You are on page 1of 94


The subject entitled ‘Electrical Protection’ is one of a series which comes under the heading
of Electrical Technology. This and other series have been prepared for our employees who
attend our trai ning programmes. .~ _..

The three ‘Fundamentals of Electricity’ manuals introduce the subject of electrical power
and provide a basis for further study of related’ topics such as power systems, generation,
distribution, motors,~control, protection a6d electrical safety. This manual has been designed
so that it can be used for self-study;~and, w.ith this in mind, a series of questions and answers
has been incorporated.

A list of the electrical power manuals is included inside the front cover, and a list of all
training manuals in the s’eries is given on the inside of the back cover.

Harry To/t
Shell Expro Training, Aberdeen.

Shell Expro is the major operator exploring for and producing oil and gas in the United
Kingdom -~ working alone and in joint ventures with third parties. In the North Sea it is the
operator for a SO/50 joint venture with Esso, where the projected output from fields already
in operation will meet more than a third of the UK’s estimated oil needs and over 12% of its
natural gas requirements.

._ --
Chapter 1’ Introduction to Protection 1

Chapter 2 Faults and Fault Levels 4

Chapter 3 Overcurrent Protection 11

Chapter 4. Discrimination 30

Chapter 5 Earth-fault and Earth-leakage Protection 34

Chapter 6 Differential Protection 42

Chapter 7 Undervoltage and System Undervoltage Protection 48

Chapter 8 Additional Forms of Protection 50

Chapter 9 Generator Protection 52

Chapter 10 Transformer Protection. 59

Chapter 11 Cable Protection 65

Chapter 12 Motor Protection 68

Chapter 13 Questions and~Answers 80




Electrical plant, machines and distribution systems must be protected against damage which
may occur through abnormal conditions arising.

Abnormal conditions may be grouped into two types:

(a) Operation outside the designed ratingsde t~bverloadifigor incorrectfunction-

ing of the system.

(b) Fault conditions due usuallyvto breakdown of some part’of the system.

Condition (a) is usually ‘chronic’ -that ii, it may persist for some time and is often accept-
able for a limited period. It may give rise to temperatures outside the design limit of the
machines and equipment, but, unless these are very excessive or very prolonged, it seldom
causes sudden or catastrophic failure. It can usually be corrected before it leads to breakdown
or a fault condition.

Condition (b) on the other hand is ‘acute’ and arises from electrical or.,mechanical failure
which, once established, produces a condition beyond.control. It usually gives rise to very
severe excess currents which will quickly cause catastrophic failure of other electrical and
mechanical plant in the system unless the fault is rapidly isolated. It may be caused by a
breakdown of insulation due to a material failure or overheating or to external conditions
such as weathe~r, or it may be due to physical damage to an item of plant or cable.

Automatic protection against conditions (a) and (b) is possible~ in electrical installations
because it is easy to measure various parameters, to detect abtiormalities, and to set in
motion the protective action the instant an abnormality arises.

Protection of an electrical system is provided for,one or more of the following principles:

(a) To maintain electrical supplies to as much of the system as possible after a

fault has been isolated.

(b) To protect the generators and other plant against damage due to abnormal
conditions and faults.

(c) To protect the consumer equipment against damage due to abnormal conditions
(e.g. overload).

(d) To isolate faulted equipment to limit the risk of fire locally.

(e) To limit damage to the cable system resulting,from a fault.

These principles will determine the type of protective equipment fitted in any installation.
It will be noted that principle (a) conflicts with the other requirements to some extent. For
example, the best way to protect a generator againstdamageby fault currents is to disconnect
it, but it would not then be available to supply other consumers. For this reason a method
of disconnection (called ‘discrimination’) has been devised. It is fairly complex and has to be
very carefully ,engineered. ,It is dealt with in Chapter 4.

Ail the protection devices in this manual are assumed to be fed from voltage transformers
(VT) or current transformers (CT). In high-voltage systems this is always the case. In low-
voltage systems, such as 44OV, voltage-operated devices are sometimes fed direct from the
busbars through fuses without an intervening VT. Except where currents are small, current-
operated devices always use CTs even on low-voltage systems.

The above five principles relate mainly to the protection of plant and equipment in the
system. There remains of course the protection of personnel.

An arc may occur at the point of the fault. Apart from possibly burning or blinding anyone
in the vicinity, an arc in a high-power system may produce enough heat to melt heavy
copper bars or even structural steel in a very short time, and a rapidly spreading fire may
result. Arcs are particularly dangerous in areas where flammable gas may be present. It is
vital therefore that the source of power which is feeding the arc should be cut off as quickly
as possible.


The fundamental principle of protection is to disconnect and isolate the faulty part of the
system so that the fault is not sustained and aggravated by a continuing flow of power into
it, and the rest of the system is not damaged and can revert to its normal state.

