration of earlier half- Utility, and with the fine-tuning of the protection that may
cycles in order to determine the phase angle drift. already be commissioned.

17.21.4 Setting Guidelines 17.22 EXAMPLES OF GENERATOR

Should loss of the Utility supply occur, it is extremely unlikely PROTECTION SETTINGS
that there will be an exact match between the output of the This section gives examples of the calculations required for
embedded generator(s) and the connected load. A small generator protection. The first is for a typical small generator
frequency change or voltage phase angle change will therefore installed on an industrial system that runs in parallel with the
occur, to which can be added any changes due to the small Utility supply. The second is for a larger generator-transformer
natural variations in loading of an isolated generator with time. unit connected to a grid system.
Once the rate of change of frequency exceeds the setting of the

17-27
Protection & Automation Application Guide

17.22.1 Protection Settings of a Small Industrial offset magnetising inrush or motor starting current waveforms
Generator with an r.m.s level close to rated current, and where there is a
high L/R time constant for the offset, the use of a 0% bias slope
Salient details of the generator, network and protection required
may give rise to maloperation. Such waveforms can be
are given in Table 17.2. The example calculations are based on
encountered when plant of similar rating to the generator is
a MiCOM P343 relay in respect of setting ranges, etc.
being energised or started. Differences between CT designs or
differing remanent flux levels can lead to asymmetric
Generator Data
saturation and the production of a differential spill current.
Rated Rated
Rated
Rated
Prime Therefore, it is appropriate to select a non-zero setting for K1,
KVA KW PF Freque Mover and a value of 5% is usual in these circumstances.
Voltage Current Speed
ncy Type

6250 5000 0.8 11000 328 50 1500

Steam 17.22.1.2 Voltage Controlled Overcurrent Protection
Turbine
This protection is applied as remote backup to the downstream
overcurrent protection in the event of protection or breaker
Generator Parameters failure conditions. This ensures that the generator will not
continue to supply the fault under these conditions.
Generator Xd Xd'
CT Ratio VT Ratio At normal voltage, the current setting must be greater than the
Type p.u. p.u.
maximum generator load current of 328A. A margin must be
Salient Pole 2.349 0.297 500/1 11000/110 allowed for resetting of the relay at this current (reset ratio =
95%) and for the measurement tolerances of the relay (5% of Is
under reference conditions), therefore the current setting is
Network Data
calculated as:
Maximum
Maximum Earth Minimum Phase 328
Earthing Downstream
Fault Fault
Phase
I vcset > × 1.05
Resistor
Current Current 0.95
Fault Current
> 362.5 A
31.7W 200A 145A 850A
The nearest settable value is 365A, or 0.73In.
The minimum phase-phase voltage for a close-up single-phase
Existing Protection to earth fault is 57%, so the voltage setting VS must be less
Overcurrent Settings Earth Fault Settings than this. A value of 30% is typically used, giving VS = 33V.
CT Ratio Characte Characte The current setting multiplying factor K must be chosen such
Setting TMS Setting TMS that KIS is less than 50% of the generator steady-state current
ristic ristic
contribution to an uncleared remote fault. This information is
200/1 SI 144A 0.176 SI 48A 0.15
not available (missing data being common in protection
Table 17.2: Data for small generator protection example
studies). However, the maximum sustained close-up phase
17.22.1.1 Differential Protection fault current (neglecting AVR action) is 145A, so that a setting
chosen to be significantly below this value will suffice. A value
Biased differential protection involves the determination of
of 87.5A (60% of the close-up sustained phase fault current) is
values for four setting values: Is1, Is2, K1 and K2 in Figure 17.5.
therefore chosen, and hence K = 0.6. This is considered to be
Is1 can be set at 5% of the generator rating, in accordance with appropriate based on knowledge of the system circuit
the recommendations for the relay, and similarly the values of
impedances. The TMS setting is chosen to co-ordinate with
Is2 (120%) and K2 (150%) of generator rating. It remains for the downstream feeder protection such that:
the value of K1 to be determined. The recommended value is
generally 0%, but this only applies where CTs that conform to • for a close-up feeder three-phase fault, that results in
IEC 60044-1 class PX (or the superseded BS 3938 Class X) are almost total voltage collapse as seen by the relay
used – i.e. CTs specifically designed for use in differential • for a fault at the next downstream relay location, if the
protection schemes. In this application, the CTs are relay voltage is less than the switching voltage
conventional class 5P CTs that meet the relay requirements in
It should also be chosen so that the generator cannot be
respect of knee-point voltage, etc. Where neutral tail and
subjected to fault or overload current in excess of the stator
terminal CTs can saturate at different times due to transiently

17-28
 Chapter 17 ⋅ Generator and Generator Transformer Protection

short-time current limits. A curve should be provided by the generator). It must therefore have a setting that grades with
manufacturer, but IEC 60034-1 demands that an AC the downstream protection. The protection is driven from the
generator should be able to pass 1.5 times rated current for at generator star-connected VT, while the downstream protection
least 30 seconds. The operating time of the downstream is current operated.
protection for a three-phase fault current of 850A is 0.682s, so
It is therefore necessary to translate the current setting of the
the voltage controlled relay element should have a minimum
downstream setting of the current-operated earth-fault
operating time of 1.09s (0.4s grading margin used as the relay
protection into the equivalent voltage for the NVD protection.
technology used for the downstream relay is not stated – see
The equivalent voltage is found from the formula:
Table 9.2). With a current setting of 87.5A, the operating time
of the voltage controlled relay element at a TMS of 1.0 is: (I pe × Z e )× 3
Veff =
VTratio
0.14
= 3.01s
 850 
0.02 48 × 31.7 × 3
  −1 =
 87.5  100
= 45.6V
Therefore a TMS of:
where:
1.09
= 0.362 Veff = effective voltage setting
3.01
is required. Use 0.375, nearest available setting. Ipe = downstream earth fault current setting
Ze = earthing resistance
17.22.1.3 Stator Earth Fault Protection
Hence a setting of 48V is acceptable. Time grading is
The maximum earth fault current, from Table 17.2, is 200A. required, with a minimum operating time of the NVD
Protection for 95% of the winding can be provided if the relay is protection of 1.13s at an earth fault current of 200A. Using
set to detect a primary earth fault current of 16.4A, and this the expression for the operation time of the NVD element:
equates to a CT secondary current of 0.033A. The nearest
relay setting is 0.04A, providing protection for 90% of the K
t= s
winding. (M − 1)
The protection must grade with the downstream earth fault where:
protection, the settings of which are also given in Table 17.2.
At an earth fault current of 200A, the downstream protection  V 
M =  
has an operation time of 0.73s. The generator earth fault
 Vsnvd 
protection must therefore have an operation time of not less
than 1.13s. At a TMS of 1.0, the generator protection relay and
operating time will be: V = voltage seen by relay
  Vsnvd = relay setting voltage
 
 0.14 s the value of K can be calculated as 3.34. The nearest settable
  200  0.02  value is 3.5, giving an operation time of 1.18s.
  − 1
  20   17.22.1.5 Loss of Excitation Protection
= 2.97s, so the required TMS is Loss of excitation is detected by a mho impedance relay
1.13 element, as detailed in Section 17.16.2. The standard settings
= 0.38 for the P340 series relay are:
2.97
Use a setting of 0.4, nearest available setting.

17.22.1.4 Neutral Voltage Displacement Protection

This protection is provided as back-up earth-fault protection
for the generator and downstream system (direct-connected

17-29
Protection & Automation Application Guide

 CTratio  that can be tolerated following load rejection. The second is a

X a = 0.5'd ×   secondary quantities high-set stage used for instantaneous tripping in the event of
 VTratio 
an intolerable overvoltage condition arising.
500 Generators can normally withstand 105% of rated voltage
= −0.5 × 0.297 × 19.36 ×
100 continuously, so the low-set stage should be set higher than
= −14.5Ω this value. A setting of 117.7V in secondary quantities
(corresponding to 107% of rated stator voltage) is typically
 CTratio 
X b = X d' ×   used, with a definite time delay of 10s to allow for transients
 VTratio  due to load switch-off/rejection, overvoltages on recovery from
500 faults or motor starting, etc.
= 2.349 × 19.36 ×
100 The second element provides protection in the event of a large
= 227Ω overvoltage, by tripping excitation and the generator circuit
breaker (if closed). This must be set below the maximum
The nearest settings provided by the relay are stator voltage possible, taking into account saturation. As the
X a = −14.5Ω open circuit characteristic of the generator is not available,
typical values must be used. Saturation will normally limit the
X b = 227Ω maximum overvoltage on this type of generator to 130%, so a
The time delay td1 should be set to avoid relay element setting of 120% (132V secondary) is typically used.
operation on power swings and a typical setting of 3s is used. Instantaneous operation is required. Generator manufacturers
This value may need to be modified in the light of operating are normally able to provide recommendations for the relay
experience. To prevent cyclical pick-up of the relay element settings. For embedded generators, the requirements of the
without tripping, such as might occur during pole-slipping local Utility may also have to be taken into account. For both
conditions, a drop-off time delay tdo1 is provided and set to elements, a variety of voltage measurement modes are
0.5s. available to take account of possible VT connections (single or
three-phase, etc.), and conditions to be protected against. In
17.22.1.6 Negative Phase Sequence Current Protection this example, a three-phase VT connection is used, and
overvoltages on any phase are to be detected, so a selection of
This protection is required to guard against excessive heating
‘Any’ is used for this setting.
from negative phase sequence currents, whatever the cause.
The generator is of salient pole design, so from IEC 60034-1,
17.22.1.8 Underfrequency Protection
the continuous withstand is 8% of rating and the I2t value is
20s. Using Equation 17.1, the required relay settings can This is required to protect the generator from sustained
found as I2>> = 0.05 and K = 8.6s. The nearest available overload conditions during periods of operation isolated from
values are I2>> = 0.05 and K = 8.6s. The relay also has a the Utility supply. The generating set manufacturer will
cooling time constant Kreset that is normally set equal to the normally provide the details of machine short-time capabilities.
value of K. To co-ordinate with clearance of heavy asymmetric The example relay provides four stages of underfrequency
system faults, that might otherwise cause unnecessary protection. In this case, the first stage is used for alarm
operation of this protection, a minimum operation time tmin purposes and a second stage would be applied to trip the set.
should be applied. It is recommended to set this to a value of The alarm stage might typically be set to 49Hz, with a time
1. Similarly, a maximum time can be applied to ensure that delay of 20s, to avoid an alarm being raised under transient
the thermal rating of the generator is not exceeded (as this is conditions, e.g. during plant motor starting. The trip stage
uncertain, data not available) and to take account of the fact might be set to 48Hz, with a time delay of 0.5s, to avoid
that the P343 characteristic is not identical with that specified tripping for transient, but recoverable, dips in frequency below
in IEC 60034. The recommended setting for tmax is 600s. this value.

17.22.1.7 Overvoltage Protection

This is required to guard against various failure modes, e.g.
AVR failure, resulting in excessive stator voltage. A two-stage
protection is available, the first being a low-set time-delayed
stage that should be set to grade with transient overvoltages

17-30
 Chapter 17 ⋅ Generator and Generator Transformer Protection

A summary of the relay settings is given in Table 17.3.

17.22.1.9 Reverse Power Protection
The relay setting is 5% of rated power. Protection Quantity Value
Is1 5%
 0.05 × 5 × 10 6 
=   Differential protection
Is2 120%
 CTratio × VTratio  K1 5%

 0.05 × 5 × 10 6
 K2 150%
=   Ise 0.04
 500 × 100  Stator earth fault
TMS 0.4
= 5W Vsnvd 48V
Neutral Voltage Displacement
This value can be set in the relay. A time delay is required to K 3.5
guard against power swings while generating at low power Xa -14.5Ω
levels, so use a time delay of 5s. No reset time delay is Xb 227Ω
required. Loss of excitation
td1 3s
td01 0.5s
Ivcset 0.73
Vs 33
Voltage controlled overcurrent
K 0.6
TMS 0.375
I2>> 0.05
K 8.6s
Negative phase sequence Kreset 8.6s
tmin 1.5s
tmax 600s
V> meas mode three-phase
V> operate mode any
V>1 setting 107%
V>1 function DT
Overvoltage
V>1 time delay 10s
V>2 setting 120%
V>2 function DT
V>2 time delay 0sec
F<1 setting 49Hz
F<1 time delay 20s
Underfrequency
F<2 setting 48Hz
F<2 time delay 0.5s
P1 function reverse power
P1 setting 5W
Reverse Power
P1 time delay 5s
P1 DO time 0s
Table 17.3: Small generator protection example – relay settings

17.22.2 Large Generator Transformer Unit Protection

The data for this unit are given in Table 17.4. It is fitted with
two main protection systems to ensure security of tripping in
the event of a fault. To economise on space, the setting
calculations for only one system, that using an MiCOM P343
relay are given. Settings are given in primary quantities

17-31
Protection & Automation Application Guide

throughout. VN
I flt =
(nR ) + (X )
2
Parameter Value Unit 2
f d + X t + nX f
Generator MVA rating 187.65 MVA
Generator MW rating 160 MW Where
Generator voltage 18 kV Iflt = minimum generator primary current for a multi-phase
Synchronous reactance 1.93 pu feeder-end fault
Direct-axis transient reactance 0.189 pu
VN = no load phase-neutral generator voltage
Minimum operating voltage 0.8 pu
Generator negative sequence capability 0.08 pu Xd = generator d-axis synchronous reactance
Generator negative sequence factor, Kg 10 Xt = generator transformer reactance
Generator third harmonic voltage under load 0.02 pu
Rf = feeder resistance
Generator motoring power 0.02 pu
alarm 1.1 Xf = feeder reactance
Generator overvoltage time delay 5 n = number of parallel generators
trip 1.3
Hence,
Generator undervoltage not required
Max pole slipping frequency 10 Hz I flt = 2893 A = 0.361I n
Generator transformer rating 360 MVA
and
Generator transformer leakage reactance 0.244 pu
Generator transformer overflux alarm 1.1 pu
V flt =
(
VN 3 (nR f ) + X t + nX f
2
( )) 2

(nR ) + (X + nX )
Generator transformer overflux alarm 1.2 pu
2 2
Network resistance (referred to 18kV) 0.56 mΩ f d + Xt f

Network reactance (referred to 18kV) 0.0199 Ω

= 1304V
System impedance angle (estimated) 80 deg
= 0.07VN
Minimum load resistance 0.8 Ω
Generator CT ratio 8000/1 A suitable value of K is therefore
Generator VT ratio 18000/120
0.361
Number of generators in parallel 2 = 0.3
Table 17.4: System data for large generator protection example
1.2 .
A suitable value of V2set is 120% of Vflt, giving a value of
17.22.2.1 Biased Differential Protection
1565V. The nearest settable value is 3000V, minimum
The settings follow the guidelines previously stated. As 100% allowable relay setting. The value of V1set is required to be
stator winding earth-fault protection is provided, high above the minimum voltage seen by the generator for a close-
sensitivity is not required and hence Is1 can be set to 10% of up phase-earth fault. A value of 80% of rated voltage is used
generator rated current. This equates to 602A, and the for V1set, 14400V.
nearest settable value on the relay is 640A (= 0.08 of rated CT
current). The settings for K1, Is2 and K2 follow the guidelines 17.2.2.3 Inadvertent Energisation Protection
in the relay manual. This protection is a combination of overcurrent with
undervoltage, the voltage signal being obtained from a VT on
17.22.2.2 Voltage Restrained Overcurrent Protection the generator side of the system. The current setting used is
The setting current Iset has to be greater than the full-load that of rated generator current of 6019A, in accordance with
current of the generator (6019A). A suitable margin must be IEEE C37.102 as the generator is for installation in the USA.
allowed for operation at reduced voltage, so use a multiplying Use 6000A nearest settable value. The voltage setting cannot
factor of 1.2. The nearest settable value is 7200A. The factor be more than 85% of the generator rated voltage to ensure
K is calculated so that the operating current is less than the operation does not occur under normal operation. For this
current for a remote end three phase fault. The steady-state application, a value of 50% of rated voltage is chosen.
current and voltage at the generator for a remote-end three-
phase fault are given by the expressions:

17-32
 Chapter 17 ⋅ Generator and Generator Transformer Protection

17.22.2.4 Negative Phase Sequence Protection The third harmonic undervoltage protection setting must be
The generator has a maximum steady-state capability of 8% of below this value, a factor of 80% is acceptable. Use a value of
rating, and a value of Kg of 10. Settings of I2cmr = 0.06 166.3V and a time delay of 0.5s. Inhibition of the element at
(=480A) and Kg = 10 are therefore used. Minimum and low generator output must be determined during
maximum time delays of 1s and 1300s are used to co-ordinate commissioning.
with external protection and ensure tripping at low levels of
negative sequence current are used.
17.22.2.7 Loss of Excitation Protection
The client requires a two-stage loss of excitation protection
17.22.2.5 Overfluxing Protection function. The first is alarm only, while the second provides
The generator-transformer manufacturer supplied the tripping under high load conditions. To achieve this, the first
following characteristics: impedance element of the P343 loss of excitation protection can
be set in accordance with the guidelines of Section 17.16.3 for a
V generator operating at rotor angles up to 120o, as follows:
Alarm > 1.1
f
X b1 = 0.5 X d = 1.666Ω
V
Trip > 1.2 inverse time characteristic X a1 = −0.75 X d' = −0.245Ω
f
Use nearest settable values of 1.669Ω and 0.253Ω. A time
Hence the alarm setting is
delay of 5s is used to prevent alarms under transient
1.05 conditions. For the trip stage, settings for high load are used,
18000 × = 315 V as given in Section 17.16.3:
60 Hz
A time delay of 5s is used to avoid alarms due to transient kV 2 18 2
X b2 = = = 1.727Ω
conditions. The trip setting is: MVA 187.65
1.2 X a2 = −0.75 X d' = −0.1406Ω
18000 × = 360 V
60 Hz The nearest settable value for Xb2 is 1.725Ω. A time delay of
A TMS value of 10 is selected, to match the withstand curve 0.5s is used.
supplied by the manufacturer.
17.22.2.8 Reverse Power Protection
17.22.2.6 100% Stator Earth Fault Protection The manufacturer-supplied value for motoring power is 2% of
This is provided by a combination of neutral voltage rated power. The recommended setting is therefore 1.6MW.
displacement and third harmonic undervoltage protection. For An instrumentation class CT is used in conjunction with the
the neutral voltage displacement protection to cover 90% of the relay for this protection, to ensure accuracy of measurement.
stator winding, the minimum voltage allowing for generator A time delay of 0.5s is used. The settings should be checked at
operation at a minimum of 92% of rated voltage is: the commissioning stage.

0.92 × 18kV × 0.1 17.22.9 Over/Under-Frequency Protection

= 956.1V
3 For under-frequency protection, the client has specified the
following characteristics:
Use a value of 935.3V, nearest settable value that ensures 90%
of the winding is covered. A 0.5s definite time delay is used to • Alarm: 59.3Hz, 0.5s time delay
prevent spurious trips. The third harmonic voltage under • 1st stage trip: 58.7Hz, 100s time delay
normal conditions is 2% of rated voltage, giving a value of:
• 2nd stage trip: 58.2Hz, 1s time delay
18kV × 0.01
= 207.8V Similarly, the overfrequency is required to be set as follows:
3
• Alarm: 62Hz, 30s time delay
• Trip: 63.5Hz, 10s time delay
These characteristics can be set in the relay directly.

17-33
Protection & Automation Application Guide

to the principles described in Chapter 16.

17.22.2.10 Overvoltage Protection
The generator manufacturers’ recommendation is: Protection Quantity Value Protection Quantity Value
reverse
• Alarm: 110% voltage for 5s Is1 8% P1 function
power
• Trip: 130% voltage, instantaneous Differential Is2 100% Reverse Power P1 setting 1.6MW
protection
K1 0% P1 time delay 0.5s
This translates into the following relay settings:
K2 150% P1 DO time 0s
• Alarm: 19800V, 5s time delay
100% Stator earth Vn3H< 166.3V Inadvertent Dead Mach I> 6000A
• Trip: 23400V, 0.1s time delay fault Vn3H delay 0.5s energisation Dead Mach V< 9000V

Neutral Voltage Vsnvd 935.3V Za 0.243W

17.22.2.11 Pole Slipping Protection Displacement Time Delay 0.5s Zb 0.656W
This is provided by the method described in Section 17.7.2.2. Xa1 -0.245W Zc 0.206W
Detection at a maximum slip frequency of 10Hz is required. Pole Slipping
Xb1 1.666W a 90°
The setting data, according to the relay manual, is as follows: Protection
td1 5s q 80°
Z A = Z n + ZT Loss of Excitation Xa2 -0.1406W T1 15ms

• Forward Reach, = 0.02 + 0.22 Xb2 1.725W T2 15ms

td2 0.5s F>1 setting 62Hz
= 0.24Ω
tDO1 0s F>1 time delay 30s
Z B = Z GEN Overfrequency
Iset 7200A F>2 setting 63.5Hz
• Reverse Reach, = 2 × X d' K 3 F>2 time delay 10s
Voltage restrained
= 0.652Ω overcurrent V1set 14400V P1 function
reverse
power
Z C = 0.9 × Z t V2set 3000V P1 setting 1.6MW
Reverse Power
• Reactance Line,, = 0.9 × 0.22 I2>> 0.06 P1 time delay 0.5s
= 0.198Ω Kg 10 P1 DO time 0s
Negative phase
where: sequence
Kreset 10 F<1 setting 59.3Hz
tmin 1s F<1 time delay 0.5s
Zt = generator transformer leakage current
tmax 1300s F<2 setting 58.7Hz
Zn = network impedance V> meas three- Underfrequency F<2 time delay 100s
mode phase
The nearest settable values are 0.243Ω, 0.656Ω, and 0.206Ω
V> operate
respectively. mode
any F<3 setting 58.2Hz

The lens angle setting, α, is found from the equation: V>1 setting 19800V F<3 time delay 1s
V>1 function DT
 1.54 − Rt min  Overvoltage
a min = 180 − 2 tan 
A −1
 V>1 time
5s
 Z A + ZB  delay
V>2 setting 23400V
and, substituting values,
V>2 function DT
α min = 62.5 
V>2 time
0.1s
delay
Use the minimum settable value of 90o. The blinder angle, θ,
Table 17.5: Relay settings for large generator protection example
is estimated to be 80o, and requires checking during
commissioning. Timers T1 and T2 are set to 15ms as 17.23 REFERENCE
experience has shown that these settings are satisfactory to
[17.1] Survey of Rate of Change of Frequency Relays and
detect pole slipping frequencies up to 10Hz.
Voltage Phase Shift Relays for Loss of Mains
This completes the settings required for the generator, and the Protection. ERA Report 95-0712R, 1995. ERA
relay settings are given in Table 17.5. Of course, additional Technology Ltd.
protection is required for the generator transformer, according

17-34
Chapter 17 ⋅ Generator and Generator Transformer Protection

17-35
INDUSTRIAL AND
COMMERCIAL POWER
SYSTEM PROTECTION
Chapter 18
INDUSTRIAL AND COMMERCIAL
POWER SYSTEM PROTECTION

18.1 Introduction 18.1 INTRODUCTION

18.2 Busbar Arrangement As industrial and commercial operations processes and plants
18.3 Discrimination have become more complex and extensive (Figure 18.1), the
requirement for improved reliability of electrical power supplies
18.4 HRC Fuses has also increased. The potential costs of outage time following
18.5 Industrial Circuit Breaker Industrial a failure of the power supply to a plant have risen dramati-
Circuit Breaker cally as well. The introduction of automation techniques into
industry and commerce has naturally led to a demand for the
18.6 Protection Relays deployment of more power system automation, to improve
18.7 Co-ordination Problems reliability and efficiency.

18.8 Fault Current Contribution from

Induction Motors
18.9 Automatic Changeover System
18.10 Voltage and Phase Reversal Protection
18.11 Power Factor Correction and Protection
of Capacitors
18.12 Examples
18.13 Reference

Figure 18.1: Large modern industrial plant

The protection and control of industrial power supply systems

must be given careful attention. Many of the techniques that
have been evolved for EHV power systems may be applied to
lower voltage systems also, but typically on a reduced scale.
However, industrial systems have many special problems that
have warranted individual attention and the development of
specific solutions.
Many industrial plants have their own generation installed.
Sometimes it is for emergency use only, feeding a limited
number of busbars and with limited capacity. This arrangement
is often adopted to ensure safe shutdown of process plant and
personnel safety. In other plants, the nature of the process
allows production of a substantial quantity of electricity, perhaps
allowing export of any surplus to the public supply system – at
either at sub-transmission or distribution voltage levels. Plants
that run generation in parallel with the public supply distribution
network are often referred to as co-generation or embedded
generation. Special protection arrangements may be demanded
for the point of connection between the private and public utility
plant (see Chapter 17 for further details).

GEGridSolutions.com 18-1
Protection & Automation Application Guide

Industrial systems typically comprise numerous cable feeders 110kV

and transformers. Chapter 16 covers the protection of
transformers and Chapters 9/10 the protection of feeders.
*
18.2 BUSBAR ARRANGEMENT 33kV
NO
The arrangement of the busbar system is obviously very
important, and it can be quite complex for some very large
industrial systems. However, in most systems a single busbar
divided into sections by a bus-section circuit breaker is
common, as shown in Figure 18.2. Main and standby drives *
A B
6kV
for a particular item of process equipment will be fed from NO

different sections of the switchboard, or sometimes from EDG

different switchboards. ∼
NO
HV supply HV supply
1 2
* NC
A B C
0.4kV
NO

NO
A B
6kV

Transformer Transformer *
1 2

2 out of 3 NO
mechanical or electrical
interlock * NC
A B C
0.4kV
NO

NO

* NC
0.4kV
A NO B C
Figure 18.2: Typical switchboard configuration for an industrial plant
Bus section C - Essential supplies
The main power system design criterion is that single outages EDG - Emergency generator
on the electrical network within the plant will not cause loss of * - Two out of three interlock

both the main and standby drives simultaneously. Considering Figure 18.3: Typical industrial power system
a medium sized industrial supply system, shown in Figure The Essential Services Boards are used to feed equipment that
18.3, in more detail, it will be seen that not only are duplicate is essential for the safe shut down, limited operation or
supplies and transformers used, but also certain important preservation of the plant and for the safety of personnel. This
loads are segregated and fed from ‘Essential Services Board(s)’ will cover process drives essential for safe shutdown, venting
(also known as ‘Emergency’ boards), distributed throughout systems, UPS loads feeding emergency lighting, process
the plant. This enables maximum utilisation of the standby control computers, etc. The emergency generator may range
generator facility. A standby generator is usually of the turbo- in size from a single unit rated 20-30kW in a small plant up to
charged diesel-driven type. On detection of loss of incoming several units of 2-10MW rating in a large oil refinery or similar
supply at any switchboard with an emergency section, the plant. Large financial trading institutions may also have
generator is automatically started. The appropriate circuit standby power requirements of several MW to maintain
breakers will close once the generating set is up to speed and computer services.
rated voltage to restore supply to the Essential Services
sections of the switchboards affected, provided that the normal 18.3 DISCRIMINATION
incoming supply is absent - for a typical diesel generator set,
Protection equipment works in conjunction with switchgear.
the emergency supply would be available within 10-20
For a typical industrial system, feeders and plant will be
seconds from the start sequence command being issued.
protected mainly by circuit breakers of various types and by
fused contactors. Circuit breakers will have their associated
overcurrent and earth fault relays. A contactor may also be

18-2
 Chapter 18 ⋅ Industrial and Commercial Power System Protection

equipped with a protection device (e.g. motor protection), but Curve of asymmetrical
associated fuses are provided to break fault currents in excess prospective short-circuit
of the contactor interrupting capability. The rating of fuses and current
selection of relay settings is carried out to ensure that Current trace
discrimination is achieved – i.e. the ability to select and isolate Ip
only the faulty part of the system.
Time
18.4 HRC FUSES Start of
The protection device nearest to the actual point of power Short-circuit
utilisation is most likely to be a fuse or a system of fuses and it Arcing
Pre-arcing
is important that consideration is given to the correct time
time
application of this important device.
The HRC fuse is a key fault clearance device for protection in Total clearance 1 Cycle
industrial and commercial installations, whether mounted in a time
distribution fuseboard or as part of a contactor or fuse-switch. Figure 18.4: HRC fuse cut-off feature
The latter is regarded as a vital part of LV circuit protection, Since the electromagnetic forces on busbars and connections
combining safe circuit making and breaking with an isolating carrying short circuit current are related to the square of the
capability achieved in conjunction with the reliable short- current, it will be appreciated that ‘cut-off’ significantly
circuit protection of the HRC fuse. Fuses combine the reduces the mechanical forces produced by the fault current
characteristics of economy and reliability; factors that are most and which may distort the busbars and connections if not
important in industrial applications. correctly rated. A typical example of ‘cut-off’ current
HRC fuses remain consistent and stable in their breaking characteristics is shown in Figure 18.5. It is possible to use
characteristics in service without calibration and maintenance. this characteristic during the design stage of a project to utilise
This is one of the most significant factors for maintaining fault equipment with a lower fault withstand rating downstream of
clearance discrimination. Lack of discrimination through the fuse, than would be the case if ‘cut-off’ was ignored. This
incorrect fuse grading will result in unnecessary disconnection may save on costs, but appropriate documentation and
of supplies, but if both the major and minor fuses are HRC maintenance controls are required to ensure that only
devices of proper design and manufacture, this need not replacement fuses of very similar characteristics are used
endanger personnel or cables associated with the plant. throughout the lifetime of the plant concerned – otherwise a
safety hazard may arise.
18.4.1 Fuse Characteristics 1000
The time required for melting the fusible element depends on
the magnitude of current. This time is known as the ‘pre-
arcing’ time of the fuse. Vaporisation of the element occurs on
melting and there is fusion between the vapour and the filling 100 1250A
710A 800A
powder leading to rapid arc extinction.
Cut off current (peak kA)

500A 630A
400A
Fuses have a valuable characteristic known as ‘cut-off’, shown 200A 315A
125A
in Figure 18.4. When an unprotected circuit is subjected to a 80A
short circuit fault, the r.m.s. current rises towards a 10
50A
‘prospective’ (or maximum) value. The fuse usually interrupts 35A
25A
the short circuit current before it can reach the prospective 16A
value, in the first quarter to half cycle of the short circuit. The
6A
rising current is interrupted by the melting of the fusible 1.0
2A
element, subsequently dying away dying away to zero during
the arcing period.

0.1
0.1 1.0 10 100 500
Figure 18.5: Typical fuse cut-off current characteristics

18-3
Protection & Automation Application Guide

18.4.2 Discrimination Between Fuses 18.4.5 Protection of Motors

Fuses are often connected in series electrically and it is The manufacturers' literature should also be consulted when
essential that they should be able to discriminate with each fuses are to be applied to motor circuits. In this application,
other at all current levels. Discrimination is obtained when the the fuse provides short circuit protection but must be selected
larger (‘major’) fuse remains unaffected by fault currents that to withstand the starting current (possibly up to 8 times full
are cleared by the smaller (‘minor’) fuse. load current), and also carry the normal full load current
continuously without deterioration. Tables of recommended
The fuse operating time can be considered in two parts:
fuse sizes for both ‘direct on line’ and ‘assisted start’ motor
• the time taken for fault current to melt the element, applications are usually given. Examples of protection using
known as the ‘pre-arcing time’ fuses are given in Section 18.12.1.
• the time taken by the arc produced inside the fuse to
extinguish and isolate the circuit, known as the ‘arcing 18.5 INDUSTRIAL CIRCUIT BREAKERS
time’ Some parts of an industrial power system are most effectively
The total energy dissipated in a fuse during its operation protected by HRC fuses, but the replacement of blown fuse
consists of ‘pre-arcing energy’ and ‘arc energy’. The values are links can be particularly inconvenient in others. In these
usually expressed in terms of I2t, where I is the current passing locations, circuit breakers are used instead, the requirement
through the fuse and t is the time in seconds. Expressing the being for the breaker to interrupt the maximum possible fault
quantities in this manner provides an assessment of the current successfully without damage to itself. In addition to
fault current interruption, the breaker must quickly disperse the
heating effect that the fuse imposes on associated equipment
during its operation under fault conditions. resulting ionised gas away from the breaker contacts, to
prevent re-striking of the arc, and away from other live parts of
To obtain positive discrimination between fuses, the total I2t equipment to prevent breakdown. The breaker, its cable or
value of the minor fuse must not exceed the pre-arcing I2t busbar connections, and the breaker housing, must all be
value of the major fuse. In practice, this means that the major constructed to withstand the mechanical forces resulting from
fuse will have to have a rating significantly higher than that of the magnetic fields and internal arc gas pressure produced by
the minor fuse, and this may give rise to problems of the highest levels of fault current to be encountered.
discrimination. Typically, the major fuse must have a rating of
The types of circuit breaker most frequently encountered in
at least 160% of the minor fuse for discrimination to be
obtained. industrial system are described in the following sections.

18.4.3 Protection of Cables by Fuses 18.5.1 Miniature Circuit Breakers (MCBs)

PVC cable is allowed to be loaded to its full nominal rating only MCBs are small circuit breakers, both in physical size but more
if it has ‘close excess current protection’. This degree of importantly, in ratings. The basic single pole unit is a small,
protection can be given by means of a fuse link having a manually closed, electrically or manually opened switch
‘fusing factor’ not exceeding 1.5, where: housed in a moulded plastic casing. They are suitable for use
on 230V a.c. single-phase/400V a.c. three-phase systems and
• Fusing factor = Minimum Fusing Current/Current for d.c. auxiliary supply systems, with current ratings of up to
Rating 125A. Contained within each unit is a thermal element, in
Cables constructed using other insulating materials (e.g. which a bimetal strip will trip the switch when excessive
paper, XLPE) have no special requirements in this respect. current passes through it. This element operates with a
predetermined inverse-time/current characteristic. Higher
18.4.4 Effect of Ambient Temperature currents, typically those exceeding 3-10 times rated current,
trip the circuit breaker without intentional delay by actuating a
High ambient temperatures can influence the capability of
magnetic trip overcurrent element. The operating time
HRC fuses. Most fuses are suitable for use in ambient
characteristics of MCBs are not adjustable. European
temperatures up to 35°C, but for some fuse ratings, derating
Standard EN 60898-2 defines the instantaneous trip
may be necessary at higher ambient temperatures.
characteristics, while the manufacturer can define the inverse
Manufacturers' published literature should be consulted for the
time thermal trip characteristic. Therefore, a typical tripping
de-rating factor to be applied.
characteristic does not exist. The maximum a.c. breaking
current permitted by the standard is 25kA.

18-4
 Chapter 18 ⋅ Industrial and Commercial Power System Protection

Single-pole units may be coupled mechanically in groups to • because of their higher ratings, MCCBs are usually
form 2, 3 or 4 pole units, when required, by assembly on to a positioned in the power distribution system nearer to
rail in a distribution board. The available ratings make MCBs the power source than the MCBs
suitable for industrial, commercial or domestic applications, for • the appropriate European specification is EN 60947-2
protecting equipment such as cables, lighting and heating
circuits, and also for the control and protection of low power Care must be taken in the short-circuit ratings of MCCBs.
motor circuits. They may be used instead of fuses on MCCBs are given two breaking capacities, the higher of which
individual circuits, and they are usually ‘backed-up’ by a device is its ultimate breaking capacity. The significance of this is that
of higher fault interrupting capacity. after breaking such a current, the MCCB may not be fit for
continued use. The lower, or service, short circuit breaking
Various accessory units, such as isolators, timers, and capacity permits continued use without further detailed
undervoltage or shunt trip release units may be combined with examination of the device. The standard permits a service
an MCB to suit the particular circuit to be controlled and breaking capacity of as little as 25% of the ultimate breaking
protected. When personnel or fire protection is required, a capacity. While there is no objection to use of MCCBs to break
residual current device (RCD) may be combined with the short-circuit currents between the service and ultimate values,
MCB. The RCD contains a miniature core balance current the inspection required after such a trip reduces the usefulness
transformer that embraces all of the phase and neutral of the device in such circumstances. It is also clearly difficult
conductors to provide sensitivity to earth faults within a typical to determine if the magnitude of the fault current was in
range of 0.05% to 1.5% of rated current, dependent on the RCD excess of the service rating.
selected. The core balance CT energises a common magnetic
trip actuator for the MCB assembly. The time-delay characteristics of the magnetic or thermal
timed trip, together with the necessity for, or size of, a back-up
It is also possible to obtain current-limiting MCBs. These types device varies with make and size of breaker. Some MCCBs are
open prior to the prospective fault current being reached, and fitted with microprocessor-controlled programmable trip
therefore have similar properties to HRC fuses. It is claimed characteristics offering a wide range of such characteristics.
that the extra initial cost is outweighed by lifetime savings in Time–delayed overcurrent characteristics may not be the same
replacement costs after a fault has occurred, plus the as the standard characteristics for dependent-time protection
advantage of providing improved protection against electric stated in IEC 60255-3. Hence, discrimination with other
shock if an RCD is used. As a result of the increased safety protection must be considered carefully. There can be
provided by MCBs fitted with an RCD device, they are tending problems where two or more MCBs or MCCBs are electrically
to replace fuses, especially in new installations. in series, as obtaining selectivity between them may be
difficult. There may be a requirement that the major device
18.5.2 Moulded Case Circuit Breakers (MCCBs) should have a rating of k times the minor device to allow
These circuit breakers are broadly similar to MCBs but have discrimination, in a similar manner to fuses – the
the following important differences: manufacturer should be consulted as to value of k. Careful
examination of manufacturers’ literature is always required at
• the maximum ratings are higher, with voltage ratings
the design stage to determine any such limitations that may be
up to 1000V a.c./1200V d.c. Current ratings of 2.5kA
imposed by particular makes and types of MCCBs. An
continuous/180kA r.m.s break are possible, dependent
example of co-ordination between MCCBs, fuses and relays is
upon power factor
given in Section 18.12.2.
• the breakers are larger, commensurate with the level of
ratings. Although available as single, double or triple 18.5.3 Air Circuit Breakers (ACBs)
pole units, the multiple pole units have a common
Air circuit breakers are frequently encountered on industrial
housing for all the poles. Where fitted, the switch for
systems rated at 3.3kV and below. Modern LV ACBs are
the neutral circuit is usually a separate device, coupled
available in current ratings of up to 6.3kA with maximum
to the multi-pole MCCB
breaking capacities in the range of 85kA-120kA r.m.s.,
• the operating levels of the magnetic and thermal depending on system voltage.
protection elements may be adjustable, particularly in
the larger MCCBs This type of breaker operates on the principle that the arc
produced when the main contacts open is controlled by
directing it into an arc chute. Here, the arc resistance is
increased and hence the current reduced to the point where

18-5
Protection & Automation Application Guide

the circuit voltage cannot maintain the arc and the current 18.5.4 Oil Circuit Breakers (OCBs)
reduces to zero. To assist in the quenching of low current arcs, Oil circuit breakers have been very popular for many years for
an air cylinder may be fitted to each pole to direct a blast of air industrial supply systems at voltages of 3.3kV and above. They
across the contact faces as the breaker opens, so reducing are found in both ‘bulk oil’ and ‘minimum oil’ types, the only
contact erosion. significant difference being the volume of oil in the tank.
Air circuit breakers for industrial use are usually withdrawable In this type of breaker, the main contacts are housed in an oil-
and are constructed with a flush front plate making them ideal filled tank, with the oil acting as the both the insulation and
for inclusion together with fuse switches and MCBs/MCCBs in the arc-quenching medium. The arc produced during contact
modular multi-tier distribution switchboards, so maximising separation under fault conditions causes dissociation of the
the number of circuits within a given floor area. hydrocarbon insulating oil into hydrogen and carbon. The
Older types using a manual or dependent manual closing hydrogen extinguishes the arc. The carbon produced mixes
mechanism are regarded as being a safety hazard. This arises with the oil. As the carbon is conductive, the oil must be
under conditions of closing the CB when a fault exists on the changed after a prescribed number of fault clearances, when
circuit being controlled. During the close-trip operation, there the degree of contamination reaches an unacceptable level.
is a danger of egress of the arc from the casing of the CB, with Because of the fire risk involved with oil, precautions such as
a consequent risk of injury to the operator. Such types may be the construction of fire/blast walls may have to be taken when
required to be replaced with modern equivalents. OCBs are installed.
ACBs are normally fitted with integral overcurrent protection,
thus avoiding the need for separate protection devices. 18.5.5 Vacuum Circuit Breakers (VCBs)
However, the operating time characteristics of the integral In recent years, this type of circuit breaker, along with CBs
protection are often designed to make discrimination with using SF6, has replaced OCBs for new installations in
MCBs/MCCBs/fuses easier and so they may not be in industrial/commercial systems at voltages of 3.3kV and above.
accordance with the standard dependent time characteristics
Compared with oil circuit breakers, vacuum breakers have no
given in IEC 60255-3. Therefore, problems in co-ordination
fire risk and they have high reliability with long maintenance
with discrete protection relays may still arise, but modern
free periods. A variation is the vacuum contactor with HRC
numerical relays have more flexible characteristics to alleviate
fuses, used in HV motor starter applications.
such difficulties. ACBs will also have facilities for accepting an
external trip signal, and this can be used in conjunction with
an external relay if desired. Figure 18.6 shows the typical
18.5.6 SF6 Circuit Breakers
tripping characteristics available. In some countries, circuit breakers using SF6 gas as the arc-
quenching medium are preferred to VCBs as the replacement
1000 for air- and oil-insulated CBs. Some modern types of
Inverse
switchgear cubicles enable the use of either VCBs or SF6-
Very Inverse insulated CBs according to customer requirements. Ratings of
100
Ultra Inverse up to 31.5kA r.m.s. fault break at 36kV and 40kA at 24kV are
Short Circuit typical. SF6-insulated CBs also have advantages of reliability
10 and maintenance intervals compared to air- or oil-insulated
Time (s)

CBs and are of similar size to VCBs for the same rating.
1
18.6 PROTECTION RELAYS
When the circuit breaker itself does not have integral
0.1
protection, then a suitable external relay will have to be
provided. For an industrial system, the most common
0.01 protection relays are time-delayed overcurrent and earth fault
1 10 20 relays. Chapter 9 provides details of the application of
Current (multiple of setting) overcurrent relays.
Figure 18.6: Typical tripping characteristics of an ACB

18-6
 Chapter 18 ⋅ Industrial and Commercial Power System Protection

Residual
CT Phase current System Type of
connections elements Notes
elements fault

A B C

Petersen coil and unearthed

(a) 3Ph. 3w Ph. - Ph. systems

A B C
(i) Ph.-Ph.
(b) 3Ph. 3w
(ii) Ph.–E* * Earth fault protection only if
earth fault current is not less
(i) Ph. - Ph. than twice primary operating
(ii) Ph. – E current
(c) 3Ph. 4w
(iii) Ph. - N

A B C

Phase elements must be in

(i) Ph. - Ph. same phases at all stations.
(d) 3Ph. 3w
(ii) Ph. – E Earth fault settings may be less
than full load

A B C
(i) Ph. - Ph. Earth fault settings may
(e) 3Ph. 3w
(ii) Ph. – E be less than full load

(i) Ph. - Ph. Earth fault settings may be less

(f) 3Ph. 4w (ii) Ph. – E than full load but must be greater
(iii) Ph. - N than largest Ph. – N load

A B C N

(i) Ph. - Ph.

Earth-fault settings may
(g) 3Ph. 4w (ii) Ph. – E be less than full load
(iii) Ph. - N

A B C N

3Ph. 3w Earth-fault settings may

(h) or Ph. - E
3Ph. 4w be less than full load

Ph. = phase ; w = wire ; E = earth ; N = neutral

Figure 18.7: Overcurrent and earth fault relay connections

18-7
Protection & Automation Application Guide

Traditionally, for three wire systems, overcurrent relays have core balance type CT generally enables more sensitive settings
often been applied to two phases only for relay element to be obtained without the risk of nuisance tripping with
economy. Even up until the last generation of static relays, residually connected phase CTs. When this method is applied
economy was still a consideration in terms of the number of to a four-wire system, it is essential that both the phase and
analogue current inputs that were provided. Two overcurrent neutral conductors are passed through the core balance CT
elements can be used to detect any interphase fault, so it was aperture. For a 3-wire system, care must be taken with the
conventional to apply two elements on the same phases at all arrangement of the cable sheath, otherwise cable faults
relay locations. The phase CT residual current connections for involving the sheath may not result in relay operation (Figure
an earth fault relay element are unaffected by such a 18.8).
convention. Figure 18.7 shows the possible relay connections
Cable gland
and limitations on settings.
Cable box
18.7 CO-ORDINATION PROBLEMS
There are a number of problems that commonly occur in
industrial and commercial networks that are covered in the Cable gland /sheath I >
ground connection
following sections.

18.7.1.1 Earth fault protection with residually-connected

CTs
For four-wire systems, the residual connection of three phase
CTs to an earth fault relay element will offer earth fault
protection, but the earth fault relay element must be set above
the highest single-phase load current to avoid nuisance
tripping. Harmonic currents (which may sum in the neutral
conductor) may also result in spurious tripping. The earth fault No operation
relay element will also respond to a phase-neutral fault for the I >

phase that is not covered by an overcurrent element where

(a) Incorrect positioning
only two overcurrent elements are applied. Where it is
required that the earth fault protection should respond only to
earth fault current, the protection element must be residually
connected to three phase CTs and to a neutral CT or to a core-
balance CT. In this case, overcurrent protection must be
applied to all three phases to ensure that all phase-neutral
faults will be detected by overcurrent protection. Placing a CT
in the neutral earthing connection to drive an earth fault relay
provides earth fault protection at the source of supply for a 4- Operation
wire system. If the neutral CT is omitted, neutral current is I >
seen by the relay as earth fault current and the relay setting
would have to be increased to prevent tripping under normal (b) Correct positioning

load conditions. Figure 18.8: CBCT connection for four-wire system

When an earth fault relay is driven from residually connected 18.7.2 Four-Wire Dual-Fed Substations
CTs, the relay current and time settings must be such that that
The co-ordination of earth fault relays protecting four-wire
the protection will be stable during the passage of transient CT
systems requires special consideration in the case of low
spill current through the relay. Such spill current can flow in
voltage, dual-fed installations. Horcher [18.1] has suggested
the event of transient, asymmetric CT saturation during the
various methods of achieving optimum co-ordination.
passage of offset fault current, inrush current or motor starting
Problems in achieving optimum protection for common
current. The risk of such nuisance tripping is greater with the
configurations are described below.
deployment of low impedance electronic relays rather than
electromechanical earth fault relays which presented
significant relay circuit impedance. Energising a relay from a

18-8
 Chapter 18 ⋅ Industrial and Commercial Power System Protection

18.7.2.1 Use of 3-pole CBs substation with both incoming supply breakers and bus section
When both neutrals are earthed at the transformers and all breaker closed.
circuit breakers are of the 3-pole type, the neutral busbar in
the switchgear creates a double neutral to earth connection, as
shown in Figure 18.9. In the event of an uncleared feeder
earth fault or busbar earth fault, with both the incoming supply
breakers closed and the bus section breaker open, the earth
fault current will divide between the two earth connections.
I > I >
Earth fault relay RE2 may operate, tripping the supply to the
RS1 RS2
healthy section of the switchboard as well as relay RE1 tripping
the supply to the faulted section. I >
Supply 1 Supply 2
RE
F1

IF/2 IF/2
RE1 RE2

IF I >
IF/2 F2
RF

Supply 1 Supply 2
IF/2 Neutral busbar Figure 18.10: Dual fed four-wire systems: use of single point neutral
earthing

Bus section CB When the substation is operated with the bus section switch
closed and either one or both of the incoming supply breakers
closed, it is possible for unbalanced neutral busbar load current
IF caused by single phase loading to operate relay RS1 and/or RS2
and inadvertently trip the incoming breaker. Interlocking the
Figure 18.9: Dual fed four-wire systems: use of 3-pole CBs trip circuit of each RS relay with normally closed auxiliary
contacts on the bus section breaker can prevent this.
If only one incoming supply breaker is closed, the earth fault
relay on the energised side will see only a proportion of the However, should an earth fault occur on one side of the busbar
fault current flowing in the neutral busbar. This not only when relays RS are already operated, it is possible for a contact
significantly increases the relay operating time but also reduces race to occur. When the bus section breaker opens, its break
its sensitivity to low-level earth faults. contact may close before the RS relay trip contact on the
healthy side can open (reset). Raising the pick-up level of
The solution to this problem is to utilise 4-pole CBs that switch
relays RS1 and RS2 above the maximum unbalanced neutral
the neutral as well as the three phases. Then there is only a
current may prevent the tripping of both supply breakers in this
single earth fault path and relay operation is not compromised.
case. However, the best solution is to use 4-pole circuit
18.7.2.2 Use of single earth electrode breakers, and independently earth both sides of the busbar.