Generally speaking this means automatically detecting the fault condition by means of a
suitable device and disconnecting the faulty section by means of a circuit-breaker or other
interrupter. For some purposes the two functions are combinrd in one item of switchgear,
as in a moulded-case circuit-breaker. In many cases protection is provided by fuses, in which
the functions are inseparable. Protection relays as devices xe described in the manual
‘Electrical Control Devices’, and fuses are discussed in Chapter 3 of this present manual.

The occurrence of a fault is indicated by various quantities, usually excess currents, depend-
ing upon the nature of the fault. The way in which the protection devices respond to faults,
in terms both of magnitude and of time, is very important for several reasons:

- The affected part of the system should be disconnected quickly, before any
av&dable damage is done.

- The protection should not operate unnecessarily. Transient disturbances are

liable to occur on most systems for many reasons connected with operation,
and most electrical plant is capable of operating safely with moderate over-
loads for short periods.

- If the amount of .equipment disconnected is to be kept to the minimum

necessary to clear the fault, the sensitivities of the various protection devices
which respond to the fault must, as far as possible, be so graded and related
that only that device needed to clear the fault actually operates. This is the
principle known as ‘discrimination’, which is discussed in Chapter 4.


The means of detecting a fault and the means oi disconnecting it are equally important.
There are three categories of devices used to disconnect faulty circuits:

Circuit-breakers. These are generally capable of interrupting the maximum fault
currents that can flow in the circuits which they feed. Since under some fault
conditions the current may rise to ten or more times the normal full-load current,
the design and selection of circuit-breakers is of great importance. Several types
of circuit-breaker are in use with different arrangements for arc suppression -
air-break, oil-break, sulphur hexafluoride (SF,) and vacuum; these are described in
detail in the manual ‘Electrical Distribution Equipment, Part A’,

Contactors. Contactors are rated to close onto the most severe’faults but have
limited breaking capacity; in most cases this is less than the maximum possible
fault current in the circuit which they feed. Therefore they have usually to be
supplemented by fuses. Contactors may be air-break or vacuum-break; they are
described in more detail in the manual ‘Electrical Distribution Equipment, Part A’.

Fuses. A fuse constitutes an intentional ‘weak--link’ in an electrical circuit and,,

suitably rated, is particularly apt for the quick interruptionofshort-circuit~currents.
Fuses, are described in Chapter 3. They are expendable and have to be replaced
after they have operated. They are very costly.


No system of protection can be designed without knowing the conditions in the network
which it.has to protect. This means that the level of fault currents at various pojnts of the
network must be known in advance so that the right type of switchgear may be installed
and a proper system of~protection worked,out.

The first tasks therefore is to calculate the ‘fault levels’ at all those points in the, network
where switching is to take place. Fortunately this is not a difficult calculation, and it is
described in Chapter 2.




An electrical network normally operates within its designed rating. Generators, transformers,
cables, transmission lines, switchboards, busbars and connected apparatus are each designed
to carry a certain maximum current. Most can carry a moderate overload for a short time
without undue overheating.

However, if a fault should develop somewhere in the system, that is to say a phase-to-phase
short-circuit or a phase-to-earth breakdown, then all connected generators will at first feed
extremely high currents into that fault, which will be limited only by the impedance of the
complete circuit from generator to fault. Fault currents can be ten or more times the normal
full-load current.

Such currents will quickly cause intense overheating of conductors and windings, leading to
almost certain breakdown unless the.y are quickly disconnected. They will also give rise to
severe mechanical forces between the current-carrying conductors or windings. All such
apparatus must be manufactured to withstand these forces. A fault current of 50 OOOA
(rms) flowing in two busbars 3 inches apart will produce between~them a peak mechanical
force of nearly half a ton-force per foot run of bars.

The pui-pose of automatic protection is to remove the fault from the system and so break
the fault current as quickly as possible. Before this can be achieved, however, the fault
current will have flowed for a finite, if small, time, and much heat energy will have been
released. Also the severe mechanical forces referred to above will already have occurred and
will have subjected all conductors to intense mechanical stress.


In order td design electrical equipment to withstand the expected thermal and mechanical
stresses, and to engineer the protective system to operate decisively and quickly, it must be
possible to calculate the maximum ~fault current to be expected anywhere in the system
under the worst possible conditions.

Phase-to-phase and phase-to-earth faults may be nietal-to-metal, but more probably they will
be arcing faults where the arc itself has some resistance which will reduce the flow of fault
current. However, for calculation purposes the worst condition is considered, and short-
circuits are assumed to be ‘bolted’ - that is, it is assumed that all three conductors are
firmly bolted together and that the fault itself has zero impedance.

In order to understand the fault conditions in an a.c. network, it will be helpful to consider
what happens in the simpler d.c. case.

2.2.1 D.C. Case

Referring to Figure 2.1, suppose the full-load current I of a d.c. generator is produced with
an external load resistance R. If E is the emf and r the internal resistance of the generator,
then the internal voltage drop is /.r an~d the terminal voltage (that is, rated voltage) V.of the
generator at full load is E - /.r. Suppose the internal drop is, say, 20% of the open-circuit
voltage E (assuming that there is no automatic voltage regulation).