A configuration sometimes adopted with four-wire dual-fed If, during a busbar earth fault or uncleared feeder earth fault,
substations where only a 3-pole bus section CB is used is to the bus section breaker fails to open when required, the
use a single earth electrode connected to the mid-point of the interlocking break auxiliary contact will also be inoperative.
neutral busbar in the switchgear, as shown in Figure 18.10. This will prevent relays RS1 and RS2 from operating and
When operating with both incoming main circuit breakers and providing back-up protection, with the result that the fault
the bus section breaker closed, the bus section breaker must must be cleared eventually by slower phase overcurrent relays.
be opened first should an earth fault occur, in order to achieve An alternative method of obtaining back-up protection could
discrimination. The co-ordination time between the earth fault be to connect a second relay RE, in series with relay RE, having
relays RF and RE should be established at fault level F2 for a an operation time set longer than that of relays RS1 and RS2.
But since the additional relay must be arranged to trip both of

18-9
Protection & Automation Application Guide

the incoming supply breakers, back-up protection would be rare cases, it may have to be taken into consideration for
obtained but busbar selectivity would be lost. correct time grading for through-fault protection
considerations, and in the calculation of peak voltage for high-
An example of protection of a typical dual-fed switchboard is
impedance differential protection schemes.
given in Section 18.12.3.
It is more important to take motor contribution into account
18.8 FAULT CURRENT CONTRIBUTION FROM when considering the fault rating of equipment (busbars,
INDUCTION MOTORS cables, switchgear, etc.). In general, the initial a.c. component
of current from a motor at the instant of fault is of similar
When an industrial system contains motor loads, the motors
magnitude to the direct-on-line starting current of the motor.
will contribute fault current for a short time. They contribute
For LV motors, 5xFLC is often assumed as the typical fault
to the total fault current via the following mechanism.
current contribution (after taking into account the effect of
When an induction motor is running, a flux, generated by the motor cable impedance), with 5.5xFLC for HV motors, unless
stator winding, rotates at synchronous speed and interacts it is known that low starting current HV motors are used. It is
with the rotor. If a large reduction in the stator voltage occurs also accepted that similar motors connected to a busbar can
for any reason, the flux in the motor cannot change be lumped together as one equivalent motor. In doing so,
instantaneously and the mechanical inertia of the machine will motor rated speed may need to be taken into consideration, as
tend to inhibit speed reduction over the first few cycles of fault 2 or 4 pole motors have a longer fault current decay than
duration. The trapped flux in the rotor generates a stator motors with a greater number of poles. The kVA rating of the
voltage equal initially to the back e.m.f. induced in the stator single equivalent motor is taken as the sum of the kVA ratings
before the fault and decaying according to the X/R ratio of the of the individual motors considered. It is still possible for
associated flux and current paths. The induction motor motor contribution to be neglected in cases where the motor
therefore acts as a generator resulting in a contribution of load on a busbar is small in comparison to the total load
current having both a.c. and d.c. components decaying (again IEC 60909 provides guidance in this respect).
exponentially. Typical 50Hz motor a.c. time constants lie in However, large LV motor loads and all HV motors should be
the range 10ms-60ms for LV motors and 60-200ms for HV considered when calculating fault levels.
motors. This motor contribution has often been neglected in
the calculation of fault levels. 18.9 AUTOMATIC CHANGEOVER SYSTEMS
Industrial systems usually contain a large component of motor Induction motors are often used to drive critical loads. In
load, so this approach is incorrect. The contribution from some industrial applications, such as those involving the
motors to the total fault current may well be a significant pumping of fluids and gases, this has led to the need for a
fraction of the total in systems having a large component of power supply control scheme in which motor and other loads
motor load. Standards relating to fault level calculations, such are transferred automatically on loss of the normal supply to
as IEC 60909, require the effect of motor contribution to be an alternative supply. A quick changeover, enabling the motor
included where appropriate. They detail the conditions under load to be re-accelerated, reduces the possibility of a process
which this should be done, and the calculation method to be trip occurring. Such schemes are commonly applied for large
used. Guidance is provided on typical motor fault current generating units to transfer unit loads from the unit
contribution for both HV and LV motors if the required data is transformer to the station supply/start-up transformer.
not known. Therefore, it is now relatively easy, using
When the normal supply fails, induction motors that remain
appropriate calculation software, to determine the magnitude
connected to the busbar slow down and the trapped rotor flux
and duration of the motor contribution, so enabling a more
generates a residual voltage that decays exponentially. All
accurate assessment of the fault level for:
motors connected to a busbar will tend to decelerate at the
• discrimination in relay co-ordination same rate when the supply is lost if they remain connected to
• determination of the required switchgear/busbar fault the busbar. This is because the motors will exchange energy
rating between themselves, so that they tend to stay ‘synchronised’
to each other. As a result, the residual voltages of all the
For protection calculations, motor fault level contribution is not motors decay at nearly the same rate. The magnitude of this
an issue that is generally important. In industrial networks, voltage and its phase displacement with respect to the healthy
fault clearance time is often assumed to occur at 5 cycles after alternative supply voltage is a function of time and the speed
fault occurrence, and at this time, the motor fault level of the motors. The angular displacement between the residual
contribution is much less than just after fault occurrence. In motor voltage and the incoming voltage will be 180° at some

18-10
 Chapter 18 ⋅ Industrial and Commercial Power System Protection

instant. If the healthy alternative supply is switched on to Figure 18.11(b) shows the residual voltage method, which is
motors which are running down under these conditions, very more common, especially in the petrochemical industry.
high inrush currents may result, producing stresses which
Two feeders are used, supplying two busbar sections
could be of sufficient magnitude to cause mechanical damage,
connected by a normally open bus section breaker. Each
as well as a severe dip in the alternative supply voltage.
feeder is capable of carrying the total busbar load. Each bus
Two methods of automatic transfer are used: section voltage is monitored and loss of supply on either
section causes the relevant incomer CB to open. Provided
• in-phase transfer system
there are no protection operations to indicate the presence of a
• residual voltage system busbar fault, the bus section breaker is closed automatically to
The in-phase transfer method is shown in Figure 18.11(a). restore the supply to the unpowered section of busbar after the
Normal and standby feeders from the same power source are residual voltage generated by the motors running down on
used. that section has fallen to a an acceptable level. This is
between 25% and 40%, of nominal voltage, dependent on the
Phase angle measurement is used to sense the relative phase characteristics of the power system. The choice of residual
angle between the standby feeder voltage and the motor voltage setting will influence the re-acceleration current after
busbar voltage. When the voltages are approximately in the bus section breaker closes. For example, a setting of 25%
phase, or just prior to this condition through prediction, a high- may be expected to result in an inrush current of around 125%
speed circuit breaker is used to complete the transfer. This of the starting current at full voltage. Alternatively, a time
method is restricted to large high inertia drives where the delay could be used as a substitute for residual voltage
gradual run down characteristic upon loss of normal feeder measurement, which would be set with knowledge of the plant
supply can be predicted accurately. to ensure that the residual voltage would have decayed
Preferred Standby sufficiently before transfer is initiated.
feeder feeder
The protection relay settings for the switchboard must take
account of the total load current and the voltage dip during the
re-acceleration period in order to avoid spurious tripping
during this time. This time can be several seconds where large
Phase inertia HV drives are involved.
High speed
angle ϕ<
relay CB
18.10 VOLTAGE AND PHASE REVERSAL
PROTECTION
Voltage relays have been widely used in industrial power
supply systems. The principle purposes are to detect
M
(a) In phase transfer method undervoltage and/or overvoltage conditions at switchboards to
disconnect supplies before damage can be caused from these
conditions or to provide interlocking checks. Prolonged
Feeder Feeder
No.1 No.2 overvoltage may cause damage to voltage-sensitive equipment
(e.g. electronics), while undervoltage may cause excessive
current to be drawn by motor loads. Motors are provided with
thermal overload protection to prevent damage with excessive
current, but undervoltage protection is commonly applied to
disconnect motors after a prolonged voltage dip. With a
Ursd < Ursd <
voltage dip caused by a source system fault, a group of motors
could decelerate to such a degree that their aggregate re-
acceleration currents might keep the recovery voltage
depressed to a level where the machines might stall. Modern
M M M M numerical motor protection relays typically incorporate voltage
protection functions, thus removing the need for discrete
(b) Residual voltage method undervoltage relays for this purpose (see Chapter 19). Older
Figure 18.11: Auto-transfer systems installations may still utilise discrete undervoltage relays, but
the setting criteria remain the same.

18-11
Protection & Automation Application Guide

Reverse phase sequence voltage protection should be applied

where it may be dangerous for a motor to be started with

Capacitor kVAr
rotation in the opposite direction to that intended. Incorrect
rotation due to reverse phase sequence might be set up
following some error after power system maintenance or
repairs, e.g. to a supply cable. Older motor control boards
might have been fitted with discrete relays to detect this
condition. Modern motor protection relays may incorporate kW V
this function. If reverse phase sequence is detected, motor ϕ2
ϕ1
starting can be blocked. If reverse phase sequence voltage kVA2 lo
a
with co d current
protection is not provided, the high-set negative phase kV mpens
A ation

Magnetising kVAr
1 l
sequence current protection in the relay would quickly detect o

Compensating kVAr
co ad
m cu
the condition once the starting device is closed – but initial pe rr
ns en
at t w
reverse rotation of the motor could not be prevented. ion ith
ou
t

18.11 POWER FACTOR CORRECTION AND

PROTECTION OF CAPACITORS
Loads such as induction motors draw significant reactive Figure 18.12: Power factor correction principle
power from the supply system, and a poor overall power factor
The following may be deduced from this vector diagram:
may result. The flow of reactive power increases the voltage-
drops through series reactances such as transformers and kW
reactors, it uses up some of the current carrying capacity of
= cos ∠ϕ1
Uncorrected power factor =
kVA 1
power system plant and it increases the resistive losses in the
power system. kW
= cos ∠ϕ 2
To offset the losses and restrictions in plant capacity they incur
Corrected power factor =
kVA 2
and to assist with voltage regulation, Utilities usually apply
tariff penalties to large industrial or commercial customers for Reduction in kVA = kVA1 - kVA2
running their plant at excessively low power factor. The If the kW load and uncorrected power factors are known, then
customer is thereby induced to improve the power factor of his the capacitor rating in kvar to achieve a given degree of
system and it may be cost-effective to install fixed or variable correction may be calculated from:
power factor correction equipment to raise or regulate the
plant power factor to an acceptable level. Capacitor kVAr = kW × (tan ∠ϕ1 − tan ∠ϕ 2 )

Shunt capacitors are often used to improve power factor. The A spreadsheet can easily be constructed to calculate the
basis for compensation is shown in Figure 18.12, where <φ1 required amount of compensation to achieve a desired power
represents the uncorrected power factor angle and <φ2 the factor.
angle relating to the desired power factor, after correction.
18.11.1 Capacitor Control
Where the plant load or the plant power factor varies
considerably, it is necessary to control the power factor
correction, since over-correction will result in excessive system
voltage and unnecessary losses. In a few industrial systems,
capacitors are switched in manually when required, but
automatic controllers are standard practice. A controller
provides automatic power factor correction, by comparing the
running power factor with the target value. Based on the
available groupings, an appropriate amount of capacitance is
switched in or out to maintain an optimum average power
factor. The controller is fitted with a ‘loss of voltage’ relay
element to ensure that all selected capacitors are disconnected

18-12
 Chapter 18 ⋅ Industrial and Commercial Power System Protection

instantaneously if there is a supply voltage interruption. When Protection equipment is required to prevent rupture of the
the supply voltage is restored, the capacitors are reconnected capacitor due to an internal fault and also to protect the cables
progressively as the plant starts up. To ensure that capacitor and associated equipment from damage in case of a capacitor
groups degrade at roughly the same rate, the controller usually failure. If fuse protection is contemplated for a three-phase
rotates selection or randomly selects groups of the same size in capacitor, HRC fuses should be employed with a current rating
order to even out the connected time. The provision of of not less than 1.5 times the rated capacitor current.
overvoltage protection to trip the capacitor bank is also
Medium voltage capacitor banks can be protected by the
desirable in some applications. This would be to prevent a
scheme shown in Figure 18.14. Since harmonics increase
severe system overvoltage if the power factor correction (PFC)
capacitor current, the relay will respond more correctly if it
controller fails to take fast corrective action.
does not have in-built tuning for harmonic rejection.
The design of PFC installations must recognise that many
Double star capacitor banks are employed at medium voltage.
industrial loads generate harmonic voltages, with the result
As shown in Figure 18.13, a current transformer in the inter
that the PFC capacitors may sink significant harmonic
star-point connection can be used to drive a protection relay to
currents. A harmonic study may be necessary to determine
detect the out-of-balance currents that will flow when
the capacitor thermal ratings or whether series filters are
capacitor elements become short-circuited or open-circuited.
required.
The relay will have adjustable current settings, and it might
contain a bias circuit, fed from an external voltage transformer,
18.11.2 Motor P.F. Correction
that can be adjusted to compensate for steady-state spill
When dealing with power factor correction of motor loads, current in the inter star-point connection.
group correction is not always the most economical method.
A
Some industrial consumers apply capacitors to selected motor
substations rather than applying all of the correction at the B

main incoming substation busbars. Sometimes, power factor C

correction may even be applied to individual motors, resulting
in optimum power factor being obtained under all conditions of
aggregate motor load. In some instances, better motor
starting may also result, from the improvement in the voltage Capacitor
regulation due to the capacitor. Motor capacitors are often bank
six-terminal units, and a capacitor may be conveniently
connected directly across each motor phase winding.
Capacitor sizing is important, such that a leading power factor
does not occur under any load condition. If excess capacitance Relay
is applied to a motor, it may be possible for self-excitation to
occur when the motor is switched off or suffers a supply
failure. This can result in the production of a high voltage or in
mechanical damage if there is a sudden restoration of supply. Alarm Trip
Since most star/delta or auto-transformer starters other than Figure 18.13: Protection of double star capacitor banks
the ‘Korndorffer’ types involve a transitional break in supply, it
Some industrial loads such as arc furnaces involve large
is generally recommended that the capacitor rating should not
inductive components and correction is often applied using
exceed 85% of the motor magnetising reactive power.
very large high voltage capacitors in various configurations.

18.11.3 Capacitor Protection Another high voltage capacitor configuration is the ‘split phase’
arrangement where the elements making up each phase of the
When considering protection for capacitors, allowance should
capacitor are split into two parallel paths. Figure 18.15 shows
be made for the transient inrush current occurring on switch-
two possible connection methods for the relay. A differential
on, since this can reach peak values of around 20 times
relay can be applied with a current transformer for each
normal current. Switchgear for use with capacitors is usually
parallel branch. The relay compares the current in the split
de-rated considerably to allow for this. Inrush currents may be
phases, using sensitive current settings but also adjustable
limited by a resistor in series with each capacitor or bank of
compensation for the unbalance currents arising from initial
capacitors.
capacitor mismatch.

18-13
Protection & Automation Application Guide

Figure 18.14: Protection of capacitor banks

18-14
 Chapter 18 ⋅ Industrial and Commercial Power System Protection

18.12 EXAMPLES
A
In this section, examples of the topics dealt with in this
Chapter are considered.

I>
18.12.1 Fuse Co-ordination
An example of the application of fuses is based on the
arrangement in Figure 18.16(a). This shows an unsatisfactory
B
scheme with commonly encountered shortcomings. It can be
seen that fuses B, C and D will discriminate with fuse A, but
the 400A sub-circuit fuse E may not discriminate, with the
I> 500A sub-circuit fuse D at higher levels of fault current.

Rating
1000A A
C

Rating Rating Rating

B C D
I> 500A 500A 500A

Rating
Rating
Alarm Trip 400A
E 30A
each
(a)

A (a) Incorrect layout giving rise to problems in discrimination

Rating
1000A A

I>

Rating Rating Rating Rating

400A 100A 500A 500A
B
E B C D

I> Rating
30A

C F
(b) Correct layout and discrimination
Figure 18.16: Fuse protection: effect of layout on discrimination
I> The solution, shown in Figure 18.16(b), is to feed the 400A
circuit E direct from the busbars. The sub-circuit fuse D may
now have its rating reduced from 500A to a value, of say
100A, appropriate to the remaining sub-circuit. This
Alarm Trip arrangement now provides a discriminating fuse distribution
scheme satisfactory for an industrial system.
(b)
Figure 18.15: Differential protection of split phase capacitor banks However, there are industrial applications where

18-15
Protection & Automation Application Guide

discrimination is a secondary factor. In the application shown Fuse

in Figure 18.17, a contactor having a fault rating of 20kA 400A
1MVA
controls the load in one sub-circuit. A fuse rating of 630A is 2000/1A
selected for the minor fuse in the contactor circuit to give LV board fault
level = 30kA
protection within the through-fault capacity of the contactor.
3300/415V
I >> MCCB
I> 400A
800A
Figure 18.18: Network diagram for protection co-ordination example -
fuse/MCCB/relay

Discrimination is required between the relay and both the fuse

and MCCB up to the 40kA fault rating of the board. To begin
400V
with, the time/current characteristics of both the 400A fuse
and the MCCB are plotted in Figure 18.19.
630A
18.12.2.1 Determination of relay current setting
Fused contactor The relay current setting chosen must not be less than the full
load current level and must have enough margin to allow the
relay to reset with full load current flowing. The latter may be
determined from the transformer rating:

kVA
Auxiliary FLC =
Circuits kV × 3
Figure 18.17: Example of back-up protection 1000
=
The major fuse of 800A is chosen, as the minimum rating that 0.4 × 3
is greater than the total load current on the switchboard. = 1443 A
Discrimination between the two fuses is not obtained, as the
pre-arcing I2t of the 800A fuse is less than the total I2t of the With the CT ratio of 2000/1A and a relay reset ratio of 95% of
the nominal current setting, a current setting of at least 80%
630A fuse. Therefore, the major fuse will blow as well as the
would be satisfactory, to avoid tripping and/or failure to reset
minor one, for most faults so that all other loads fed from the
with the transformer carrying full load current. However, choice
switchboard will be lost. This may be acceptable in some
of a value at the lower end of this current setting range would
cases. In most cases, however, loss of the complete
move the relay characteristic towards that of the MCCB and
switchboard for a fault on a single outgoing circuit will not be
discrimination may be lost at low fault currents. It is therefore
acceptable, and the design will have to be revised.
prudent to select initially a relay current setting of 100%.
18.12.2 Grading of Fuses/MCCBs/Overcurrent Relays
18.12.2.2 Relay characteristic and time multiplier selection
An example of an application involving a moulded case circuit
An EI characteristic is selected for the relay to ensure
breaker, fuse and a protection relay is shown in Figure 18.18.
discrimination with the fuse (see Chapter 9 for details). From
A 1MVA 3.3kV/400V transformer feeds the LV board via a
Figure 18.19, it may be seen that at the fault level of 40kA the
circuit breaker, which is equipped with an Alstom MiCOM
fuse will operate in less than 0.01s and the MCCB operates in
P14x numerical relay having a setting range of 8-400% of
approximately 0.014s. Using a fixed grading margin of 0.4s,
rated current and fed from 2000/1A CTs.
the required relay operating time becomes 0.4 + 0.014 =
0.414s. With a CT ratio of 2000/1A, a relay current setting of
100%, and a relay TMS setting of 1.0, the extremely inverse
curve gives a relay operating time of 0.2s at a fault current of
40kA. This is too fast to give adequate discrimination and
indicates that the EI curve is too severe for this application.
Turning to the VI relay characteristic, the relay operation time
is found to be 0.71s at a TMS of 1.0. To obtain the required
relay operating time of 0.414s:

18-16
 Chapter 18 ⋅ Industrial and Commercial Power System Protection

0.414 starting current of 7 x FLC.

TMS setting = = 0.583
0.71
Use a TMS of 0.6, nearest available setting. 1.6 MVA
11/0.4kV
The use of a different form of inverse time characteristic makes Z=6.25%
it advisable to check discrimination at the lower current levels I>> I>>
also at this stage. At a fault current of 4kA, the relay will Relay C1 I> 2500/1 2500/1 I> Relay C2

operate in 8.1s, which does not give discrimination with the Trip Trip
2500/1 NO
MCCB. A relay operation time of 8.3s is required. To A1 A2 0.4kV
50kA rms
overcome this, the relay characteristic needs to be moved
away from the MCCB characteristic, a change that may be I>> Trip
I>
achieved by using a TMS of 0.625. The revised relay
characteristic is also shown in Figure 18.19. Relay B
I>>
300/1 Relay A
10.0

Motor cable

M
160kW
Characteristic for relay
Figure 18.20: Relay grading example for dual-fed switchboard

The transformer impedance is to IEC standards. The LV

switchgear and bus bars are fault rated at 50kA rms. To
1.0
simplify the analysis, only the phase-fault LV protection is
considered.
Operating time (s)

18.12.3.1 General considerations

Analysis of many substations configured as in Figure 18.20
shows that the maximum fault level and feeder load current is
Fuse obtained with the bus-section circuit breaker closed and one of
0.1 the infeeding CBs open. This applies so long as the
switchboard has a significant amount of motor load. The
contribution of motor load to the fault level at the switchboard
MCCB is usually larger than that from a single infeeding transformer,
as the transformer restricts the amount of fault current infeed
from the primary side. The three-phase break fault level at the
switchboard under these conditions is assumed to be 40kA
rms.
0.01
1000 10,000 100,000 Relays C are not required to have directional characteristics
Operating current (A) to 415V base
(see Section 9.14.3) as all three circuit breakers are only closed
Original relay characteristic momentarily during transfer from a single infeeding
Revised relay characteristic transformer to two infeeding transformers configuration. This
Figure 18.19: Grading curves for Fuse/MCCB/relay grading example transfer is normally an automated sequence, and the chance of
a fault occurring during the short period (of the order of 1s)
18.12.3 Protection of a Dual-Fed Substation when all three CBs are closed is taken to be negligibly small.
As an example of how numerical protection relays can be used Similarly, although this configuration gives the largest fault
in an industrial system, consider the typical large industrial level at the switchboard, it is not considered from either a
substation of Figure 18.20. Two 1.6MVA, 11/0.4kV switchboard fault rating or protection viewpoint.
transformers feeding a busbar whose bus-section CB is It is assumed that modern numerical relays are used. For
normally open. The LV system is solidly earthed. The largest simplicity, a fixed grading margin of 0.3s is used.
outgoing feeder is to a motor rated 160kW, 193kVA, and a

18-17
Protection & Automation Application Guide

18.12.3.2 Motor protection relay settings acceleration did not need to be considered.
From the motor characteristics given, the overcurrent relay The high-set setting needs to be above the full load current
settings (Relay A) can be found using the guidelines set out in and motor starting/re-acceleration transient current, but less
Chapter 19 as: than the fault current by a suitable margin. A setting of
Thermal element: 12.5kA is initially selected. A time delay of 0.3s has to used to
ensure grading with relay A at high fault current levels; both
• current setting: 300A relays A and B may see a current in excess of 25kA for faults
• time constant: 20 mins on the cable side of the CB feeding the 160kW motor. The
relay curves are shown in Figure 18.21.
Instantaneous element:
Relay A
• current setting: 2.32kA
Relay B

These are the only settings relevant to the upstream relays. Re-acceleration
Relay A setting
Relay B setting
18.12.3.3 Relay B settings 1000
Relay B settings are derived from consideration of the loading
and fault levels with the bus-section breaker between busbars 100
A1 and A2 closed. No information is given about the load split
between the two busbars, but it can be assumed in the 10

absence of definitive information that each busbar is capable

Time (s)

of supplying the total load of 1.6MVA. With fixed tap 1

transformers, the bus voltage may fall to 95% of nominal under

these conditions, leading to a load current of 2430A. The 0.1

IDMT current setting must be greater than this, to avoid relay

operation on normal load currents and (ideally) with aggregate 0.01

starting/re-acceleration currents. If the entire load on the 100 1000 10000 100000

busbar was motor load, an aggregate starting current in excess Current (A) referred to 0.4kV
of 13kA would occur, but a current setting of this order would Figure 18.21: Grading of relays A and B
be excessively high and lead to grading problems further
upstream. It is unlikely that the entire load is motor load 18.12.3.4 Relays C settings
(though this does occur, especially where a supply voltage of
The setting of the IDMT element of relays C1 and C2 has to be
690V is chosen for motors – an increasingly common practice)
suitable for protecting the busbar while grading with relay B.
or that all motors are started simultaneously (but simultaneous
The limiting condition is grading with relay B, as this gives the
re-acceleration may well occur). What is essential is that relay
longest operation time for relays C.
B does not issue a trip command under these circumstances –
i.e. the relay current/time characteristic is in excess of the The current setting has to be above that for relay B to achieve
current/time characteristic of the worst-case starting/re- full co-ordination, and a value of 3250A is suitable. The TMS
acceleration condition. It is therefore assumed that 50% of the setting using the SI characteristic is chosen to grade with that
total bus load is motor load, with an average starting current of relay B at a current of 12.5kA (relay B instantaneous
of 600% of full load current (= 6930A), and that re-acceleration setting), and is found to be 0.45. The high-set element must
takes 3s. A current setting of 3000A is therefore initially used. grade with that of relay B, so a time delay of 0.62sec is
The SI characteristic is used for grading the relay, as co- required. The current setting must be higher than that of relay
ordination with fuses is not required. The TMS is required to B, so use a value of 15kA. The final relay grading curves and
be set to grade with the thermal protection of relay A under settings are shown in Figure 18.22.
‘cold’ conditions, as this gives the longest operation time of
Relay A, and the re-acceleration conditions. A TMS value of
0.41 is found to provide satisfactory grading, being dictated by
the motor starting/re-acceleration transient. Adjustment of
both current and TMS settings may be required depending on
the exact re-acceleration conditions. Note that lower current
and TMS settings could be used if motor starting/re-

18-18
 Chapter 18 ⋅ Industrial and Commercial Power System Protection

Re-acceleration Relay A 18.13 REFERENCES

Relay A setting
Relay B setting
Relay B
[18.1] Overcurrent Relay Co-ordination for Double Ended
Relay C setting Relay C
Substations. George R Horcher. IEEE. Vol. 1A-14
Relay A
Parameter
Ith
Value Parameter
300A Time Const
Value
1200s
Parameter Value Parameter
I>> 2320A tdinst
Value
0
No. 6 1978.
Relay B I> 2500A TMS 0.175 I>> 12500A tdinst 0.32s
Relay C I> 2750A TMS 0.25 I>> 15000 tdinst 0.62s

(a) Relay settings

1000

100

10
Time (s)

I> and I>>

I> and I>>
0.1

Ith and I>>

0.01
100 1000 10000 100000
Current (A) referred to 0.4kV

(b) Grading curves

Figure 18.22: Final relay grading curves

18.12.3.5 Comments on grading

While the above grading may appear satisfactory, the
protection on the primary side of the transformer has not been
considered. IDMT protection at this point will have to grade
with relays C and with the through-fault short-time withstand
curves of the transformer and cabling. This may result in
excessively long operation times. Even if the operation time at
the 11kV level is satisfactory, there is probably a Utility infeed
to consider, which will involve a further set of relays and
another stage of time grading, and the fault clearance time at
the Utility infeed will almost certainly be excessive. One
solution is to accept a total loss of supply to the 0.4kV bus
under conditions of a single infeed and bus section CB closed.
This is achieved by setting relays C such that grading with
relay B does not occur at all current levels, or omitting relay B
from the protection scheme. The argument for this is that
network operation policy is to ensure loss of supply to both
sections of the switchboard does not occur for single
contingencies. As single infeed operation is not normal, a
contingency (whether fault or maintenance) has already
occurred, so that a further fault causing total loss of supply to
the switchboard through tripping of one of relays B is a second
contingency. Total loss of supply is therefore acceptable. The
alternative is to accept a lack of discrimination at some point
on the system, as already noted in Chapter 9. Another solution
is to employ partial differential protection to remove the need
for Relay A, but this is seldom used. The strategy adopted will
depend on the individual circumstances.

18-19
AC MOTOR PROTECTION
Chapter 19
AC MOTOR PROTECTION

19.1 Introduction 19.1 INTRODUCTION

19.2 Modern Relay Design There are a wide range of a.c. motors and motor characteristics
19.3 Thermal (Overload) Protection in existence, because of the numerous duties for which they
are used. All motors need protection, but fortunately, the more
19.4 Start/Stall Protection fundamental problems affecting the choice of protection are
19.5 Short-Circuit Protection independent of the type of motor and the type of load to which
it is connected. There are some important differences between
19.6 Earth Fault Protection the protection of induction motors and synchronous motors,
19.7 Negative Phase Sequence Protection and these are fully dealt with in section 19.3.
19.8 Faults in Rotor Windings Motor characteristics must be carefully considered when
applying protection; while this may be regarded as stating the
19.9 RTD Temperature Detection obvious, it is emphasised because it applies more to motors than
19.10 Bearing Failures to other items of power system plant. For example, the starting
and stalling currents/times must be known when applying
19.11 Undervoltage Protection overload protection, and furthermore the thermal withstand of
19.12 Loss-of-load Protection the machine under balanced and unbalanced loading must be
clearly defined.
19.13 Additional Protection for
Synchronous Motors The conditions for which motor protection is required can be
divided into two broad categories: imposed external conditions
19.14 Motor Protection Examples and internal faults. Table 19.1 provides details of most likely
faults that require protection

External Faults Internal faults

Unbalanced supplies Bearing failures
Undervoltages Winding faults
Single phasing Overloads
Reverse phase sequence

Figure 18.1: Large modern industrial plant

19.2 MODERN RELAY DESIGN

The design of a modern numerical motor protection relay must
be adequate to cater for the protection needs of any one of the
vast range of motor designs in service, many of the designs
having no permissible allowance for overloads. A relay offering
comprehensive protection will have the following set of features:

SYNCHRONOUS AND ASYNCHRONOUS MOTORS

.. thermal protection
.. extended start protection
.. stalling protection
.. number of starts limitation
.. short circuit protection

GEGridSolutions.com 19-1
Protection & Automation Application Guide

• earth fault protection The temperature T at any instant is given by:

• winding RTD measurement/trip
 −
t

• negative sequence current detection 
T = Tmax 1 − e τ 

 
• undervoltage protection
• loss-of-load protection where:

• auxiliary supply supervision Tmax = final steady state temperature

Synchronous motors only τ = heating time constant
• out-of-step protection Temperature rise is proportional to the current squared:
• loss of supply protection
 −
t

In addition, relays may offer options such as circuit breaker  2
T = KI 1 − e τ 
R 
condition monitoring as an aid to maintenance.  
Manufacturers may also offer relays that implement a reduced
where:
functionality to that given above where less comprehensive
protection is warranted (e.g. induction motors of low rating). I R = current which, if flowing continuously, produces
The following sections examine each of the possible failure temperature Tmax in the motor
modes of a motor and discuss how protection may be applied Therefore, it can be shown that, for any overload current I ,
to detect that mode.
the permissible time t for this current to flow is:
19.3 THERMAL (OVERLOAD) PROTECTION  
The majority of winding failures are either indirectly or directly  
caused by overloading (either prolonged or cyclic), operation  1 
t = t log e  2 
on unbalanced supply voltage, or single phasing, which all lead  1 −  I R   
through excessive heating to the deterioration of the winding    I   
insulation until an electrical fault occurs. The generally  
accepted rule is that insulation life is halved for each 10ºC rise In general, the supply to which a motor is connected may
in temperature above the rated value, modified by the length of contain both positive and negative sequence components, and
time spent at the higher temperature. As an electrical both components of current give rise to heating in the motor.
machine has a relatively large heat storage capacity, it follows Therefore, the thermal replica should take into account both of
that infrequent overloads of short duration may not adversely these components, a typical equation for the equivalent
affect the machine. However, sustained overloads of only a current being:
few percent may result in premature ageing and insulation
failure. I eq = (I 1
2
+ KI 22 )
Furthermore, the thermal withstand capability of the motor is where
affected by heating in the winding prior to a fault. It is
therefore important that the relay characteristic takes account I1 = positive sequence current
of the extremes of zero and full-load pre-fault current known
respectively as the 'Cold' and 'Hot' conditions.
I 2 = negative sequence current

The variety of motor designs, diverse applications, variety of K = negative sequence rotor resistance / positive sequence
possible abnormal operating conditions and resulting modes of rotor resistance at rated speed.
failure result in a complex thermal relationship. A generic A typical value of K is 3.
mathematical model that is accurate is therefore impossible to
Finally, the thermal replica model needs to take into account
create. However, it is possible to develop an approximate
the fact that the motor will tend to cool down during periods of
model if it is assumed that the motor is a homogeneous body,
light load, and the initial state of the motor. The motor will
creating and dissipating heat at a rate proportional to
have a cooling time constant, τ r , that defines the rate of
temperature rise. This is the principle behind the ‘thermal
replica’ model of a motor used for overload protection. cooling. Hence, the final thermal model can be expressed as:

19-2
 Chapter 19 ⋅ AC Motor Protection

t = Te1 log e
(K − A )
2 2 100 000

(K − 1)
2

Equation 19.1
10 000
where: Te1 = Te2 = 60min

Te1 = heating time constant Te1 = Te2 = 54min

Te1 = Te2 = 48min
1000
I eq Te1 = Te2 = 42min
K=
I th Te1 = Te2 = 36min

A 2 = initial state of motor (cold or hot) 100

Operating time (seconds)

I th = thermal setting current
Equation 19.1 takes into account the ‘cold’ and ‘hot’ 10 Te1 = Te2 = 30min

characteristics defined in IEC 60255, part 8. Te1 = Te2 = 24min

Te1 = Te2 = 18min
Some relays may use a dual curve characteristic for the Te1 = Te2 = 12min
Te1 = Te2 = 6min
heating time constant, and hence two values of the heating 1
Te1 = Te2 = 1min
time constant are required. Switching between the two values
takes place at a pre-defined motor current. This may be used
to obtain better tripping performance during starting on 0
motors that use a star-delta starter. During starting, the 1 10

motor windings carry full line current, while in the ‘run’ Thermal equivalent current Ieq in terms of the
condition, they carry only 57% of the current seen by the relay. current thermal threshold Iq>
Similarly, when the motor is disconnected from the supply, the Figure 19.1: Thermal overload characteristic curves from cold – initial
heating time constant Te1 is set equal to the cooling time thermal state 0%

constant Te 2 . 19.4 START/STALL PROTECTION

Since the relay should ideally be matched to the protected When a motor is started, it draws a current well in excess of
motor and be capable of close sustained overload protection, a full load rating throughout the period that the motor takes to
wide range of relay adjustment is desirable together with good run-up to speed. While the motor starting current reduces
accuracy and low thermal overshoot. somewhat as motor speed increases, it is normal in protection
practice to assume that the motor current remains constant
Typical relay setting curves are shown in Figure 19.1.
throughout the starting period. The starting current will vary
depending on the design of the motor and method of starting.
For motors started DOL (direct-on-line), the nominal starting
current can be 4-8 times full-load current. However, when a
1
star-delta starter is used, the line current will only be of
3
the DOL starting current.
Should a motor stall whilst running, or fail to start, due to
excessive loading, the motor will draw a current equal to its
locked rotor current. It is not therefore possible to distinguish
between a stall condition and a healthy start solely on the basis
of the current drawn. Discrimination between the two
conditions must be made based on the duration of the current
drawn. For motors where the starting time is less than the
safe stall time of the motor, protection is easy to arrange.
However, where motors are used to drive high inertia loads,

19-3
Protection & Automation Application Guide

the stall withstand time can be less than the starting time. In 100

these cases, an additional means must be provided to enable

discrimination between the two conditions to be achieved.
Relay current setting

19.4.1 Excessive Start Time/Locked Rotor Protection

Motor starting current
A motor may fail to accelerate from rest for a number of
Relay time setting
reasons: 10

• loss of a supply phase

Time (s)
• mechanical problems
• low supply voltage
• excessive load torque
1
A large current will be drawn from the supply, and cause
extremely high temperatures to be generated within the motor.
This is made worse by the fact that the motor is not rotating,
and hence no cooling due to rotation is available. Winding
damage will occur very quickly – either to the stator or rotor
windings depending on the thermal limitations of the 0.1
particular design (motors are said to be stator or rotor limited 0.1 1 10
Current (p.u.)
in this respect). The method of protection varies depending on
whether the starting time is less than or greater than the safe Figure 19.2: Relay setting for successful start: start time<stall time
stall time. In both cases, initiation of the start may be sensed
19.4.1.2 Start time ≥ safe stall time
by detection of the closure of the switch in the motor feeder
(contactor or CB) and optionally current rising above a starting For this condition, a definite time overcurrent characteristic by
current threshold value – typically 200% of motor rated itself is not sufficient, since the time delay required is longer
current. For the case of both conditions being sensed, they than the maximum time that the motor can be allowed to
may have to occur within a narrow aperture of time for a start carry starting current safely. An additional means of detection
to be recognised. of rotor movement, indicating a safe start, is required. A
speed-sensing switch usually provides this function. Detection
Special requirements may exist for certain types of motors of a successful start is used to select the relay timer used for
installed in hazardous areas (e.g. motors with type of the safe run-up time of the motor. This time can be longer
protection EEx ‘e’) and the setting of the relay must take these than the safe stall time, as there is both a (small) decrease in
into account. Sometimes a permissive interlock for machine current drawn by the motor during the start and the rotor fans
pressurisation (on EEx ‘p’ machines) may be required, and this begin to improve cooling of the machine as it accelerates. If a
can be conveniently achieved by use of a relay digital input and start is sensed by the relay through monitoring current and/or
the in-built logic capabilities. start device closure, but the speed switch does not operate, the
relay element uses the safe stall time setting to trip the motor
19.4.1.1 Start time < safe stall time before damage can occur. Figure 19.3(a) illustrates the
Protection is achieved by use of a definite time overcurrent principle of operation for a successful start, and Figure 19.3(b)
characteristic, the current setting being greater than full load for an unsuccessful start.
current but less than the starting current of the machine. The
time setting should be a little longer than the start time, but
less than the permitted safe starting time of the motor. Figure
19.2 illustrates the principle of operation for a successful start.

19-4
 Chapter 19 ⋅ AC Motor Protection

The stall protection would be expected to operate and defeat

1 the object of the re-acceleration scheme. A motor protection
CB closed
0 relay will therefore recognise the presence of a voltage dip and
Time recovery, and inhibit stall protection for a defined period. The
1
Current undervoltage protection element (section 19.11) can be used
0
Time to detect the presence of the voltage dip and inhibit stall
Speed 1 protection for a set period after voltage recovery. Protection
switch against stalled motors in case of an unsuccessful re-
information 0
Time acceleration is therefore maintained. The time delay setting is
Trip 1
dependent on the re-acceleration scheme adopted and the
command
0 characteristics of individual motors. It should be established
Time after performing a transient stability study for the re-
(a) Successful start acceleration scheme proposed.

1 19.4.3 Number of Starts Limitation

CB closed
0 Any motor has a restriction on the number of starts that are
Time allowed in a defined period without the permitted winding, etc.
1
Current temperatures being exceeded. Starting should be blocked if
0 the permitted number of starts is exceeded. The situation is
Time
Speed 1 complicated by the fact the number of permitted ‘hot’ starts in
switch
information 0 a given period is less than the number of ‘cold’ starts, due to
Time the differing initial temperatures of the motor. The relay must
Trip 1 Stall time setting
maintain a separate count of ‘cold’ and ‘hot’ starts. By
command
0 making use of the data held in the motor thermal replica, ‘hot’
Time and ‘cold’ starts can be distinguished.
(b) Unsuccessful start
To allow the motor to cool down between starts, a time delay
Figure 19.3: Relay settings for start time> stall time may be specified between consecutive starts (again
distinguishing between ‘hot’ and ‘cold’ starts). The start inhibit
19.4.2 Stall Protection is released after a time determined by the motor specification.
Should a motor stall when running or be unable to start
The overall protection function is illustrated in Figure 19.4.
because of excessive load, it will draw a current from the
supply equivalent to the locked rotor current. It is obviously Motor start
desirable to avoid damage by disconnecting the machine as
quickly as possible if this condition arises. On

Motor stalling can be recognised by the motor current

exceeding the start current threshold after a successful start – Off
i.e. a motor start has been detected and the motor current has
dropped below the start current threshold within the motor
Supervising time
safe start time. A subsequent rise in motor current above the Start
inhibition
motor starting current threshold is then indicative of a stall Supervising time
condition, and tripping will occur if this condition persists for
On
greater than the setting of the stall timer. An instantaneous
overcurrent relay element provides protection.
Off
In many systems, transient supply voltage loss (typically up to
tn Inhibit
2 seconds) does not result in tripping of designated motors. start
tn = time associated with
They are allowed to re-accelerate upon restoration of the the set maximum
time
number of starts
supply. During re-acceleration, they draw a current similar to
Start lockout
the starting current for a period that may be several seconds.
It is thus above the motor stall relay element current threshold. Figure 19.4: Number of starts limitation

19-5
Protection & Automation Application Guide

In this example, the maximum number of starts within the phase, etc.) or the cause (cyclic overheating, etc.), the
Supervising Time has been reached, therefore the Inhibit Start presence of the surrounding metallic frame and casing will
Time is initiated. The remaining time is greater than the ensure that it rapidly develops into a fault involving earth.
Inhibit Start Time, so the start inhibition remains for a duration Therefore, provision of earth fault protection is very important.
equal to the supervising time minus the tn. The type and sensitivity of protection provided depends largely
on the system earthing, so the various types will be dealt with
19.5 SHORT-CIRCUIT PROTECTION in turn. It is common, however, to provide both instantaneous
Motor short-circuit protection is often provided to cater for and time-delayed relay elements to cater for major and slowly
major stator winding faults and terminal flashovers. Because developing faults.
of the relatively greater amount of insulation between phase
windings, faults between phases seldom occur. As the stator 19.6.1 Solidly-Earthed System
windings are completely enclosed in grounded metal the fault Most LV systems fall into this category, for reasons of
would very quickly involve earth, which would then operate the personnel safety. Two types of earth fault protection are
instantaneous earth fault protection. A single definite time commonly found – depending on the sensitivity required.
overcurrent relay element is all that is required for this
For applications where a sensitivity of > 20% of motor
purpose, set to about 125% of motor starting current. The time
continuous rated current is acceptable, conventional earth
delay is required to prevent spurious operation due to CT spill
fault protection using the residual CT connection of Figure
currents, and is typically set at 100ms. If the motor is fed from
19.5 can be used. A lower limit is imposed on the setting by
a fused contactor, co-ordination is required with the fuse, and
possible load unbalance and/or (for HV systems) system
this will probably involve use of a long time delay for the relay
capacitive currents.
element. Since the object of the protection is to provide rapid
fault clearance to minimise damage caused by the fault, the Upstream

protection is effectively worthless in these circumstances. It is a

therefore only provided on motors fed via circuit breakers.

Differential (unit) protection may be provided on larger HV
motors fed via circuit breakers to protect against phase-phase c b
a b c
and phase-earth faults, particularly where the power system is
resistance-earthed. The differential protection can be made Direction
of current
quite sensitive allowing early detection of faults, thus damage Ia

to the motor can be minimised. The normal definite time

overcurrent protection would not be sufficiently sensitive, and
sensitive earth fault protection may not be provided. The user Ib
may wish to avoid the detailed calculations required of
capacitance current in order to set sensitive non-directional
earth fault overcurrent protection correctly on HV systems Ic
(Chapter 9) or there may be no provision for a VT to allow
application of directional sensitive earth fault protection. There
is still a lower limit to the setting that can be applied, due to Ia+Ib+Ic
spill currents from CT saturation during starting, while on
some motors, neutral current has been found to flow during
starting, even with balanced supply voltages, that would cause IED

the differential protection to operate. For details on the Downstream

application of differential protection, refer to Chapter 10. Figure 19.5: Residual CT connection for earth fault protection
However, non-directional earth fault overcurrent protection
Care must be taken to ensure that the relay does not operate
will normally be cheaper in cases where adequate sensitivity
from the spill current resulting from unequal CT saturation
can be provided.
during motor starting, where the high currents involved will
almost certainly saturate the motor CT’s. It is common to use
19.6 EARTH FAULT PROTECTION a stabilising resistor in series with the relay, with the value
One of the most common faults to occur on a motor is a stator being calculated using the formula:
winding fault. Whatever the initial form of the fault (phase-

19-6
 Chapter 19 ⋅ AC Motor Protection

I st
Rstab = (Rct + kRl + Rr )
I0 Contact
breaking
Equation 19.2 capacity
Fuse
where:

Time
I st = starting current referred to CT secondary

I 0 = relay earth fault setting (A) Earth Fault relay

Rstab =stabilising resistor value (ohms)

Rct =dc resistance of CT secondary (ohms)

Current
Rl =CT single lead resistance (ohms) (a) Incorrect

Rr = relay resistance (ohms)

k = CT connection factor (1 for star point at CT, 2 for star
Contact
point at relay). breaking
The effect of the stabilising resistor is to increase the effective capacity
setting of the relay under these conditions, and hence delay Fuse
tripping. When a stabilising resistor is used, the tripping
Time

characteristic should normally be instantaneous. An

alternative technique, avoiding the use of a stabilising resistor Earth Fault relay
is to use a definite time delay characteristic. The time delay
used will normally have to be found by trial and error, as it
must be long enough to prevent maloperation during a motor
start, but short enough to provide effective protection in case
Current
of a fault.
(b) Correct
Co-ordination with other devices must also be considered. A
Figure 19.6: Grading of relay with fused contactor
common means of supplying a motor is via a fused contactor.
The contactor itself is not capable of breaking fault current If a more sensitive relay setting is required, it is necessary to
beyond a certain value, which will normally be below the use a core-balance CT (CBCT). This is a ring type CT, through
maximum system fault current – reliance is placed on the fuse which all phases of the supply to the motor are passed, plus
in these circumstances. As a trip command from the relay the neutral on a four-wire system. The turns ratio of the CT is
instructs the contactor to open, care must be taken to ensure no longer related to the normal line current expected to flow,
that this does not occur until the fuse has had time to operate. so can be chosen to optimise the pick-up current required.
Figure 19.6(a) illustrates incorrect grading of the relay with the Magnetising current requirements are also reduced, with only
fuse, the relay operating first for a range of fault currents in a single CT core to be magnetised instead of three, thus
excess of the contactor breaking capacity. Figure 19.6(b) enabling low settings to be used. Figure 19.7 illustrates the
illustrates correct grading. To achieve this, it may require the application of a core-balance CT, including the routing of the
use of an intentional definite time delay in the relay. cable sheath to ensure correct operation in case of core-sheath
cable faults.

19-7
Protection & Automation Application Guide

phases and it is necessary to locate and clear the fault before a

19.6.2 Resistance-Earthed Systems
second occurs. Two methods are possible:
These are commonly found on HV systems, where the
• detection of the resulting unbalance in system charging
intention is to limit damage caused by earth faults through
currents
limiting the earth-fault current that can flow. Two methods of
resistance earthing are commonly used: • residual overvoltage.

19.6.2.1 Low resistance earthing 19.6.3.1 System charging current unbalance

In this method, the value of resistance is chosen to limit the Sensitive earth fault protection using a core-balance CT is
fault current to a few hundred amps – values of 200A-400A required for this scheme. For details of the technique, please
being typical. With a residual connection of line CT’s, the refer to Section 9.18.2.
minimum sensitivity possible is about 10% of CT rated primary
current, due to the possibility of CT saturation during starting. 19.6.3.2 Residual voltage method
For a core-balance CT, the sensitivity that is possible using a A single earth fault results in a rise in the voltage between
simple non-directional earth fault relay element is limited to system neutral and earth, which may be detected by a relay
three times the steady-state charging current of the feeder. measuring the residual voltage of the system (normally zero for
The setting should not be greater than about 30% of the a perfectly balanced, healthy system). Thus, no CTs are
minimum earth fault current expected. Other than this, the required, and the technique may be useful where provision of
considerations in respect of settings and time delays are as for an extensive number of core-balance CTs is impossible or
solidly earthed systems. difficult, due to physical constraints or on cost grounds. The
VTs used must be suitable for the duty, thus 3-limb, 3-phase
19.6.2.2 High resistance earthing VTs are not suitable, and the relay usually has alarm and trip
In some HV systems, high resistance earthing is used to limit settings, each with adjustable time delays. The setting voltage
the earth fault current to a few amps. In this case, the system must be calculated from knowledge of system earthing and
capacitive charging current will normally prevent conventional impedances, an example for a resistance-earthed system is
sensitive earth fault protection being applied, as the magnitude shown in Figure 19.7.
of the charging current will be comparable with the earth fault Grading of the relays must be carried out with care, as the
current in the event of a fault. The solution is to use a residual voltage will be detected by all relays in the affected
sensitive directional earth fault relay. A core balance CT is section of the system. Grading has to be carried out with this
used in conjunction with a VT measuring the residual voltage in mind, and will generally be on a time basis for providing
of the system, with a relay characteristic angle setting of alarms (1st stage), with a high set definite time trip second
+45ºC (see Chapter 9 for details). The VT must be suitable for stage to provide backup.
the relay and therefore the relay manufacturer should be
consulted over suitable types – some relays require that the VT
must be able to carry residual flux and this rules out use of a 3-
limb, 3-phase VT. A setting of 125% of the single phase
capacitive charging current for the whole system is possible
using this method. The time delay used is not critical but must
be fast enough to disconnect equipment rapidly in the event of
a second earth fault occurring immediately after the first.
Minimal damage is caused by the first fault, but the second
effectively removes the current limiting resistance from the
fault path leading to very large fault currents.
An alternative technique using residual voltage detection is
also possible, and is described in the next section.

19.6.3 Insulated Earth System

Earth fault detection presents problems on these systems since
no earth fault current flows for a single earth fault. However,
detection is still essential as overvoltages occur on sound

19-8
 Chapter 19 ⋅ AC Motor Protection

S R F 0.5
E
 R1 p + R2 p 
2

( )
ZS ZL
2
N ∼ =   + X 1 p + X 2' p 
ZE
 (2 − s )  
A-G

Hence, at standstill (s=1.0), impedance

S VA-G
S
R VA-G
S
R
[(
= R1 p + R2' p ) + (X
2
1p + X 2' p )]
2 0.5

G,F G,F G,F

VC-G VB-G VC-G VB-G VC-G VB-G

The motor negative sequence impedance at slip s
0.5
 R2' n 
2

VRES VB-G VRES
VB-G
VRES VB-G =  R1n +
S 
(
 + X 1n + X 2' n )2

 
VA-G VA-G

VC-G VC-G VC-G

and, at normal running speed, the impedance

Z SO + 3Z E
V RES × 3E 0.5
2Z S1 + Z SO + 2Z L1 + Z LO + 3Z E  R2' n 
2

Figure 19.7: Residual voltage earth-fault protection for resistance- =  R1n +
2 
(
 + X 1n + X 2' n )2

earthed system  

19.6.4 Petersen Coil Earthed System

where:
Earthing of a HV power system using a reactor equal to the
system shunt capacitance is known as Petersen Coil (or • suffix p indicates positive sequence quantities
resonant coil) earthing. With this method, a single earth fault and
results in zero earth fault current flowing (for perfect balance
between the earthing inductance and system shunt • suffix n indicates negative sequence quantities
capacitance), and hence the system can be run in this state for
a substantial period of time while the fault is located and R1 + R’2 j(X1 + X’2)
corrected. The detailed theory and protection method is
explained in Section 9.19.
[(1 - s)/s] x R’2
19.7 NEGATIVE PHASE SEQUENCE
PROTECTION
Negative phase sequence current is generated from any
unbalanced voltage condition, such as unbalanced loading, (a) Positive phase sequence equivalent circuit
loss of a single phase, or single-phase faults. The latter will
normally be detected by earth-fault protection, however, a R1 + R’2 j(X1 + X’2)
fault location in a motor winding may not result in the earth
fault protection operating unless it is of the sensitive variety.
The actual value of the negative sequence current depends on [(s - 1)/(2 - s)] x R’2
the degree of unbalance in the supply voltage and the ratio of
the negative to the positive sequence impedance of the
machine. The degree of unbalance depends on many factors,
(b) Negative phase sequence equivalent circuit
but the negative sequence impedance is more easily
Figure 19.8: Induction motor equivalent circuits
determined. Considering the classical induction motor
equivalent circuit with magnetising impedance neglected of The resistance can be neglected as it is small compared with
Figure 19.8: the reactance. Thus the negative sequence reactance at
running speed is approximately equal to the positive sequence
Motor positive sequence impedance at slip s:

19-9
Protection & Automation Application Guide

reactance at standstill. An alternative more meaningful way of 19.8 FAULTS IN ROTOR WINDINGS
expressing this is: On wound rotor machines, some degree of protection against
positive seq. impedance starting current faults in the rotor winding can be given by an instantaneous
= stator current overcurrent relay element. As the starting
negative seq. impedance rated current
current is normally limited by resistance to a maximum of
A typical LV motor starting current is 6 x full load current twice full load, the instantaneous unit can safely be set to
(FLC). Therefore, a 5% negative sequence voltage (due to, say, about three times full load if a slight time delay of
unbalanced loads on the system) would produce a 30% approximately 30 milliseconds is incorporated. It should be
negative sequence current in the machine, leading to excessive noted that faults occurring in the rotor winding would not be
heating. For the same motor, negative sequence voltages in detected by any differential protection applied to the stator.
excess of 17% will result in a negative sequence current larger
than rated full load current. 19.9 RTD TEMPERATURE DETECTION
Negative sequence current is at twice supply frequency. Skin RTDs are used to measure temperatures of motor windings or
effect in the rotor means that the heating effect in the rotor of shaft bearings. A rise in temperature may denote overloading
a given negative sequence current is larger than the same of the machine, or the beginning of a fault in the affected part.
positive sequence current. Thus, negative sequence current A motor protection relay will therefore usually have the
may result in rapid heating of the motor. Larger motors are capability of accepting a number of RTD inputs and internal
more susceptible in this respect, as the rotor resistance of such logic to initiate an alarm and/or trip when the temperature
machines tends to be higher. Protection against negative exceeds the appropriate setpoint(s). Occasionally, HV motors
sequence currents is therefore essential. are fed via a unit transformer, and in these circumstances,
some of the motor protection relay RTD inputs may be
Modern motor protection relays have a negative sequence assigned to the transformer winding temperature RTDs, thus
current measurement capability, in order to provide such providing overtemperature protection for the transformer
protection. The level of negative sequence unbalance depends without the use of a separate relay.
largely upon the type of fault. For loss of a single phase at
start, the negative sequence current will be 50% of the normal 19.10 BEARING FAILURES
starting current. It is more difficult to provide an estimate of
the negative sequence current if loss of a phase occurs while There are two types of bearings to be considered: the anti-
running. This is because the impact on the motor may vary friction bearing (ball or roller), used mainly on small motors
widely, from increased heating to stalling due to the reduced (up to around 350kW), and the sleeve bearing, used mainly on
torque available. large motors.

A typical setting for negative sequence current protection must The failure of ball or roller bearings usually occurs very quickly,
take into account the fact that the motor circuit protected by causing the motor to come to a standstill as pieces of the
the relay may not be the source of the negative sequence damaged roller get entangled with the others. There is
current. Time should be allowed for the appropriate protection therefore very little chance that any relay operating from the
to clear the source of the negative sequence current without input current can detect bearing failures of this type before the
introducing risk of overheating to the motor being considered. bearing is completely destroyed. Therefore, protection is
This indicates a two stage tripping characteristic, similar in limited to disconnecting the stalled motor rapidly to avoid
principle to overcurrent protection. A low-set definite time- consequential damage. Refer to Section 19.4 on stall
delay element can be used to provide an alarm, with an IDMT protection for details of suitable protection.
element used to trip the motor in the case of higher levels of Failure of a sleeve bearing can be detected by means of a rise
negative sequence current, such as loss-of-phase conditions at in bearing temperature. The normal thermal overload relays
start, occurring. Typical settings might be 20% of CT rated cannot give protection to the bearing itself but will operate to
primary current for the definite time element and 50% for the protect the motor from excessive damage. Use of RTD
IDMT element. The IDMT time delay has to be chosen to temperature detection, as noted in Section 19.9, can provide
protect the motor while, if possible, grading with other suitable protection, allowing investigation into the cause of the
negative sequence relays on the system. Some relays may not bearing running hot prior to complete failure.
incorporate two elements, in which case the single element
should be set to protect the motor, with grading being a 19.11 UNDERVOLTAGE PROTECTION
secondary consideration. Motors may stall when subjected to prolonged undervoltage

19-10
 Chapter 19 ⋅ AC Motor Protection

conditions. Transient undervoltages will generally allow a motor torque occurs. Other conditions that may cause this
motor to recover when the voltage is restored, unless the condition are a fall in the applied voltage to stator or field
supply is weak. windings. Such a fall may not need to be prolonged, a voltage
dip of a few seconds may be all that is required. An out-of-
Motors fed by contactors have inherent undervoltage
step condition causes the motor to draw excessive current and
protection, unless a latched contactor is used. Where a
generate a pulsating torque. Even if the cause is removed
specific undervoltage trip is required, a definite time
promptly, the motor will probably not recover synchronism, but
undervoltage element is used. If two elements are provided,
eventually stall. Hence, it must be disconnected from the
alarm and trip settings can be used. An interlock with the
supply.
motor starter is required to block relay operation when the
starting device is open, otherwise a start will never be The current drawn during an out-of-step condition is at a very
permitted. The voltage and time delay settings will be system low power factor. Hence a relay element that responds to low
and motor dependent. They must allow for all voltage dips power factor can be used to provide protection. The element
likely to occur on the system during transient faults, starting of must be inhibited during starting, when a similar low power
motors, etc. to avoid spurious trips. As motor starting can factor condition occurs. This can conveniently be achieved by
result in a voltage depression to 80% of nominal, the voltage use of a definite time delay, set to a value slightly in excess of
setting is likely to be below this value. Re-acceleration is the motor start time.
normally possible for voltage dips lasting between 0.5-2
The power factor setting will vary depending on the rated
seconds, depending on system, motor and drive
power factor of the motor. It would typically be 0.1 less than
characteristics, and therefore the time delay will be set bearing
the motor rated power factor i.e. for a motor rated at 0.85
these factors in mind.
power factor, the setting would be 0.75.

19.12 LOSS-OF-LOAD PROTECTION 19.13.2 Protection against Sudden Restoration of

Loss-of-load protection has a number of possible functions. It Supply
can be used to protect a pump against becoming unprimed, or
If the supply to a synchronous motor is interrupted, it is
to stop a motor in case of a failure in a mechanical
essential that the motor breaker be tripped as quickly as
transmission (e.g. conveyor belt), or it can be used with
possible if there is any possibility of the supply being restored
synchronous motors to protect against loss-of-supply
automatically or without the machine operator’s knowledge.
conditions. Implementation of the function is by a low forward
This is necessary in order to prevent the supply being restored
power relay element, interlocked with the motor starting
out of phase with the motor generated voltage.
device to prevent operation when the motor is tripped and thus
preventing a motor start. When starting against a very low Two methods are generally used to detect this condition, in
load (e.g. a compressor), the function may also need to be order to cover different operating modes of the motor.
inhibited for the duration of the start, to prevent maloperation.
19.13.2.1 Underfrequency protection
The setting will be influenced by the function to be performed
by the relay. A time delay may be required after pickup of the The underfrequency relay element will operate in the case of
element to prevent operation during system transients. This is the supply failing when the motor is on load, which causes the
especially important for synchronous motor loss-of supply motor to decelerate quickly. Typically, two elements are
protection. provided, for alarm and trip indications. The underfrequency
setting value needs to consider the power system
19.13 ADDITIONAL PROTECTION FOR characteristics. In some power systems, lengthy periods of
operation at frequencies substantially below normal occur, and
SYNCHRONOUS MOTORS
should not result in a motor trip. The minimum safe operating
The differences in construction and operational characteristics frequency of the motor under load conditions must therefore
of synchronous motors mean that additional protection is be determined, along with minimum system frequency.
required for these types of motor. This additional protection is
discussed in the following sections. 19.13.2.2 Low forward power protection
This can be applied in conjunction with a time delay to detect a
19.13.1 Out-of-Step Protection
loss-of-supply condition when the motor may share a busbar
A synchronous motor may decelerate and lose synchronism with other loads. The motor may attempt to supply the other
(fall out-of-step) if a mechanical overload exceeding the peak loads with power from the stored kinetic energy of rotation. A

19-11
Protection & Automation Application Guide

low-forward-power relay can detect this condition. A time multiple of τ 1 . With a cooling time constant of 75min,
delay will be required to prevent operation during system
τ R = 3 ×τ1 .
transients leading to momentary reverse power flow in the
motor.
19.14.1.2 Short circuit protection
19.14 MOTOR PROTECTION EXAMPLES Following the recommendations of Section 19.5, with a
starting current of 550% of full load current, the short-circuit
This section gives examples of the protection of HV and LV
element is set to 1.25 x 5.5 x 211A = 1450A.
induction motors.
In terms of the relay nominal current In, the setting value is
19.14.1 Protection of a HV Motor 1450
= 5.8 I n
Table 19.2 gives relevant parameters of a HV induction motor 250
to be protected. Using an Alstom MiCOM P241 motor
There is a minimum time delay of 100ms for currents up to
protection relay, the important protection settings are
120% of setting to allow for transient CT saturation during
calculated in the following sections.
starting and 40ms above this current value. These settings are
Quantity Value satisfactory.
Rated output 1000kW CMR
Rated Voltage 3.3kV
19.14.1.3 Earth-fault protection
Rated frequency 50Hz It is assumed that no CBCT is fitted. A typical setting of 30% of
Rated power factor/efficiency 0.9/0.92
motor rated current is used, leading to an earth fault relay
setting of
Stall withstand time cold/hot 20/7 sec
Starting current 550% DOL 211
0.3 × = 0.25I n
Permitted starts cold/hot 3/2 250
CT ratio 250/1
A stabilising resistor is required, calculated in accordance with
Start time at 100% voltage 4 sec
Equation 19.2 to prevent maloperation due to CT spill current
Start time at 80% voltage 5.5 sec
during starting as the CTs may saturate. With the stabilising
Heating/cooling time constant 25/75 mins resistor present, instantaneous tripping is permitted. The
System earthing Solid alternative is to omit the stabilising resistor and use a definite
Control device Circuit Breaker time delay in association with the earth-fault element.
Table 19.2: Example motor data However, the time delay must be found by trial and error
during commissioning.
19.14.1.1 Thermal protection
The current setting ITH is set equal to the motor full load 19.14.1.4 Locked rotor/Excessive start time protection
current, as it is a CMR rated (Continuous Maximum Rated) The current element must be set in excess of the rated current
motor. Motor full load current can be calculated as 211A, of the motor, but well below the starting current of the motor
therefore (in secondary quantities): to ensure that a start condition is recognised (this could also
be achieved by use of an auxiliary contact on the motor CB
211
I TH = = 0.844 wired to the relay). A setting of 500A (2In) is suitable. The
250 associated time delay needs to be set to longer than the start
Use a value of 0.85, nearest available setting. time, but less than the cold stall time. Use a value of 15s.
The relay has a parameter, K, to allow for the increased
19.14.1.5 Stall protection
heating effect of negative sequence currents. In the absence of
any specific information, use K=3. The same current setting as for locked rotor protection can be
used – 500A. The time delay has to be less than the hot stall
Two thermal heating time constants are provided, τ 1 and τ 2 . time of 7s but greater than the start time by a sufficient
τ 2 is used for starting methods other than DOL, otherwise it margin to avoid a spurious trip if the start time happens to be
a little longer than anticipated. Use a value of 6.5s.
is set equal to τ 1 . τ 1 is set to the heating time constant,
hence τ 1 = τ 2 .=25min. Cooling time constant τ R is set as a The protection characteristics for sections 19.14.1.1-5 are
shown in Figure 19.8.

19-12
 Chapter 19 ⋅ AC Motor Protection

Motor tripping characteristics. 19.14.2 Protection of an LV Motor

1000
LV motors are commonly fed via fused contactors and
therefore the tripping times of a protection relay for
100 overcurrent must be carefully co-ordinated with the fuse to
ensure that the contactor does not attempt to break a current
10 in excess of its rating. Table 19.3(a) gives details of an LV
motor and associated fused contactor.
Time (sec)

1 Cold Thermal
Hot Thermal
S/C
Parameter Symbol Value Unit
Locked Rotor
Stall (a) LV motor example
0.1 Start Current (100% V)
Start Current (80% V) Standard IEC 60034
Motor Voltage 400 V
0.01
0.01 1 10
Motor kW 75 kW
Ith/I (p.u.)

Figure 19.9: Protection characteristics for motor protection example Motor kVA 91.45 kVA
Motor FLC 132 A
19.14.1.6 Negative phase sequence protection Starting Current 670 %
Two protection elements are provided, the first is definite time- Starting Time 4.5 sec
delayed to provide an alarm. The second is an IDMT element Contactor rating 300 A
used to trip the motor on high levels of negative sequence Contactor breaking capacity 650 A
current, such as would occur on a loss of phase condition at Fuse rating 250 A
starting. In accordance with Section 19.7, use a setting of 20%
(b) Relay settings
with a time delay of 30s for the definite time element and 50%
Overcurrent Disabled -
with a TMS of 1.0 for the IDMT element. The resulting
Overload setting Ib 4.4 A
characteristic is shown in Figure 19.10. The motor thermal
protection, as it utilises a negative sequence component, is Overload time delay I>t 15 sec

used for protection of the motor at low levels of negative Unbalance I2 20 %

sequence current. Unbalance time delay I2>t 25 sec
Loss of phase time delay <Ip 5 sec
10
Table 19.3: LV motor protection setting example

19.14.2.1 CT ratio
Time (sec)

1
The relay is set in secondary quantities, and therefore a
suitable CT ratio has to be calculated. From the relay manual,
a CT with 5A secondary rating and a motor rated current in the
range of 4-6A when referred to the secondary of CT is required.
0.1 Use of a 150/5A CT gives a motor rated current of 4.4A when
10 Current (A) 10000 referred to the CT secondary, so use this CT ratio.
Figure 19.7: Motor protection example- negative sequence protection
characteristic 19.14.2.2 Overcurrent (short-circuit) protection
19.14.1.7 Other protection considerations The fuse provides the motor overcurrent protection, as the
protection relay cannot be allowed to trip the contactor on
If the relay can be supplied with a suitable voltage signal, stall
overcurrent in case the current to be broken exceeds the
protection can be inhibited during re-acceleration after a voltage
contactor breaking capacity. The facility for overcurrent
dip using the undervoltage element (set to 80-85% of rated
protection within the relay is therefore disabled.
voltage). Undervoltage protection (set to approximately 80%
voltage with a time delay of up to several seconds, dependent on
19.14.2.3 Thermal (overload) protection
system characteristics) and reverse phase protection can also be
implemented to provide extra protection. Unless the drive is The motor is an existing one, and no data exists for it except
critical to the process, it is not justifiable to provide a VT specially the standard data provided in the manufacturer's catalogue.
to enable these features to be implemented. This data does not include the thermal (heating) time constant
of the motor.

19-13
Protection & Automation Application Guide

In these circumstances, it is usual to set the thermal protection definite time relay element, a setting of 20% (from Section
so that it lies just above the motor starting current. 19.7) is appropriate, with a time delay of 25s to allow for short
high-level negative sequence transients arising from other
The current setting of the relay, I b , is found using the formula
causes.
In
Ib = 5 × 19.14.2.5 Loss of phase protection
Ip
The relay has a separate element for this protection. Loss of a
Where phase gives rise to large negative sequence currents, and
therefore a much shorter time delay is required. A definite time
I n = motor rated primary current delay of 5s is considered appropriate.
I p = CT primary current The relay settings are summarised in Table 19.3(b).

Hence,

132
Ib = 5 × = 4.4 A
150
With a motor starting current of 670% of nominal, a setting of
the relay thermal time constant with motor initial thermal
state of 50% of 15s is found satisfactory, as shown in Figure
19.11.

250A

Contactor

CT

Relay
Cable

Figure 19.9: Example of an industrial motor management relay

M (Multilin 339 Motor Management System)
~
(a) LV motor protection - contactor fed example
1000

100
Time (sec)

10

Trip time
Start current
1
0 1 2 3 4 5 6 7 8
I/Ib
(b) Relay trip characteristic
Figure 19.81: Motor protection example – contactor-fed motor

19.14.2.4 Negative sequence (phase unbalance)

protection
The motor is built to IEC standards, which permit a negative
sequence (unbalance) voltage of 1% on a continuous basis.
This would lead to approximately 7% negative sequence current
in the motor (Section 19.7). As the relay is fitted only with a

19-14
Chapter 19 ⋅ AC Motor Protection

19-15
SYSTEM INTEGRITY
PROTECTION SCHEMES
Chapter 20
SYSTEM INTEGRITY PROTECTION
SCHEMES

20.1 Introduction 20.1 INTRODUCTION

20.2 Summary of System Integrity Operating electric power systems is continually becoming
Protection Schemes more complex. Increasing challenges are posed as the systems
20.3 Time-Synchronised System are operated closer to their stability limits. Electrical power
systems continue to expand, integrating new components
Integrity Protection Schemes
such as renewable energy sources, distributed generation and
20.4 Non-Synchronised System independent power producers, in an increasingly open energy
Integrity Protection Schemes market. Regulatory pressure has power system operators’
attention focused to grow return on investment of their assets
20.5 References whilst power consumption is increasing and many power
system infrastructures are ageing.
In the wake of deregulation, electricity markets load the grid
in a less predictable and more dynamic way, while generation
companies’ decisions do not always account for transmission
constraints.
Stability margins are narrowing since the system is being pushed
to operate closer to its limits, and forecasting operational data
is getting more difficult as grids connect over increasingly larger
geographical footprints.
The result is that when things do go wrong, they can have a
dramatic effect with cascade tripping and blackouts.

20.2 BLACKOUTS
In 1965, 25 million people in parts of the United States and
Canada lost their electricity supply for about 12 hours. The
failure was attributed to a maintenance error and carries the
reputable distinction of being the first large scale blackout. It
is not an isolated event. Leaving aside ‘planned’ disruptions
caused by, for example civil unrest such as the miners’ strike in
the UK in 1974, there are many such events that have caused
disruption that is, at the least inconvenient, generally commer-
cially damaging, and at worst, life-threatening.
In 1978, 80 percent of France was affected by a blackout caused
by the breakdown of a transmission line.
In 1989, a geomagnetic storm caused an outage that left 6
million people in Québec, Canada without power for nine

GEGridSolutions.com 20-1
Protection & Automation Application Guide

In 1999 a major blackout was triggered in Brazil following a include energy management system technology and system
lightning strike. Over 90 million people were affected. integrity protection schemes also known as special protection
schemes, remedial action schemes, or wide area schemes.
In August 2003, a wide-area power collapse in the north of the
United States of America and central Canada affected 50 million Electrical power systems are designed and operated to
people. It is estimated to have cost billions of dollars, and at withstand contingencies that are likely to occur as a result of a
least eleven fatalities were attributed to the loss of power. single event. Such events are sometimes referred to as class 1,
or N-1 type contingencies. For these contingencies protection is
During the miners' strike of 1974, energy and electrical power
likely to be afforded by relays applied to individual items of plant.
systems had been at the forefront of the political agenda in the
UK. The devastation caused by the North American blackout A combination of simultaneous events could be classified as a
put electrical power systems onto the world political agenda. class 2, or N-2 type contingency, and a cascading event as a
Supply of electrical energy had become, to many, no longer a class 3, or N-3 contingency.
luxury, but a basic human need.
For N-2 and N-3 contingencies, system level action is
Large-scale blackouts have continued since that time in required. This could involve generator tripping (possibly on a
countries including Australia, Japan, Peru, Greece, Russia and massive scale), load shedding or, ultimately, coordinated
Italy, but technology is being developed and deployed to help system separation, to preserve system stability and/or prevent
prevent future occurrences. serious damage to transmission and generation equipment.

Scenarios of major blackouts are hardly ever the same. The As well as the N- contingencies, another way of looking at the
starting event may be a generation-load imbalance, a short system operation is by means of the state diagram introduced
circuit, human actions (or inaction), unexpected grid topology by Fink and Carlsen [20.1] as shown in Figure 20.1.
changes, lack of voltage support, or natural causes such as
lightning storms or untrimmed trees. What they all share is a Normal
resulting failure process that cascades.

Occasionally, to prevent network components from

overloading, parts of a power system automatically disconnect
Restore Alert
or shut down to avoid damage and isolate the problem
component from the rest of the network. Under certain
conditions, the shutdown of the component can cause
significant current fluctuations or overloads in the remaining Extreme Emergency
segments of the network. These fluctuations may propagate
and amplify, leading to uncontrollable power system dynamic Figure 20.1: Fink Carlsen diagram

oscillations and cascading failure, which spreads to wider parts In this model, conventional protection and control is likely to
of the network and, in some cases, the entire grid. be effective in the ‘alert’ and ‘emergency’ states where the
load capacity and generating capacity remain matched. In the
20.1.2 Preserving Supply ‘extreme’ state, they are no longer matched and system
Although rare, these widespread blackouts and their collection integrity protection schemes are required.
of critical consequences on business, public safety (lighting), Remedial actions performed by system integrity protection
public comfort (heating/air conditioning), telecommunications, schemes are the second and third lines in the plan to protect
etc., have driven the industry to develop automatic control the system.
systems to mitigate the occurrence of such large-scale failures
caused by unexpected events. Modern power systems are 20.1.3 Enabling Technologies
designed and equipped to be resistant to this sort of cascading System integrity protection schemes have advanced thanks to
failure. Facilities and tools, which embed such resistance, technological developments, which have been widely adopted

20-2
 Chapter 20 ⋅ System Integrity Protection Schemes

by the industry. Three technological developments in be catastrophic and so system integrity protection schemes
particular that have brought great benefit and advances in the must be carefully designed. Four main design criteria
field of measurement, protection, control and automation of applicable to system integrity protection schemes are [20.2]:
electrical power systems are:
• Dependability:– The certainty that the system integrity
• advances in embedded computing protection scheme operates when required, that is, in
all cases where emergency controls are required to
• wide-area communications and
avoid a collapse.
• accurate time synchronisation
• Security:– The certainty that the system integrity
This chapter explores at how these enabling technologies can protection scheme will not operate when not required,
enhance the measurement of power system quantities, i.e. does not apply emergency controls unless they are
enabling wide-area monitoring and control, and ‘bigger necessary to avoid a collapse.
picture’ protection solutions. • Selectivity:– The ability to select the correct and
minimum action to perform the intended function.
20.2 SUMMARY OF SYSTEM INTEGRITY That is, to avoid using disruptive controls such as load
shedding if they are not necessary to avoid a collapse.
PROTECTION SCHEMES
• Robustness:– The ability of the system integrity
System integrity protection schemes (or 'special protection
protection scheme to provide dependability, security
schemes', 'remedial action schemes', and in some cases
and selectivity over the full range of dynamic and
'wide-area schemes'), are tools to protect the system. Unlike
steady state operating conditions that it will encounter.
conventional protection relays where the purpose is to isolate
faulted plant, system integrity protection schemes initiate Conformance to these criteria is critical to the deployment of
actions to correct the system. Typical system stress scenarios system integrity protection schemes.
that a system integrity protection scheme might be required to
Under normal system operation, the measures that a system
act upon include:
integrity protection scheme may initiate will be inactive. They
• transient instability are activated when some system disturbance or stress
• voltage degradation condition occurs.
• frequency degradation In terms of detection principles, system integrity protection
• thermal overloading schemes may be classified in two different categories: event-
• loss of synchronism based and response based. Event-based systems act to detect
• large power swings the cause(s) of a disturbance; Response-based systems
• cascade of overloads respond to the effects. Typically, event based systems will
automatically activate remedial actions and will operate
To implement effective, intelligent system integrity protection
quickly. Response based systems may activate the remedial
schemes appropriate for the prevailing power system condition,
actions either automatically or manually, and longer operating
it is essential that real time system data is made available.
times may apply.
Using this data, a system integrity protection scheme can
System integrity protection schemes may exist in stand-alone
initiate:
devices, using local quantities only to provide fast acting
• generator tripping decisions, or they may take quantities from a wider area of the
• load rejection system. The latter will have longer operating times as it relies
• controlled system separation on the communication of data between source of acquisition
and decision-maker. Additionally, system integrity protection
• voltage clamping
schemes may be made at the highest system level and be part
• modification to the operational state of FACTS devices
of the energy management suite as shown in Figure 20.2.
As with conventional protection the consequences of failing to
operate when required, or operating when not required could

20-3
Protection & Automation Application Guide
Centralised

Energy management
system integrity
protection scheme

Wide-area system GPS GPS

integrity protection Receiver Receiver
scheme

IED system integrity Communications link

Local

protection scheme IED at IED at

end A end B
Milliseconds Seconds Minutes
Figure 20.3: GPS timing synchronisation for current differential
Figure 20.2: System integrity protection schemes residency and typical protection.
operating times.
The basis of the scheme described by Figure 20.3 is that the GPS
Examples of IED (Intelligent Electronic Device) system integrity
signals are picked up by antennae connected to synchronising
protection schemes include detecting changes to topology, and
units. These synchronising units decode the information to
detecting loss of synchronism.
produce a highly accurate timing signal that synchronises the
An example of a wide-area system integrity protection scheme current signals used by the relays to perform the current
is monitoring angular stability across the system and taking differential algorithm. The timing signal presented to the relays is
actions necessary to correct. Energy management system in the form of a pulse that is transmitted at the start of every
integrity protection schemes can be used to take a more second. This one pulse per second (1pps) is more than sufficiently
‘static’ view, looking at long term phenomena. accurate for multi-ended current differential protection and it is a
logical progression to use the same accuracy of GPS timing
Implementing system integrity protection schemes requires
information to synchronise the measurements of power system
real time data. When that data is coming from different
quantities across the whole grid.
sources it is crucial that the data is synchronised, and that it
can be communicated effectively and efficiently. It is the The GPS 1pps signal can be used to synchronise the
possibility to accurately synchronise devices and to rapidly measurement of anything from anywhere in the world. In the
communicate the data that makes time-synchronised system electrical power system it can be used to synchronise the
integrity protection schemes possible. measurement of voltage and current from key nodes on the
grid to enhance the monitoring and control of the grid.
20.3 TIME-SYNCHRONISED SYSTEM
So called phasor measurement units (PMU) are synchronized
INTEGRITY PROTECTION SCHEMES
using the GPS timing to produce time-synchronised phasor
As discussed briefly in chapter 10, timing information from the
representations of the current and voltage signals that they
global positioning satellite system (GPS) can be used to
take as input. These time-synchronised phasor representations,
synchronise current differential unit protection and merging
or synchrophasors, can be collected from different nodes on
units (chapter 24). The GPS system consists of a number of
the grid and compared to get a wide-area system-level view.
satellites in space that are continually transmitting
information. Current and voltage synchrophasors from phasor
measurement units can be used to observe the state of the
The basic principle of using the GPS system to synchronise
system and improve the performance of different system level
current differential protection is shown in Figure 20.3. This is
applications. Using synchrophasors to observe, measure, or
an example of a communication-assisted scheme with a
monitor the system is sometimes termed ‘wide-area
practical application.
monitoring’ or WAM. Control actions taken on the basis of the
interpretation of these WAMs is sometimes termed ‘wide-area
control’, or WAC, and protection action taken on the WAMs is

20-4
 Chapter 20 ⋅ System Integrity Protection Schemes

sometimes termed ‘wide-area protection’, or WAP. System 20.3.1.1 Synchrophasors for Oscillatory Stability
integrity protection schemes based on wide-area technology Synchrophasors can be used for oscillatory stability detection
are essentially control schemes. It is perhaps unwise to try to designed to detect slower inter-area system oscillations. These
differentiate the application of wide-area technology to the are characterised by power oscillations between two areas of
separate fields of monitoring, control, and protection, rather to generation, running at slightly different speeds and following
combine them into a single heading of ‘wide-area monitoring, small system perturbations such as load switching or tap
protection and control’, or WAMPAC. changes. If there is insufficient system damping, these
oscillations do not die away and can escalate leading to
20.3.1 Wide-area Monitoring, Protection and Control
potential out-of-step conditions if no pre-emptive action is
(WAMPAC) System Integrity Protection Schemes
taken. If the phasor measurement units feature frequency
A wide-area scheme to accommodate system integrity tracking that can accurately measure and track small changes
protection could consist of phasor measurement units to in signal frequencies, this can provide an accurate
generate the synchrophasor data signals. These signals need measurement for the fast detection of small signal oscillations
to be collected together before system level applications are to initiate damping actions.
exercised. Collection of synchrophasors is typically performed
by a phasor data concentrator (PDC) which prepares the
synchrophasor data for the application.

Wide-area technology based on synchrophasors can be used

to enhance state estimation and provides opportunities for
WAMPAC schemes as shown in Figure 20.4.

Portion of grid
de-energised? Adaptive
PMUs restoration
~minutes

3rd
Slow voltage
Figure 20.5: PMU data viewed in a simple visualisation tool
Voltage
decline?
PMUs
instability
prediction Synchronised measurements allow the comparison of the phase
~milliseconds
~10s of
seconds of the oscillatory modes to determine the parts of the system
Wide-Area State
1st Estimation 3rd that are oscillating with respect to each other, and enable
~ seconds

Angular
stabilising actions to be initiated as shown in Figure 20.6.
Substation instability
SE Abrupt frequency prediction
decline? ~milliseconds
2nd
Oscillatory
Inter-area and local instability Tune PSS, SVC, HVDC,
3rd
oscillations prediction taps, etc.
~minutes
Poorly damped Oscillatory
Substation oscillations? instability Key
SE
prediction
~milliseconds PSS: Power System Stabiliser
~minutes
SVC: Static Var Compensator
Immediate response to unexpected HVDC: High Voltage DC
1st events in the grid
Figure 20.6: Oscillatory instability prediction
Figure 20.4: WAMPAC opportunities

Figure 20.4 shows applications based on the use of 20.3.1.2 Synchrophasors for Voltage Stability
synchrophasors and state estimation (SE) for oscillatory Synchrophasors can be used to predict voltage instability. By
stability, voltage stability, transient stability (or angular comparing voltage and power transfers on the line against
instability prediction), and system restoration. specific stability margins, operators can be alerted to slow
The technology can also offer event triggering based on voltage declines and take remedial actions such as switching
overcurrent, under/overvoltage, under/over-frequency and in reactors as shown in Figure 20.7.
under/over-rate of change of frequency.

20-5
Protection & Automation Application Guide

accurate time synchronisation signal, time synchronised

Capacitor switching
Voltage
Slow voltage decline
instability
Reactor switching
SVC setting
phasors, known as synchrophasors can be produced. The
prediction
or load level?
~10s of
Undervoltage shedding
Remedial actions potential benefits of synchrophasor technology have long been
seconds
Fast unit runback
recognised, but it is only since enabling technologies have
Key
SVC: Static Var Compensator developed that their effective application became reality.
Figure 20.7: Voltage instability prediction
In 1893, Charles Proteus Steinmetz presented a paper on a
20.3.1.3 Synchrophasors for Transient Stability simplified mathematical description of the waveforms of
Synchrophasors can be used to defend transient stability by alternating electricity. Steinmetz called his representation a
predicting angular instability. By being measured in real-time phasor, and phasor notation became widely used in the field of
rather than estimated, generator and bus-phase angles can be electrical power systems (see section 3.4.1). The time
compared and stability threats can be presented to an operator synchronisation of phasors in what were to become known as
for remedial action such as controlled islanding as shown in phasor measurement units evolved and in 1988 the first
Figure 20.8. phasor measurement units based on Steinmetz’s technique,
were developed. The first commercial phasor measurement
Controlled islanding
Generator tripping
unit was launched in 1992.
Dynamic braking
Angular
instability
Underfrequency.
shedding
The deployment of phasor measurement units into electrical
Frequency decline
prediction Fast load shedding
~milliseconds FACTS, HVDC controls power systems, however, was limited initially as two main
Remedial actions
Special protection issues could not be addressed:
Key Fast unit runback
FACTS: Flexible AC Transmission System
HVDC: High Voltage DC • Accuracy of time synchronisation
Figure 20.8: Angle instability prediction • Cost effective, high-speed, communications
Figure 20.9 shows the relative phase angles from two parts of
With the global positioning satellite system, the ability to time
the North American grid taken on August 14th 2003.
synchronise anywhere in the world to an accuracy of better
than a micro-second is achievable. Developments in the
telecommunication markets have brought fast Ethernet to the
substation, and with these developments, the hurdles are
overcome, and the potential of synchrophasors can be
exploited to the maximum.

20.3.3 Synchrophasor Definition

A phasor is a vector representation on the complex plane of a
Figure 20.9: Phase angle divergence prior to blackout
sinusoidal quantity such as an alternating current.

The phase angles from the two parts of the system can be seen A synchrophasor is a phasor calculated from sampled values
diverging. A major blackout followed. This is the type of event that are referenced to a common timing signal. The timing
that the synchrophasor technology can help prevent. But what signal is global, meaning that a synchrophasor measured
exactly is a synchrophasor? anywhere in the world will be referenced to the same time
signal. This enables synchrophasors to be communicated for
20.3.2 Synchronised Phasor Measurements collation and comparison. This is the fundamental concept
A phasor is a representation of a sinusoidal quantity, behind wide-area special protection schemes.
representing the sinusoid as a vector rotating on the complex The IEEE 1344 standard for synchrophasors was published in
plane, defined in terms of amplitude and phase. As steady 1995. It was reaffirmed in 2001, but in 2005 it was
state power system quantities are sinusoids, it follows that superseded by IEEE C37.118. This recognised the
they can be represented by phasors. When augmented by an developments that would facilitate the effective application of

20-6
 Chapter 20 ⋅ System Integrity Protection Schemes

synchrophasors to electrical power systems. IEEE C37.118 Figure 20.10: Time and frequency domain representation of
synchrophasors
describes the standards for measurement of synchrophasors,
the method of quantifying the measurements, as well as the The measured angle θ is valid in the range of –π to +π
testing and certification requirements for verifying accuracy. It radians.
also prescribes the data transmission format and protocol for A primary requirement for synchrophasors is the accuracy of
synchrophasor communication. The more recent 2011 edition the measured phasor. The IEEE C37.118 (2005) standard
of IEE C37.118 classifies phasor measurement units into M defines the total permitted vector error (TVE) for the static
Class (for very accurate measurements in steady-state) and P condition at nominal frequency as:
class (for tracking dynamic system conditions).

The standard prescribes that all measurements are referenced TVE =

( X r ( n ) − X r )2 + ( X i ( n ) − X i )2
2 2
to an accurate one pulse per second (1pps) signal. Generally it Xr + Xi
is anticipated that this will be derived from the global Where Xr(n) and Xi(n) are the measured real and imaginary
positioning satellite system, but it can be from any external components and Xr and Xi are the reference values.
source provided the accuracy is in accordance with the
Levels of accuracy are specified for the permissible total vector
requirements.
error as shown in Table 20.1 .
Phasor representation defines the complex exponential as a
point or vector on the complex plane according to the Range of influence quantity change with respect
to reference and maximum allowable TVE (in %)
equation: for each compliance level
Influence Reference
jθ quantity Condition Level 0 Level 1
e = cos θ + j sin θ
TVE TVE
The IEEE C37.118 standard specifies that the angle θ is zero Range Range
(%) (%)
degrees when the maximum of the signal to be measured Signal frequency Fnominal +/- 0.5Hz 1 +/- 5Hz 1
coincides with the leading edge of GPS 1pps pulse. It follows Signal
100% rated 80% -120% rated 1 10% -120% rated 1
that it will be –π/2 radians (-90 degrees) if the positive zero magnitude
Phase angle 0 radians +/-π radians 1 +/-π radians 1
crossing coincides with the 1pps pulse, +π/2 radians (+90
Harmonic <0.2% 1%, any harmonic 1%, any harmonic
degrees) if the negative zero crossing coincides with the 1pps 1 1
distortion (THD) up to 50th up to 50th
pulse, and π radians (180 degrees) when the reverse polarity
Out-of-band <0.2% of input
peak of the signal to be measured coincides with the 1pps 1% of input signal 1% of input signal
interfering signal signal 1 1
magnitude magnitude
at freq fi * magnitude
pulse.
* f − f > Fs , where Fs = phasor reporting rate and f0 = Fnominal
The relationship of time domain and frequency domain signals i 0 2

relative to the 1pps signal is further described in Figure 20.10, Table 20.1: Influencing quantities and allowable error limits for
compliance levels 0 – 1.
where X r (n) and X i (n) are the RMS real and imaginary
Devices offering compliance level 1 perform better in SIPS
filtered components at a particular instance and θ is the phase schemes, whereas level 0 devices are intended more for
angle as per the IEEE C37.118 standard definition. measurement purposes.
θ
Peak
This measurement accuracy will vary with the magnitude and
ms
frequency of the input signal and may be depicted graphically
Xi(n)
θ
as shown in Figure 20.11. The graph depicts a device with a
Xr(n) total vector error typically less than 0.4% (y-axis), consistent for
system frequency excursions off-nominal (z-axis), and
maintained for all except very low current magnitudes (x-axis).
1 PPS

20-7
Protection & Automation Application Guide

Total Vector Error across, say, an Ethernet network. An example of a typical data
frame is shown in Table 20.2.
TVE (%)

Field function Field size (bytes)

Message synchronisation 2
Frame size definition 2
PMU ID code* 2
Real time timing information 8
Status information 2
Phasor data values Typically 48
System frequency measurement 2
Rate of change of system frequency 2
Digital data values 2
Frequency Deviation Magnitude (%)
Frame checksum 2
Total Typically 72
Figure 20.11: Graphical representation of synchrophasor total vector
error Table 20.2: Synchrophasor data frame format

* The PMU ID code is set to uniquely associate the

20.3.4 Synchrophasor Communication
synchrophasor values to the PMU responsible for producing
As well as the generation of synchrophasors, the IEEE C37.118 them.
standard also defines how the synchrophasor values are to be
The rate at which the frames are sent is defined.
communicated.
• For 50Hz operation the frame repetition rate can be set
Whilst the standard does not mandate that Ethernet
to 10, 25 or 50 frames per second
communications are used for synchrophasor transmission, it is
• For 60Hz operation the frame repetition rate can be set
Ethernet that often enables wide-area phasor measurement
to 10, 12, 15, 20, 30, or 60 frames per second.
technology, and that is reflected in this section. Some
knowledge of Ethernet communications is assumed here, but if The faster the rate of frame repetition, the shorter the latency
needed, a more detailed discussion on substation in responding to a developing system condition, allowing faster
communications, including Ethernet, is presented in chapter SIPS and state estimation reaction.
24: ‘The Digital Substation’. Clearly the more times per second that the data is sent, the
Four message types associated with synchrophasor greater the impact on the Ethernet communications.
communications are described in the standard:
20.3.4.1 Mapping Synchrophasors to Ethernet
• Data
When mapped to Ethernet the synchrophasor phasor data
• Configuration
frames can use either transmission control protocol (TCP) or
• Header user datagram protocol (UDP) according to preference. TCP is
• Command a connection-oriented protocol. It manages message
All message types have a similar frame structure to package acknowledgement, re-transmission and time out. As such it
the information. The “Configuration”, “Header” and can be considered to be reliable and ordered, but carrying
“Command” types are used to set up the communications and overheads. UDP is a simpler protocol that broadcasts
are used infrequently. The “Data” type is used for the real- messages from the transmitter without checking the state of
time transfer of synchrophasor values, and that will have the the receiver. As such it can be considered less reliable and not
largest impact on the communications system. ordered, but lean. The choice will come down to the specific
requirements of the application, and the protocol options
According to the standard, synchrophasor signals are
provided by the selected PMU.
encapsulated into data packets, or frames, for transmission

20-8
 Chapter 20 ⋅ System Integrity Protection Schemes

The TCP or UDP mapped messages are written to and read frame rate is thus:
from using standard internet protocol (IP) input-output
• 1072 * 50 = 53.6kbps for a maximum frame rate of 50
functions. These functions apply a numerical identifier for the frames per second (TCP)
data structures of the terminals of the communications. The
• 1072 * 60 = 64.32kbps for a maximum frame rate of
terminal is referred to as a port and the numerical identifier is 60 frames per second (TCP)
called the port number. The port number should be set to
• 800 * 50 = 40kbps for a maximum frame rate of 50
align with the recipient phasor data concentrator (see section frames per second (UDP)
20.3.7). • 800 * 60 = 48kbps for a maximum frame rate of 60
Other settings that apply as the phasor message is mapped to frames per second (UDP)
the Ethernet layer are the IP address, subnet mask, and These figures are not absolute, but they can be of benefit in
gateway address. planning network requirements when installing and operating
phasor measurement units.
20.3.4.2 Communication Bandwidth Requirements
It can be useful to understand how much loading the addition 20.3.5 Phasor Measurement Units
of a phasor measurement unit can have on a communications Synchrophasors are generated in phasor measurement units.
system. A typical frame length of 72 bytes has already been A typical phasor measurement unit architecture is shown in
mentioned. For a typical application with a frame length of 72 Figure 20.12.
bytes and a repetition rate of 60 per second, the minimum
bandwidth requirement for these raw frames is, therefore,
4320 bytes per second.

Ethernet conforms to the OSI (Open System Interconnect) 7

layer model (covered in chapter 24). Some layers require extra
1 PPS
Measurement
information to be added and so, in addition to the raw data Digitised
Protection and
GPS
A/D processor
measurement
samples Receiver
frames, there is an overhead associated with the OSI 7-layer Analogue processors
electrical IRIG-B
model.
signals
Ethernet
IEEE C37.118 data
For TCP/IP mapping the overhead is: PMU processor
Data frames over
• TCP 24 bytes per frame serial or Ethernet
(TCP/UDP)
• IP 20 bytes per frame Figure 20.12: Typical Phasor measurement unit architecture
• MAC 18 bytes per frame Electrical signals to be converted into synchrophasors are
So the TCP/IP overhead is 62 bytes of data per frame. presented at the analogue inputs. They are sampled and
converted into numerical values in an analogue to digital
For UDP/IP mapping the overhead is:
converter (A/D). The sampling is controlled by the
• Source port 2 bytes measurement processor which is synchronised to the GPS
• Destination Port 2 bytes system by a 1pps signal from a time synchronisation unit. The
• UDP length 2 bytes numerical values representing the digitised samples are passed
• UDP checksum 2 bytes to the measurement processor. The measurement processor
• IP 20 bytes per frame determines the amplitude and phase quantities required for the
synchrophasors according to their displacement from the 1pps
So the UDP/IP overhead is 28 bytes of data per frame. signal. The measurement processor also calculates the derived
As the C37.118 frame contains 72 bytes, the total number of quantities that are required to be presented as synchrophasors.
bytes per frame is 134 for TCP and 100 for UDP. These include sequence components, frequency, and rate-of-
change of frequency, etc.
With 8 bits per byte, the resulting bandwidth of the maximum

20-9
Protection & Automation Application Guide

An IRIG-B signal from the time synchronisation unit allows the more conservative to apply dedicated time synchronisation
measurement processor to accurately time-stamp the devices into the substation using connections that are separate
synchrophasor values before passing them to the Ethernet to the LAN. A typical time synchronisation device might
processor. The Ethernet processor frames the synchrophasor provide a one-pulse-per-second (1pps) signal, derived from
values according to IEEE C37.118 before transmitting over the GPS signals, in order to provide precise alignment of the
either TCP/IP or UDP/IP. synchrophasors. To combat the effects of electrical interference
in the substation environment, the 1pps signal is generally
20.3.5.1 Physical Deployment of Phasor Measurement presented as a fibre-optic signal. The accuracy is typically
Units better than 0.1μs. Although the 1pps signal is very accurate
There is no need to install phasor measurement units at every (Figure 20.13), the pulse on it only signifies that another
busbar in the system as this will not necessarily contribute to second has passed (or is about to start). In order that the
more efficient system operation. Indeed, too many phasor phasor measurement unit can provide an accurate time stamp
measurement units could cause congestion on the for the synchrophasors, it is necessary to provide an additional
communications system, making data less manageable. ‘real’ time signal. This is achieved with an IRIG-B output,
which delivers coordinated universal time (UTC) for very
Phasor measurement units should be connected so that inter-
accurate time stamping. It is quite likely that the time
area oscillations and other critical dynamics will be correctly
synchronisation device will feature multiple outputs and be
monitored. They should be deployed on the network at critical
capable of synchronising multiple PMUs and merging units,
buses to ensure that a sufficient picture of the system
providing substation-wide timing information.
dynamics is available to the control centre to enable decisions
to be made. System visualisation using synchrophasors from
Seconds
various parts of the system can show, for example, the 1s
angular, frequency and voltage differences between groups of
<50ns
generators. To do this, synchrophasors from various locations
need to be brought together, and one method of achieving this Light <50ns
ON
is by means of phasor data concentrators, which is covered in 200ms ON

section 20.3.7. Output

Permissible
area for leading
edge of optical
20.3.6 Time Synchronisation Light signal
OFF
Phasor measurement units require accurate time Time
Figure 20.13: 1pps accuracy
synchronisation. There are two preferred technology options for
time synchronisation of PMU's - IEEE 1588 or 1 PPS. The GPS system consists of more than 20 satellites rotating in
orbit above the earth. The satellites are not in geo-stationary
IEEE-1588 describes a high-precision time protocol that can
orbit, and so the constellation that can be ‘seen’ is constantly
achieve sub-microsecond accuracy for time synchronisation to
changing. In general, up to 7 satellites can be ‘seen’ at any
be used in measurement systems. It is an Ethernet protocol
time. When GPS time synchronising receivers are energised,
that can be connected to a local area network (LAN). Timing is
they need to initialise their timing algorithms. Generally, this
derived from the global positioning satellite (GPS) system.
requires that at least 4 satellites are in view for a short while.
Since the synchronism is delivered by the Ethernet LAN, only
one GPS receiver is required to synchronise all devices on the In order to receive the signals from the GPS satellites, an
LAN. This offers potential cost benefits when compared to antenna is required. This needs to be mounted externally. A
using dedicated time synchronisation devices, especially as the mounting pole is generally used as shown Figure 20.14.
PMUs are typically connected to the Ethernet for the
transmission of synchrophasors. Given the criticality of
delivering the time synchronisation signal, however, it may be

20-10
 Chapter 20 ⋅ System Integrity Protection Schemes

measurement units (PMU) dispersed throughout the electricity

system and connected to an energy management system via
phasor data concentrators (PDC). Phasor data concentrators
are devices which collect data from a number of phasor
measurement units. They convert such data into a single data
stream for onward transmission. Phasor measurement units
are connected to a phasor data concentrator at either the
substation or the control centre level.

Phasor data concentrators connect to supervisory control and

data acquisition (SCADA) systems and/or energy management
systems (EMS) at the central control facility to provide the
NOTE synchrophasor information for the wide-area applications.
Earth cable no longer than 10m
Earth cable 10mm2 cross section

Figure 20.14: Antenna connection details

Substation A Substation B
It is vital that the receiving antenna has the best possible
GPS GPS
mounting location to maintain reception with as many Receiver Receiver

satellites as possible, thus the ideal mounting is shown in PMU PMU

Figure 20.15.
PMU PMU
150° vertical view 360° all-round view

From other
PDC substations
Antenna

Visualisation
15° mask angle Software

Figure 20.15: GPS antenna required line of sight

The antenna should be mounted so that it is above the roofline Figure 20.16: PMU and PDC architecture

of the building to which it is attached, ensuring that there is a The complete network requires rapid transfer of data within
360º horizontal view and a 150º vertical view of the sky (a the frequency of sampling of the synchrophasor data.
masking angle of 15º with respect to the horizon is the
The phasor data concentrator correlates the data, and controls
maximum that should be permitted). Ideally there should be
and monitors the phasor measurement units. It also serves to
no obstructions in view such as metal structures or buildings.
archive the data, which is essential in the event that post-
If the masking angle is greater than 15º with respect to the
disturbance analysis might be undertaken in order to
horizon (i.e. if is a large obstruction) the antenna must be re-
determine the root causes of cascade events. At the central
sited or mounted on a longer antenna pole.
control facility, the SCADA/EMS system presents system-wide
Due to the requirement for ensuring such a clear view of the data on all generators and substations in the system with
GPS constellation, the antenna is likely to be the highest frequent regular updates. These updates provide for a
component in the vicinity. As such it is exposed to lightning, visualisation of the system which facilitates response-based
and suitable protection is required. A lightning arrestor is schemes.
recommended for protection.
20.3.8 System Visualisation
20.3.7 Phasor Data Concentrators System visualisation enhances wide-area schemes. Figure
As shown in Figure 20.16, a phasor network consists of phasor 20.17, Figure 20.18, and Figure 20.19, show visualisations of

20-11
Protection & Automation Application Guide

electrical networks as presented on an advanced energy and the options available to them for remedial action.
management system.
System visualisation of real-time data realised thanks to wide-
area technology provides power system operators greater
awareness, and hence more precise control of the system.
Underpinning this wide-area technology is the capability to
accurately time synchronise the phasor measurement units.

20.4 NON-SYNCHRONISED SYSTEM

INTEGRITY PROTECTION SCHEMES
The system integrity protection schemes described in the
previous section require highly accurate time synchronisation
and fast communications infrastructure. One might be
Figure 20.17: Stability state: annunciating voltage contours, weak
elements and possible remedial actions tempted to consider that system integrity protection can only
be afforded where these facilities are available. This is not,
however, the case.

The numerical busbar protection outlined in chapter 15 can be

considered as a distributed protection scheme as information is
gathered from peripheral units at various points on the system
for transmission to a central unit where protection monitoring
and control is managed. This shows how distributed
intelligence can work, and by increasing the geographical
spread of other suitable devices, wide-area schemes are the
Figure 20.18: Visualisation of inter-area oscillations: frequency, severity
and damping next logical step.

Recognising however that sometimes information is available

only from a single point in the system, system integrity
protection schemes have been developed that require neither
communications, nor accurate time synchronisation.

By taking signal inputs from single points in the power system,

and by using special filters and techniques such as fuzzy logic,
predictive out-of-step or loss of synchronism conditions can be
detected and consequently acted upon to maintain system
stability. Similar techniques can be used to detect and act on
changes to system topology such as loss of power corridor.

By adding in busbar protection and relabeling the horizontal

axis in Figure 20.2 , the communication requirement of system
integrity protection can be highlighted, as shown in Figure
Figure 20.19: Angular separation across transmission lines: alerts to
dispatchers 20.20.
The figures show how phasor measurement units are providing
EMS software with synchrophasor values and how virtual phase-
meters connected to the lines give a clear picture of the real-time
phase differences between buses. If the angular displacement
exceeds a preset value, dispatchers are informed of the violation

20-12
 Chapter 20 ⋅ System Integrity Protection Schemes

Energy management reactors, but these bring added complexity. Protection against
system integrity system disturbances may be provided to some extent by
protection scheme
frequency relays or voltage relays set to initiate actions such as
Centralised

Wide-area system
load-shedding on under-frequency conditions, but these may
integrity protection
scheme not be sufficient in themselves for all applications and
sophisticated system integrity protection schemes have been
Busbar
protection developed to complement them. Two particular examples that
do not require communications are described in sections
IED system integrity entitled 20.4.2 System Topology Changes and 20.4.3 System
Local

protection scheme
Synchronism Stability. The former describes an example of an
Communication requirements event-based system integrity protection scheme looking for
Figure 20.20: Communications requirements for system integrity changes in system topology. The latter describes an example of
protection schemes. a response-based system integrity protection scheme looking
This shows ‘local’ IED system integrity protection schemes for system instability.
that do not require any communications, and these form the
Both of these system integrity protection schemes are able to
subject of the rest of this chapter.
anticipate system-threatening conditions using local variables
only. They are based on Kalman filters and fuzzy-logic multi-
20.4.1 Single-point special protection
criteria algorithms. These system integrity protection schemes
In countries where large sources of power generation are
are applicable to large power systems prone to transient and
geographically far from the principal loads, so-called
dynamic instabilities.
transmission corridors of power provide the routes for the
transmission of that power. One example of this is the Hydro- 20.4.2 System Topology Changes
Québec network in Canada as shown in Figure 20.21.
Detecting topology changes such as open line conditions
allows timely initiation of system defence plans for severe
contingencies such as a loss of transmission corridor. In the
absence of communications, signalling an open line condition
from a remote terminal is not possible and so, detecting a loss
of corridor caused by a remote terminal action using only
locally available signals can enable system stabilising actions
to be initiated.

20.4.2.1 Open Line Detection

Open line detection, like any other protection, needs to be
secure, reliable, and sufficiently fast.

To be secure, the detector should not operate in cases of low or

zero power transfer on the monitored line, distortion on line
voltages and currents during faults, fast variations in the
power flow, reversal of the power flow, power swings, sub-
synchronous resonances resulting from the parallel
combination of shunt reactors and series capacitors, nor
Figure 20.21: Hydro-Quebec 735kV series-compensated transmission
system (reproduced courtesy of Hydro-Quebec) should it operate in the presence of harmonics.
Reliability of transmission of power on such systems can be To be reliable, the detector should detect line opening at the
achieved by the addition of series compensation and shunt remote end, line closing at the local end when already open at

20-13
Protection & Automation Application Guide

the remote end, and line closing or opening with a shunt a better basis for open line detection. Observing the rate of
reactor connected. change of active and reactive power for various system events
produces interesting results. The rate of change of power
The typical speed required for open line detection operation is
(dP/dt, or DeltaP) can be studied for various system events
two cycles or less.
and two in particular are illustrated in Figure 20.23 and Figure
As with many other aspects of electrical power systems, open 20.24.
line operation can be characterised by both transient and
steady state conditions. Observation of the transient state
aspects and steady state aspects of open line conditions 60pu/min

reveals that transient events generally are associated with line

DeltaP
opening, whilst steady state phenomena are more closely
linked to closed lines. 0.5pu/min

An open line condition is very similar to a closed line which has

no active power transfer. If the relationship of critical
measured variables such as the voltage and the reactive power
2pu/min
are observed when the active power of a connected line
P Ramp

disappears, certain characteristics become apparent. Figure

20.22 shows that, for a given line length, an operating point 1pu/min

(consisting of the angle shift and the voltage ratio between its
ends) can be found so that, as seen from the sending end,
Time(s)
both active and reactive powers disappear.
Figure 20.23: Rate of change of power (DeltaP) filter response to power
flow reversal

1pu

0.5pu
DeltaP

0.01pu

Figure 20.22: Voltage ratio and angle trajectories vary with line length
for sending end active (Ps) and reactive (Qs) powers both at zero.

Although this critical operating point can be calculated, it is Time(s)

very sensitive: For short lines, say 30km, the typical Figure 20.24: Rate of change of power (DeltaP) at breaker opening
sensitivities are 45% MW per degree of angle shift and 25%
Figure 20.23 shows power transfer ramping from a positive
MVAr per unit of voltage ratio.
direction to a negative direction, thus passing through zero
It is therefore difficult to accurately detect this condition based without apparently any significant effect. As the power ramp
on power or current measurement alone, and thus it cannot crosses zero, however, it can be seen that DeltaP reaches a
form a reliable basis for detecting open lines. maximum value a few cycles later. Although the maximum
value of DeltaP varies according to the power slope, the
Studying the transient state aspects of opening a line provides
maximum is reached at nearly the same time.

20-14
 Chapter 20 ⋅ System Integrity Protection Schemes

Figure 20.24 shows the effect on DeltaP of suddenly opening a Line

Voltages
line at various values of surge impedance loading. In all cases,

Open Line Detection Pre-

Fuzzy Logic Engine

when the line is opened, DeltaP increases sharply from its zero Line

Anti-alias filter

Kalman filters

Output logic
Open Line

processor
Currents
steady-state value to its maximum value. Detection
Reactor Flag
It is apparent that observing DeltaP provides the capability for Currents

fast reliable detection of line opening.

Figure 20.26: Open-line detection functional block diagram
20.4.2.2 Implementing Open Line Detection
Anti-aliasing filters limit the bandwidth of the input signals
SUBSTATION before high speed data acquisition converts the input signals
BUSBARS
into numeric values for provision to the principal filtering unit.

The signals are filtered to extract the fundamental phasor

D D
LINE BREAKERS components.
CT1 CT2
These vectors are used as the inputs into a Kalman Filter. A
v

ICT1 ICT2
Kalman Filter is used since it can extract the fundamental
CT3
SHUNT REACTOR components and harmonic components. But unlike filters such
LINE
D
v as the Discrete Fourier Transform (DFT), a Kalman Filter can
IREACTOR
DISCONNECTOR
provide signal components at spectral frequencies other than
CVT LINE
the fundamental and its harmonics. Kalman Filters are well
suited to the application of protecting compensated lines, since
they can be used to detect, for example, the negative sequence
component at fundamental as well as resonances that may
Figure 20.25: Connections for open line detection. transiently appear when a line is opened at both ends while its
Figure 20.25 shows typical inputs used by an open line shunt reactors are connected. The outline of a Kalman filter is
detection algorithm. The inputs are taken from a capacitive shown in Figure 20.27, and the state diagram is shown in
voltage transformer (CVT) and three sets of current Figure 20.28.
transformers (CT1, CT2 associated with the circuit breakers X̂ klk
Vk +- Gk ++
and CT3 associated with the shunt reactor).

The functional diagram of the open line detection is shown in

Ck Ak-1 Delay
Figure 20.26. It consists of five principal elements. X̂ klk -1

Figure 20.27: Kalman Filter block diagram

• Anti-alias filtering
Startup
• Kalman filtering
Initial state Corrector
• Pre-processing Kalman gain (for noise and
probability), state estimate
• Fuzzy logic Predictor and error covariance
State & error correction
• Output logic covariance prediction

Figure 20.28: Kalman Filter state diagram.

The pre-processing unit takes the fundamental frequency

phasor components of the inputs, the derived sequence and
spectral components, and uses them to determine a number of
decisions about the system based on quantities that reflect the
‘openness’ or ‘closedness’ of the system such as DeltaP, active
power, reactive power, etc.. These decisions are used by the

20-15
Protection & Automation Application Guide

Fuzzy logic to produce a conclusion as to the state of the line described by common language terms such as "Normal",
(open or closed). "Small", "Large", "Very Large", etc.

20.4.2.3 Fuzzy Logic “SMALL” Active power “LARGE” DeltaP

Line closed with Line open with
Conventional, or Boolean, logic is commonly used in power full certainty full certainty
system protection. It is based on signals being in one state or 1 1

the other (i.e. ON or OFF). Fuzzy logic is based on signals

Probability μ

Probability μ
0.5 0.5
being in more than one state and where there may be overlap
Active Power
between states. Consider Figure 20.29, which shows a Delta P
0 0
θL θC θS θL θC θS
number of states for the active power.
Figure 20.30: Fuzzification of typical decision features
Membership Figure 20.30 illustrates this fuzzification process for two typical
variables: the Active Power and the DeltaP. The first feature is
defined as "SMALL" when its crisp value is below a given
1
Nil Low Normal threshold (θL) while the DeltaP feature is defined as "LARGE"
for crisp values higher than (θS).

Active Power A set of rules is applied to these fuzzy variables according to

0 basic principles. "Reliability Rules" are used specifically to
1.0 pu (Crisp Input)
assess the open-line condition and "Security Rules" target
Figure 20.29: Fuzzy Logic Sets
closed-line situations.
The value 1.0pu is resident in both the “Low” state as well as
the “Normal” state. It is not resident in the “Nil” state. An example for each type of rule is given below:

Its membership of the “Nil” state is 0, its membership of the • Security: if (ActivePower is not SMALL) LINE is
CLOSED
“Normal” state is 1, and its membership of the “Low” state is
somewhere between 0 and 1. This process is known as • Reliability: if (DeltaP is LARGE) LINE is OPEN
“fuzzification”. Although called “fuzzy”, it is, in fact, a The fuzzy logic engine contains a number of these rules for
repeatable process to which rules can be applied. detecting the open-line status under various circumstances
A fuzzy logic approach can resolve conflicts that may arise in (reliability rules), and a number of rules for detecting the
doubtful cases such as during line energisation at no-load, or closed-line status (security rules). In each category, the rules
where capacitive voltage transformers induce electro-magnetic can be further classified in terms of transient and steady-state
transients suggesting that the active power is not zero (line rules. For instance, the reliability rule above is a transient one
closed) while the line is actually open. Fuzzy logic is ideally while the security rule is a steady state.
suited to resolving such conflicts since the decision is based on The last stage of the process is called defuzzification. This
several criteria with adaptable weighting factors. consists of converting the evaluation results of the criteria,
The fuzzy decision system consists of three main steps: which are applied by the rules, into a single decision. Each
criterion states that the line is open or closed with a given
• Fuzzification of the selected decisions features from the probability and defuzzification reconciles all partial decisions
pre-processor
using an aggregation procedure into a single decision.
• Fuzzy logic inference on these features, using rules or
criteria The output from the open line detector is the flag that is
• Crisp decision sent to the output asserted when an open line is detected.

Fuzzification is a necessary step prior to a fuzzy-logic based 20.4.2.4 Open Line Detection Summary
reasoning system. It consists in transforming the crisp
In the context of interconnected and heavily loaded power
variables provided by the pre-processor into categories easily

20-16
 Chapter 20 ⋅ System Integrity Protection Schemes

systems, fast detection of topology change is considered an 20.4.3.1 Conventional Out-of-Step Protection
efficient means to instigate remedial actions for the defence of
a power system against severe events. Open Line Detection, Blinder unit

based on intelligent devices operating using locally-measured

power system quantities can provide such detection for a
typical topology change that could impact system stability.

20.4.3 System Synchronism Stability

In general, occurrence of extreme contingencies in a power
system can lead to voltage, frequency or angular instabilities.
Angular or frequency instabilities are related more to the
incapacity of the system to provide appropriate real power
flows between generation areas, or between generation and Figure 20.31: Impedance loci during power swings
load areas to maintain system equilibrium. Such instabilities
Detection of power swings has been briefly introduced in
may lead to either under/overfrequency situations or loss-of-
chapters 11 and 17.
synchronism conditions.
Classical approaches for transmission line applications make
In the first case, the frequency will be the same for the whole
use of auxiliary elements in the R, X plane in combination with
system but out of the normal range, whilst in the second case,
timers. Figure 20.31 illustrates this principle with two mho
area or sub-system frequencies will differ and induce loss-of-
relays (outer and inner zones) and a blinder unit. Zinitial is the
synchronism conditions characterised by periodic zero voltage,
pre-disturbance apparent impedance seen by a relay located at
or virtual faults. Anticipation of such loss of synchronism and
one end of a transmission line. Whilst a fault would cause an
subsequent fast remedial action, may allow system stability to
instantaneous movement of apparent impedance Zapp from the
be maintained.
load area to the line impedance ZL, a power swing will cause
Prediction of an impending loss of synchronism, or 'out-of- Zapp to move at a progressive rate and will be detected if the
step' condition, using signals from just a single point in the sequential pickup of the outer and inner mho units takes more
system can be achieved using conventional or novel than a predetermined time. It may be considered unstable if
algorithms. the blinder line is crossed, and stable if Zapp returns to the load
area. Typically, this scheme will trip upon unstable swings and
Unlike conventional out-of-step relaying introduced in the
block for stable ones.
earlier chapter on distance protection, the novel algorithms
provide a system integrity protection scheme designed to When tripping is required, excessive breaker stresses will be
predict out-of-step conditions on a transmission system. This reduced if the two systems to be separated do not move close
prediction can be used to initiate actions to preserve power to anti-phase. It is therefore highly desirable either to
system stability such as engaging surge arrestors in order to anticipate the unstable swing or to delay the trip signal until
protect primary equipment against over-voltages before they after the pole slip. Anticipating the instability can rarely be
exceed the insulation rating, or islanding the system in done with such a scheme however, because the Zapp trajectory
advance into areas of equilibrium resulting in the minimum is hard to predict for all conditions prevailing in a real system
loss of load or generation. and tripping on a stable swing is not desirable. On the other
hand, with a delayed trip signal, the magnitude and duration
of voltage excursions will be more severe and will decrease the
chances of island survival.

These difficulties can be overcome with more sophisticated

schemes for predicting the dynamic behaviour of power
swings.

20-17
Protection & Automation Application Guide

20.4.3.2 Predictive Out-of-Step 20.4.4 Acknowledgements

Rather than waiting for the out-of-step condition to occur, the GE gratefully acknowledges the research, development and
novel predictive out-of-step protection provides early warning expert assistance provided by Hydro Québec during the joint
of impending system instability. development of the DLO (MiCOM P846 Open Line Detector
and RPS (MiCOM P848 Predictive Out-Of-Step) products.
Predictive out-of-step protection can be achieved with a
special algorithm called an RPS algorithm. The
20.5 REFERENCES
implementation of this algorithm is similar to that of the
[20.1] L. H. Fink and K. Carlsen, “Operating Under Stress and
detection of open lines as already described. Kalman filters
Strain,” IEEE Spectrum, vol. 15, pp. 48–53, March
and fuzzy logic engines are used to apply multiple criteria that 1978.
are indicative of impending loss of synchronism. [20.2] CIGRE Task Force 38.02.19: “System Protection
Schemes in Power Networks”, Final Draft v 5.0, 2000.
A specially designed Kalman filter extracts the voltage and
[20.3] R Grondin, S Richards, et al, “Loss of Synchronism
current positive-sequence phasors used as the primary
Detection. A Strategic Function for Power System
variables in these applications. Protection”, CIGRE B5-205, 2006.
A transmission line model is used to compute values for the
remote voltage and current phasors Vr and Ir, given the local
phasors Vs and Is. These sets are then used to derive decisional
variables such as line angle and frequency shifts, local and
remote Vcosφ and derivative quantities.

The RPS variables are used to track and anticipate the stable
or unstable evolution of power swings across the transmission
line. Three angular variables, one concavity index, two voltage
and six Vcosφ variables are derived and processed to prepare
them for the fuzzy-logic algorithm.

Any one of these variables individually can provide an

indication of system stability, but given the wide range of
operating conditions found in large power systems, the
performance of single-variable-based algorithms is limited
because of the margin for error in any one of them. Fuzzy logic
enables multi-criteria algorithms to take advantage of the
strength of all the variables by considering them all together to
assess the stability condition. The construction of the fuzzy
logic to implement the RPS algorithm is similar to that already
described, with fuzzification giving linguistic categories to the
signals and the fuzzy logic applying weighted rules to
determine the RPS output, indicating whether the system is
predicted to go out-of-step or not.

In practical applications, the out-of-step prediction is raised

almost half a second before the instability, giving time to
initiate remedial actions to prevent the event and protect the
system.

20-18
Chapter 20 ⋅ System Integrity Protection Schemes

20-19
RELAY TESTING
AND COMMISSIONING
Chapter 21
RELAY TESTING AND
COMMISSIONING

21.1 Introduction 21.1 INTRODUCTION

21.2 Electrical Type Tests The testing of protection equipment schemes presents a number
21.3 Electromagnetic Compatibility Tests of problems. This is because the main function of protection
equipment is solely concerned with operation under system
21.4 Product Safety Type Tests fault conditions, and cannot readily be tested under normal
21.5 Environmental Type Tests system operating conditions. This situation is aggravated by the
increasing complexity of protection schemes, the use of relays
21.6 Software Type Tests with extensive functionality implemented in software, and the
21.7 Dynamic Validation Type Testing trend towards peer-to-peer logic implemented via Ethernet (as
opposed to more traditional hardwired approaches).
21.8 Production Testing
The testing of protection equipment may be divided into four
21.9 Commissioning Tests stages:
21.10 Secondary Injection Test Equipment .. type tests
21.11 Secondary Injection Testing .. routine factory production tests
21.12 Primary Injection Tests .. commissioning tests
21.13 Testing of Protection Scheme Logic .. periodic maintenance tests
21.14 Tripping and Alarm Annunciation Tests
21.1.1 TYPE TESTS
21.15 Periodic Maintenance Tests
21.16 Protection Scheme Design for Type tests are required to prove that a relay meets the published
specification and complies with all relevant standards. Since the
Maintenance principal function of a protection relay is to operate correctly
under abnormal power conditions, it is essential that the perfor-
mance be assessed under such conditions. Comprehensive
type tests simulating the operational conditions are therefore
conducted at the manufacturer’s works during the development
and certification of the equipment.
The standards that cover most aspects of relay performance are
IEC 60255 and IEEE C37.90. However compliance may also
involve consideration of the requirements of IEC 61000, 60068
and 60529, while products intended for use in the EU also have
to comply with the requirements of Directives (see Appendix C).
Since type testing of a digital or numerical relay involves testing
of software as well as hardware, the type testing process is very
complicated and more involved than a static or electrome-
chanical relay.

GEGridSolutions.com 21-1
Protection & Automation Application Guide

21.1.2 Routine Factory Production Tests Element Range Step Size

These are conducted to prove that relays are free from defects Directionality Forward/Reverse/Non-directional

during manufacture. Testing will take place at several stages RCA -95° to +95° 1°
Characteristic DT/IDMT
during manufacture, to ensure problems are discovered at the
Definite Time Delay 0 - 100s 0.01s
earliest possible time and hence minimise remedial work. The
IEC Standard Inverse
extent of testing will be determined by the complexity of the
IEC Very Inverse
relay and past manufacturing experience. IEC IDMT Time Delay
IEC Extremely Inverse
UK Long Time Inverse
21.1.3 Commissioning Tests Time Multiplier Setting (TMS) 0.025 - 1.2 0.005

These tests are designed to prove that a particular protection IEEE Moderately Inverse
IEEE Very Inverse
scheme has been installed correctly prior to setting to work.
IEEE IDMT Time Delay IEEE Extremely Inverse
All aspects of the scheme are thoroughly checked, from US-C08 Inverse
installation of the correct equipment through wiring checks US-C02 Short Time Inverse
and operation checks of the individual items of equipment, Time Dial (TD) 0.5 - 15 0.1
finishing with testing of the complete scheme. IEC Reset Time (DT only) 0 - 100s 0.01s
IEEE Reset Time IDMT/DT
21.1.4 Periodic Maintenance Checks IEEE DT Reset Time 0 - 100s 0.01s

These are required to identify equipment failures and IEEE Moderately Inverse
IEEE Very Inverse
degradation in service, so that corrective action can be taken.
IEEE IDMT Reset Time IEEE Extremely Inverse
Because a protection scheme only operates under fault US-C08 Inverse
conditions, defects may not be revealed for a significant period US-C02 Short Time Inverse

of time, until a fault occurs. Regular testing assists in Table 21.1: Overcurrent relay element specification

detecting faults that would otherwise remain undetected until Test no. Description
a fault occurs. Three phase non-directional pick up and drop off accuracy over complete current
Test 1
setting range for both stages

21.2 ELECTRICAL TYPE TESTS Test 2

Three phase directional pick up and drop off accuracy over complete RCA setting
range in the forward direction, current angle sweep
Various electrical type tests must be performed, as follows: Three phase directional pick up and drop off accuracy over complete RCA setting
Test 3
range in the reverse direction, current angle sweep
21.2.1 Functional Tests Three phase directional pick up and drop off accuracy over complete RCA setting
Test 4
range in the forward direction, voltage angle sweep
The functional tests consist of applying the appropriate inputs
Three phase directional pick up and drop off accuracy over complete RCA setting
to the relay under test and measuring the performance to Test 5
range in the reverse direction, voltage angle sweep
determine if it meets the specification. They are usually carried Test 6 Three phase polarising voltage threshold test
out under controlled environmental conditions. The testing Test 7 Accuracy of DT timer over complete setting range
may be extensive, even where only a simple relay function is Test 8 Accuracy of IDMT curves over claimed accuracy range
being tested., as can be realised by considering the simple Test 9 Accuracy of IDMT TMS/TD
overcurrent relay element of Table 21.1. Test 10 Effect of changing fault current on IDMT operating times
Test 11 Minimum Pick-Up of Starts and Trips for IDMT curves
To determine compliance with the specification, the tests listed
Test 12 Accuracy of reset timers
in Table 21.2 are required to be carried out. This is a time
Test 13 Effect of any blocking signals, opto inputs, VTS, Autoreclose
consuming task, involving many engineers and technicians.
Test 14 Voltage polarisation memory
Hence it is expensive. Table 21.2: Overcurrent relay element functional type tests

Element Range Step Size When a modern numerical relay with many functions is
I>1 0.08 - 4.00In 0.01In considered, each of which has to be type-tested, the functional
I>2 0.08 - 32In 0.01In type-testing involved is a major issue. In the case of a recent

21-2
 Chapter 21 ⋅ Relay Testing and Commissioning

relay development project, it was calculated that if one person they comply with the product specification. Particular
had to do all the work, it would take 4 years to write the withstand and endurance type tests have to be carried out
functional type-test specifications, 30 years to perform the using d.c., since the normal supply is via a station battery.
tests and several years to write the test reports that result.
Automated techniques/ equipment are clearly required, and 21.2.7 Insulation Resistance
are covered in Section 21.7.2. The insulation resistance test is carried out according to IEC
60255-27, i.e. 500V d.c. ± 10%, for a minimum of 5 seconds.
21.2.2 Rating Tests This is carried out between all circuits and case earth, between
Rating type tests are conducted to ensure that components are all independent circuits and across normally open contacts.
used within their specified ratings and that there are no fire or The acceptance criterion for a product in new condition is a
electric shock hazards under a normal load or fault condition of minimum of 100MΩ. After a damp heat test the pass
the power system. This is in addition to checking that the criterion is a minimum of 10MΩ.
product complies with its technical specification. The following
are amongst the rating type tests conducted on protection 21.2.8 Auxiliary Supplies
relays, the specified parameters are normally to IEC 60255-1. Digital and numerical protection relays normally require an
auxiliary supply to provide power to the on-board
21.2.3 Thermal Withstand microprocessor circuitry and the interfacing opto-isolated input
The thermal withstand of VTs, CTs and output contact circuits circuits and output relays. The auxiliary supply can be either
is determined to ensure compliance with the specified a.c. or d.c., supplied from a number of sources or safe supplies
continuous and short-term overload conditions. In addition to - i.e. batteries, UPSs, etc., all of which may be subject to
functional verification, the pass criterion is that there is no voltage dips, short interruptions and voltage variations. Relays
detrimental effect on the relay assembly, or circuit are designed to ensure that operation is maintained and no
components, when the product is subjected to overload damage occurs during a disturbance of the auxiliary supply.
conditions that may be expected in service. Thermal withstand
Tests are carried out for both a.c. and d.c. auxiliary supplies
is assessed over a time period of 1s for CTs and 10s for VTs.
and include mains variation both above and below the nominal
21.2.4 Relay Burden rating, supply interruptions derived by open circuit and short
circuit, supply dips as a percentage of the nominal supply,
The burdens of the auxiliary supply, optically isolated inputs,
repetitive starts. The duration of the interruptions and supply
VTs and CTs are measured to check that the product complies
dips range from 2ms to 60s intervals. A short supply
with its specification. The burden of products with a high
interruption or dip up to 20ms, possibly longer, should not
number of input/output circuits is application specific i.e. it
cause any malfunction of the relay. Malfunctions include the
increases according to the number of optically isolated input
operation of output relays and watchdog contacts, the reset of
and output contact ports which are energised under normal
microprocessors, alarm or trip indication, acceptance of
power system load conditions. It is usually envisaged that not
corrupted data over the communication link and the corruption
more than 50% of such ports will be energised concurrently in
of stored data or settings. For a longer supply interruption, or
any application.
dip in excess of 50ms, the relay self recovers without the loss
of any function, data, settings or corruption of data. No
21.2.5 Relay Inputs
operator intervention is required to restore operation after an
Relay inputs are tested over the specified ranges. Inputs
interruption or dip in the supply.
include those for auxiliary voltage, VT, CT, frequency, optically
isolated digital inputs and communication circuits. In addition to the above, the relay is subjected a number of
repetitive starts or a sequence of supply interruptions. Again
21.2.6 Relay Output Contacts the relay is tested to ensure that no damage or data corruption
Protection relay output contacts are type tested to ensure that has occurred during the repetitive tests.

21-3
Protection & Automation Application Guide

Specific tests carried out on d.c. auxiliary supplies include • high frequency disturbance test
reverse polarity, a.c. waveform superimposed on the d.c. • fast transient test
supply and the effect of a rising and decaying auxiliary voltage. • surge immunity test
All tests are carried out at various levels of loading of the relay
• power frequency interference test
auxiliary supply.
• electrostatic discharge test

21.3 ELECTROMAGNETIC COMPATIBILITY • conducted and radiated emissions tests

TESTS • conducted and radiated immunity tests
There are numerous tests that are carried out to determine the • magnetic field tests
ability of relays to withstand the electrical environment in
21.3.1 D.C Interrupt Test
which they are installed. The substation environment is a very
severe environment in terms of the electrical and This is a test to determine the maximum length of time that
electromagnetic interference that can arise. There are many the relay can withstand an interruption in the auxiliary supply
sources of interference within a substation, some originating without de-energising, e.g. switching off, and that when this
internally, others being conducted along the overhead lines or time is exceeded and it does transiently switch off, that no
cables into the substation from external disturbances. The maloperation occurs.
most common sources are: It simulates the effect of a loose fuse in the battery circuit, or a
• switching operations short circuit in the common d.c. supply, interrupted by a fuse.
Another source of d.c. interruption is if there is a power system
• system faults
fault and the battery is supplying both the relay and the circuit
• lightning strikes
breaker trip coils. When the battery energises the coils to
• conductor flashover
initiate the circuit breaker trip, the voltage may fall below the
• telecommunication operations e.g. mobile phones
required level for operation of the relay and hence a d.c.
A whole suite of tests are performed to simulate these types of interrupt occurs. The test is specified in IEC 60255-11 and
interference, and they fall under the broad umbrella of what is comprises interruptions of 10, 20, 30, 50, 100, 200, 300, 500,
known as EMC, or Electromagnetic Compatibility tests. 1000, and 5000ms. For interruptions lasting up to and
including 20ms, the relay must not de-energise of maloperate,
Broadly speaking, EMC can be defined as:
while for longer interruptions it must not maloperate. Many
‘The ability of different equipment to co-exist in the same modern products are capable of remaining energised for
electromagnetic environment’ interruptions up to 50ms
It is not a new subject and has been tested for by the military The relay is powered from a battery supply, and both short
ever since the advent of electronic equipment. EMC can cause circuit and open circuit interruptions are carried out. Each
real and serious problems, and does need to be taken into interruption is applied 10 times, and for auxiliary power
account when designing electronic equipment. supplies with a large operating range, the tests are performed
EMC tests determine the impact on the relay under test of at minimum, maximum, and other voltages across this range,
high-frequency electrical disturbances of various kinds. Relays to ensure compliance over the complete range.
manufactured or intended for use in the EU have to comply
21.3.2 A.C. Ripple on D.C. Supply
with Directive 2004/108/EC in this respect. To achieve this, in
addition to designing for statutory compliance to this Directive, This test (IEC 60255-11) determines that the relay is able to
the following range of tests are carried out: operate correctly with a superimposed a.c. voltage on the d.c.
supply. This is caused by the station battery being charged by
• d.c. interrupt test
the battery charger, and the relevant waveform is shown in
• a.c. ripple on d.c. supply test Figure 21.1. It consists of a 15% peak-to-peak ripple
• d.c. ramp test superimposed on the d.c. supply voltage.

21-4
 Chapter 21 ⋅ Relay Testing and Commissioning

60.00

50.00

40.00
Voltage (V)

Voltage
30.00 0
Time
20.00

10.00

0.00
1
88
175
262
349
436
523
610
697
784
871
958
1045
1132
1219
1306
1393
Figure 21.2: High Frequency Disturbance Test waveform
Time (ms)
Figure 21.1: A.C. ripple superimposed on d.c. supply test 21.3.5 Fast Transient Test
For auxiliary power supplies with a large operating range, the The Fast Transient Test simulates the HV interference caused
tests are performed at minimum, maximum, and other by disconnector operations in GIS substations or breakdown of
voltages across this range, to ensure compliance for the the SF6 insulation between conductors and the earthed
complete range. The interference is applied using a full wave enclosure. This interference can either be inductively coupled
rectifier network, connected in parallel with the battery supply. onto relay circuits or can be directly introduced via the CT or
The relay must continue to operate without malfunction during VT inputs. It consists of a series of 15ms duration bursts at
the test. 300ms intervals, each burst consisting of a train of 50ns wide
pulses with very fast (5ns typical) rise times (Figure 21.3),
21.3.3 D.C. Ramp Down/Ramp Up with a peak voltage magnitude of 4kV.
This test simulates a failed station battery charger, which
Burst period, 300 ms
would result in the auxiliary voltage to the relay slowly ramping
V Burst duration (1/15 ms)
down. The ramp up part simulates the battery being recharged
after discharging. The relay must power up cleanly when the
voltage is applied and not maloperate. t

There is no international standard for this test, so individual

V 5 ns rise time, 50 ns pulse width
manufacturers can decide if they wish to conduct such a test
and what the test specification shall be.
Repetition period t
21.3.4 High Frequency Disturbance Test
Figure 21.3 Fast Transient Test waveform
The High Frequency Disturbance Test simulates high voltage The product is energised in both normal (quiescent) and
transients that result from power system faults and plant tripped modes for this test. It must not maloperate when the
switching operations. It consists of a 1MHz decaying interference is applied in common mode via the integral
sinusoidal waveform, as shown in Figure 21.2. The coupling network to each circuit in turn, for 60 seconds.
interference is applied across each independent circuit Interference is coupled onto communications circuits, if
(differential mode) and between each independent circuit and required, using an external capacitive coupling clamp.
earth (common mode) via an external coupling and switching
network. The product is energised in both normal (quiescent) 21.3.6 Surge Immunity Test
and tripped modes for this test, and must not maloperate The Surge Immunity Test simulates interference caused by
when the interference is applied for a 2 second duration. major power system disturbances such as capacitor bank
switching and lightning strikes on overhead lines within 5km
of the substation. The test waveform has an open circuit
voltage of 4kV for common mode surges and 2kV for

21-5
Protection & Automation Application Guide

differential mode surges. The waveshape consists on open of the relay, with the cover both on and off. Two types of
circuit of a 1.2/50µs rise/fall time and a short circuit current of discharges are applied, air discharge and contact discharge. Air
8/20µs rise/fall time. The generator is capable of providing a discharges are used on surfaces that are normally insulators,
short circuit test current of up to 2kA, making this test and contact discharges are used on surfaces that are normally
potentially destructive. The surges are applied sequentially conducting. IEC 60255-22-2 is the relevant standard this test,
under software control via dedicated coupling networks in both for which the test parameters are:
differential and common modes with the product energised in
• cover on: Class 4, 8kV contact discharge, 15kV air
its normal (quiescent) state. The product shall not maloperate discharge
during the test, shall still operate within specification after the • cover off: Class 3, 6kV contact discharge, 8kV air
test sequence and shall not incur any permanent damage. discharge

21.3.7 Power Frequency Interference In both cases above, all the lower test levels are also tested.
This test simulates the type of interference that is caused when The discharge current waveform is shown in Figure 21.4.
there is a power system fault and very high levels of fault
100
current flow in the primary conductors or the earth grid. This 90
Rise Time = 0.7 to 1.0 ns.
causes 50 or 60Hz interference to be induced onto control and 80
Current specified for 30 ns and 60 ns
Current, % of Peak

70
communications circuits. 60
50
There is no international standard for this test, but one used by 40

some Utilities is: 30

20

• 500V r.m.s., common mode 10

0
• 250V r.m.s., differential mode 0 10 20 30 40 50 60 70 80 90

Time, ns
applied to circuits for which power system inputs are not Figure 21.4: ESD Current Waveform
connected.
The test is performed with single discharges repeated on each
Tests are carried out on each circuit, with the relay in the test point 10 times with positive polarity and 10 times with
following modes of operation: negative polarity at each test level. The time interval between
successive discharges is greater than 1 second. Tests are
• current and voltage applied at 90% of setting, (relay not
tripped) carried out at each level, with the relay in the following modes
of operation:
• current and voltage applied at 110% of setting, (relay
tripped) • current and voltage applied at 90% of setting, (relay not
• main protection and communications functions are tripped)
tested to determine the effect of the interference • current and voltage applied at 110% of setting, (relay
tripped)
The relay shall not maloperate during the test, and shall still
• main protection and communications functions are
perform its main functions within the claimed tolerance.
tested to determine the effect of the discharge
21.3.8 Electrostatic Discharge Test To pass, the relay shall not maloperate, and shall still perform
This test simulates the type of high voltage interference that its main functions within the claimed tolerance.
occurs when an operator touches the relay’s front panel after
being charged to a high potential. This is exactly the same 21.3.9 Conducted and Radiated Emissions Tests
phenomenon as getting an electric shock when stepping out of These tests arise primarily from the essential protection
a car or after walking on a synthetic fibre carpet. requirements of the EU directive on EMC. These require
manufacturers to ensure that any equipment to be sold in the
In this case the discharge is only ever applied to the front panel
countries comprising the European Union must not interfere

21-6
 Chapter 21 ⋅ Relay Testing and Commissioning

with other equipment. To achieve this it is necessary to Screened room

measure the emissions from the equipment and ensure that Ante-chamber
they are below the specified limits.
Access panel
E.U.T. Support/analysis
Conducted emissions are measured only from the equipment’s Impedance network equipment

power supply ports and are to ensure that when connected to

a mains network, the equipment does not inject interference
back into the network which could adversely affect the other
(a) Conducted EMC emissions test arrangement
equipment connected to the network.

Radiated emissions measurements are to ensure that the

10m
interference radiated from the equipment is not at a level that
could cause interference to other equipment. This test is
normally carried out on an Open Area Test Site (OATS) where
there are no reflecting structures or sources of radiation, and Antenna
therefore the measurements obtained are a true indication of E.U.T.

the emission spectrum of the relay. An example of a plot

obtained during conducted emissions tests is shown in Figure Turntable

21.5. Earth Plane

100 Quasi-peak limits

90 Average limits (b) Radiated Emissions test arrangement on an OATS
80 Typical trace
E.U.T. = Equipment under test
Emissions Level, dBuV

70
60
Figure 21.6: EMC test arrangements
50
When performing these two tests, the relay is in a quiescent
40
30 condition, that is not tripped, with currents and voltages
20
applied at 90% of the setting values. This is because for the
10
0 majority of its life, the relay will be in the quiescent state and
0.1 1 10 100
Frequency, MHz the emission of electromagnetic interference when the relay is
Figure 21.5: EMC Conducted Emissions test - example test plot tripped is considered to be of no significance. Tests are
The test arrangements for the conducted and radiated conducted in accordance with IEC 60255-25 and EN 55022,
emissions tests are shown in Figure 21.6. and are detailed in Table 21.3.

Frequency Range Specified Limits Test Limits

Radiated 30 - 230MHz 30dB(µV/m) at 30m 40dB(µV/m) at 10m
Radiated 230 - 1000MHz 37dB(µV/m) at 30m 47dB(µV/m) at 10m
79dB(µV) quasi-peak 79dB(µV) quasi-peak
Conducted 0.15 - 0.5MHz
66dB(µV) average 66dB(µV) average
73dB(µV) quasi-peak 73dB(µV) quasi-peak
Conducted 0.5 - 30MHz
60dB(µV) average 60dB(µV) average

Table 21.3: Test criteria for Conducted and Radiated Emissions tests

21.3.10 Conducted and Radiated Immunity Tests

These tests are designed to ensure that the equipment is
immune to levels of interference that it may be subjected to.
The two tests, conducted and radiated, arise from the fact that
for a conductor to be an efficient antenna, it must have a

21-7
Protection & Automation Application Guide

length of at least ¼ of the wavelength of the electromagnetic particular, the frequencies of mobile communications
wave it is required to conduct. equipment used by personnel working in the substation. This is
to ensure that when working in the vicinity of a relay, the
If a relay were to be subjected to radiated interference at
personnel should be able to operate their radios/mobile phones
150kHz, then a conductor length of at least
without fear of relay maloperation.
300 × 10 6
λ= = 500m IEC 60255-22-3 specifies the radiated immunity tests to be
(
150 × 10 3 × 4 ) conducted (ANSI/IEEE C37.90.2 is used for equipment
compliant with North American standards), with signal levels
of:
would be needed to conduct the interference. Even with all the
cabling attached and with the longest PCB track length taken • IEC: Class III, 10V/m, 80MHz -2700MHz
into account, it would be highly unlikely that the relay would • ANSI/IEEE: 35V/m 25MHz - 1000MHz with no
be able to conduct radiation of this frequency, and the test modulation, and again with 100% pulse modulation
therefore, would have no effect. The interference has to be IEC 60255-22-6 is used for the conducted immunity test, with
physically introduced by conduction, hence the conducted a test level of:
immunity test. However, at the radiated immunity lower
frequency limit of 80MHz, a conductor length of approximately • Class III, 10V r.m.s., 150kHz - 80MHz.
1.0m is required. At this frequency, radiated immunity tests
21.3.11 Power Frequency Magnetic Field Tests
can be performed with the confidence that the relay will
conduct this interference, through a combination of the These tests are designed to ensure that the equipment is
attached cabling and the PCB tracks. immune to magnetic interference. The three tests, steady
state, pulsed and damped oscillatory magnetic field, arise from
Although the test standards state that all 6 faces of the the fact that for different site conditions the level and
equipment should be subjected to the interference, in practice waveshape is altered.
this is not carried out. Applying interference to the sides and
top and bottom of the relay would have little effect as the 21.3.11.1 Steady state magnetic field tests
circuitry inside is effectively screened by the earthed metal These tests simulate the magnetic field that would be
case. However, the front and rear of the relay are not experienced by a device located within close proximity of the
completely enclosed by metal and are therefore not at all well power system. Testing is carried out by subjecting the relay to
screened, and can be regarded as an EMC hole. a magnetic field generated by two induction coils. The relay is
Electromagnetic interference when directed at the front and rotated such that in each axis it is subjected to the full
back of the relay can enter freely onto the PCBs inside. magnetic field strength. IEC 61000-4-8 is the relevant
When performing these two tests, the relay is in a quiescent standard, using a signal level of:
condition, that is not tripped, with currents and voltages
• Level 5: 300A/m continuous and 1000A/m short
applied at 90% of the setting values. This is because for the
duration
majority of its life, the relay will be in the quiescent state and
the coincidence of an electromagnetic disturbance and a fault The test arrangement is shown in Figure 21.7.
is considered to be unlikely.

However, spot checks are performed at selected frequencies

when the main protection and control functions of the relay
are exercised, to ensure that it will operate as expected, should
it be required to do so.

The frequencies for the spot checks are in general selected to

coincide with the radio frequency broadcast bands, and in

21-8
 Chapter 21 ⋅ Relay Testing and Commissioning

essential requirements are that the relay is safe and will not
cause an electric shock or fire hazard under normal conditions
Induction Coil and in the presence of a single fault. A number of specific tests
to prove this may be carried out, as follows.

21.4.1 Dielectric Voltage Withstand

Induction Coil E.U.T.
Dielectric Voltage Withstand testing is carried out as a routine
test i.e. on every unit prior to despatch. The purpose of this test
is to ensure that the product build is as intended by design.
This is done by verifying the clearance in air, thus ensuring that
Ground Plane the product is safe to operate under normal use conditions.
The following tests are conducted unless otherwise specified in
the product documentation:

• 2.0kV r.m.s., 50/60Hz for 1 minute between all

E.U.T. - Equipment Under Test terminals and case earth and also between independent
Figure 21.7: Power frequency magnetic field test arrangement circuits, in accordance with IEC 60255-27. Some
communication circuits are excluded from this test, or
To pass the steady-state test, the relay shall not maloperate,
have modified test requirements e.g. those using D-
and shall still perform its main functions within the claimed
type connectors
tolerance. During the application of the short duration test, the
• 1.5kV r.m.s., 50/60Hz for 1 minute across normally
main protection function shall be exercised and verified that open contacts intended for connection to tripping
the operating characteristics of the relay are unaffected. circuits, in accordance with ANSI/IEEE C37.90
• 1.0kV r.m.s., 50/60Hz for 1 minute across the normally
21.3.11.2 Pulsed magnetic field
open contacts of watchdog or changeover output
These tests simulate the magnetic field that would be relays, in accordance with IEC 60255-27
experienced by a device located within close proximity of the
power system during a transient fault condition. According to The routine dielectric voltage withstand test time may be
IEC 61000-4-9, the generator for the induction coils shall shorter than for the 1 minute type test time, to allow a
reasonable production throughput, e.g. for a minimum of 1
produce a 6.4/16µs waveshape with test level 5, 100A/m with
second at 110% of the voltage specified for 1 minute.
the equipment configured as for the steady state magnetic field
test. The relay shall not maloperate, and shall still perform its
21.4.2 Insulation Withstand for Overvoltages
main functions within the claimed tolerance during the test.
The purpose of the High Voltage Impulse Withstand type test
21.3.11.3 Damped oscillatory magnetic field is to ensure that circuits and their components will withstand
overvoltages on the power system caused by lightning. Three
These tests simulate the magnetic field that would be
positive and three negative high voltage impulses, 5kV peak,
experienced by a device located within close proximity of the
are applied between all circuits and the case earth and also
power system during a transient fault condition. IEC 61000-4-
between the terminals of independent circuits (but not across
10 specifies that the generator for the coil shall produce an
normally open contacts). As before, different requirements
oscillatory waveshape with a frequency of 0.1MHz and 1MHz,
apply in the case of circuits using D-type connectors.
to give a signal level in accordance with Level 5 of 100A/m,
and the equipment shall be configured as in Figure 21.7. The test generator characteristics are as specified in IEC
60255-27 and are shown in Figure 21.8. No disruptive
21.4 PRODUCT SAFETY TYPE TESTS discharge (i.e. flashover or puncture) is allowed.
A number of tests are carried out to demonstrate that the
If it is necessary to repeat either the Dielectric Voltage or High
product is safe when used for its intended application. The

21-9
Protection & Automation Application Guide

Voltage Impulse Withstand tests these should be carried out at 21.4.6 Other Regional or Industry Norms
75% of the specified level, in accordance with IEC 60255-27, Many products intended for application in North America will
to avoid overstressing insulation and components. additionally need certification to the requirements of UL
(Underwriter's Laboratory), or CUL for Canada. Products for
application in mining or explosive environments will often be
required to demonstrate an ATEX claim

21.5 ENVIRONMENTAL TYPE TESTS

Various tests have to be conducted to prove that a relay can
Voltage

withstand the effects of the environment in which it is

expected to work. They consist of: the following tests:

• temperature
• humidity
• enclosure protection
• mechanical
5kV peak Duration (50 %) = 50µs These tests are described in the following sections.
Figure 21.8: Test generator characteristics for insulation withstand test
21.5.1 Temperature Test
21.4.3 Single Fault Condition Assessment
Temperature tests are performed to ensure that a product can
An assessment is made of whether a single fault condition withstand extremes in temperatures, both hot and cold, during
such as an overload, or an open or short circuit, applied to the transit, storage and operating conditions. Storage and transit
product may cause an electric shock or fire hazard. In the case conditions are defined as a temperature range of –25oC to
of doubt, type testing is carried out to ensure that the product +70oC and operating as –25oC to +55oC. Many products now
is safe. claim operating temperatures of +70°C or even higher.

21.4.4 Earth Bonding Impedance Dry heat withstand tests are performed at 70°C for 96 hours
Class 1 products that rely on a protective earth connection for with the relay de-energised. Cold withstand tests are
safety are subjected to an earth bonding impedance (EBI) type performed at –40°C for 96 hours with the relay de-energised.
test. This ensures that the earth path between the protective Operating range tests are carried out with the product
earth connection and any accessible earthed part is sufficiently energised, checking all main functions operate within tolerance
low to avoid damage in the event of a single fault occurring. over the specified working temperature range –25°C to +55°C.
The test is conducted using a test voltage of 12V maximum
21.5.2 Humidity Test
and a test current of twice the recommended maximum
protective fuse rating. After 1 minute with the current flowing The humidity test is performed to ensure that the product will
withstand and operate correctly when subjected to 93% relative
in the circuit under test, the EBI shall not exceed 0.1Ω.
humidity at a constant temperature of 40°C for 56 days. Tests
21.4.5 CE Marking are performed to ensure that the product functions correctly
within specification after 21 and 56 days. After the test, visual
A CE mark on the product, or its packaging, shows that
inspections are made for any signs of unacceptable corrosion
compliance is claimed against relevant EU directives e.g. Low
Voltage Directive 2006/95/EC and Electromagnetic and mould growth.
Compatibility (EMC) Directive 2004/108/EC.
21.5.3 Cyclic Temperature/Humidity Test
This is a short-term test that stresses the relay by subjecting it
to temperature cycling in conjunction with high humidity.

21-10
 Chapter 21 ⋅ Relay Testing and Commissioning

The test does not replace the 56 day humidity test, but is used or growth, and photographs are usually taken of each PCB and
for testing extension to ranges or minor modifications to prove the case as a record of this.
that the design is unaffected.
21.5.4 Enclosure Protection Test
The applicable standard is IEC60068-2-30 and test conditions of:
Enclosure protection tests prove that the casing system and
• +25°C ±3°C and 95% relative humidity/+55°C ±2°C connectors on the product protect against the ingress of dust,
• 95% relative humidity moisture, water droplets (striking the case at pre-defined
are used, over the 24 hour cycle shown in Figure 21.9. angles) and other pollutants. An ‘acceptable’ level of dust or
water may penetrate the case during testing, but must not
100 96%
90 95% impair normal product operation, safety or cause tracking
Relative humidity %

95%
90%
80
70 15min
80% across insulated parts of connectors.
15min
15min Start of
temperature
fall 21.5.5 Mechanical Tests
End of temperature rise

Upper Temperature +2°C

Time Mechanical tests simulate a number of different mechanical
+55°C
Upper Temperature -2°C
conditions that the product may have to endure during its
lifetime. These fall into two categories
Ambient Temperature °C

• response to disturbances while energised

• response to disturbances during transportation (de-
energised state)
± 0.5h
+28°C Tests in the first category are concerned with the response to
+25°C
3h +22°C Time
vibration, shock and seismic disturbance. The tests are
3h

12h± 0.5h 6h designed to simulate normal in-service conditions for the

24h product, for example earthquakes. These tests are performed in
Figure 21.9: Cyclic temperature/humidity test profile all three axes, with the product energised in its normal
(quiescent) state. During the test, all output contacts are
For these tests the relay is placed in a humidity cabinet, and
continually monitored for change using contact follower
energised with normal in-service quantities for the complete
circuits. Vibration levels of 1g, over a 10Hz-150Hz frequency
duration of the tests. In practical terms this usually means
sweep are used. Seismic tests use excitation in a single axis,
energising the relay with currents and voltages such that it is
using a test frequency of 35Hz and peak displacements of
10% from the threshold for operation. Throughout the duration
7.5mm and 3.5mm in the x and y axes respectively below the
of the test the relay is monitored to ensure that no unwanted
crossover frequency and peak accelerations of 2.0gn and 1.0gn
operations occur.
in these axes above the crossover frequency.
Once the relay is removed from the humidity cabinet, its
The second category consists of vibration endurance, shock
insulation resistance is measured to ensure that it has not
withstand and bump tests. They are designed to simulate the
deteriorated to below the claimed level. The relay is then
longer-term affects of shock and vibration that could occur
functionally tested again, and finally dismantled to check for
during transportation. These tests are performed with the
signs of component corrosion and growth.
product de-energised. After these tests, the product must still
The acceptance criterion is that no unwanted operations shall operate within its specification and show no signs of
occur including transient operation of indicating devices. After permanent mechanical damage. Equipment undergoing a
the test the relay’s insulation resistance should not have seismic type test is shown in Figure 21.10, while the
significantly reduced, and it should perform all of its main waveform for the shock/bump test is shown in Figure 21.11
protection and communications functions within the claimed
tolerance. The relay should also suffer no significant corrosion

21-11