If now the external load R is decreased so that the current doubles to 2/, then the internal
drop increases to 2/x, or 40%. If R is further reduced so that the current is trebled to 3/,
then the internal drop increases to 3/x, or 60%. Similarly a current of 4/ will cause an
internal drop of 80%, and 5/ would produce 100%.

A 100% internal drop means that the whole emf is used in overcoming the internal drop,
and there is no voltage left between the terminals - that is, they are effectively at short-
circuit. Put the other way, a dead short-circuit across the generator produces B current five
times (1 + 20% or 1 + 0.2) full-load current. The generator is then said to have an internal
resistance of 20%. This is an alternative way of expressing it instead of in ohms.

2.2.2 A.C. Case

The same argument applies to the ax. generator shown in Figure 2.2, except that, instead of
external and internal resistance, there is now impedance. However, in all ax. generators the
internal impedance is almost wholly reactive, and it is therefore customary to talk of a
generator’s ‘reactance’ x and to ignore the resistance. It is, like the resistance in the d.c. case;
expressed as a percentage. Therefore a generator with a reactance of 20% will deliver 1 + 0.20,
or five times, full-load current on short-circuit. This method of using percentages rather than
ohms avoids having to con,sider the size (kVA) or voitage of the particular generator. The
above applies whatever its size or voltage.


An a.c. genes-ator has in fact a varying reactance, which increases as the fault proceeds, due
to its complicated magnetic behaviour. For fault calculation purposes however the lowest
reactance at the beginning of the fault is always taken; it is the ‘subtransient’ reactance and
is typically between 8% and 20% on most generators. This is the most severe condition.

A similar argument applies to transformers where the reactances are typically between 4%
and 10%. A transformer reactance however is fixed and does not vary as the fault proceeds.

5000kVA BASE

u 15%
‘5% 15%
-15% 2 7%%)F

P =

50% t
7= .5;!+,
Q Q b lo.575 at0 I






Figure 2.3(a) shows a typical, but simple, network comprising two generators, two trans-
formers and an WV and LV distribution system. A fault at a point P on one of the HV
feeders would, if all HV breakers were closed, be fed by both generators in parallel. A fault
at a point Q on the LV system would be fed by both generators (as before), but they would
be in series with one transformer- if the LV section breaker web-e open, or with both trans-
formers in parallel if it were closed.

The exact calculation should, strictly, also take into account the resistances of thegeneratol-s
and transformers as well as the impedances of the connecting cables, but for a rough calcu-
lation with platform-sized lengths of cable these can be disregat-ded.

Figures 2.3(b) and 2.3(c) show the reactance equivalent of each of the elements of the
network, with the percentage reactance placed against each. Since the size of each generator
is 5 OOOkVA the impedance of all other elements, such as that of the transformers, mustbe
raised to th~is ‘base’. So, though the transformers are each rated 5% at 5OOkVA, they are
entered as 50% at 5 OOOkVA, giving the same short-circuit current. The ‘adjusted’ reactances
are shown in Figures 2.3(b) and (c) in red. it should be noted that any figure, such as
100 OOOkVA, may be chosen as a base; it makes no difference to the result. Choosing as a
base the kVA of the largest generator is merely a convenience. ,(For onshore grid network
calculations 100 OOOkVA base is usually chosen. In those cases generator and transformer
resistances and cable impedances cannot be ignored.)

The adjusted reactances are then resolved by ordinary series-parallel network methods until
they become a single reactance. Thus in Figure 2.3(b) the reactance up to the point P with
both generators connected is two 15% in parallel, equivalent to one 7%%. To the chosen
base of 5 OOOkVA the fault level at P is:

5 006
- = 67 OOOkVA, or 67MVA

(Note. For the purpose of calculation, percentages are expressed as ‘per unit’. Thus
7%% = 0.075 p.u.)

For point Q with one transformerconnected there is a further series reactance of 50% to
add, making 57%% in all. The fauhlevel at Q would be:.

m ‘= 8 7OOkVA, or 8.7MVA

For point Q with two transformers in parallel (Figure 2.3(c)) there is a further series reac-
tance of 25% to add (derived from the two 50% in parallel), making 32%% in all. The fault
level at Q would then be:

= 1,5 4OOkVA, or 15.4MVA

If the generators had been of different sizes - say 5 OOOkVA and 2 5OOkVA, each with
reactance 15% - the larger would have been chosen as ‘base’ and the smaller raised to it -
that is, call it 5 OOOkVA at 30%, and proceed as before:

This calculation, though much simplified, illustrates the basic method of making fault calcu-
lations. It shows too the advantage of regarding all reactances as percentages; the actual
voltage levels have not come into the calculation. It also illustrates the considerable reducing
effect of transformers on a system fault level.

The fault levels so calculated would apply respectively to the whole HV system and the whole
LV system, and they are usually marked in MVA on drawings. The switchgear at each level
must be capable of breaking the currents appropriate to those levels, and all conductors,
busbars, cables, etc, must be able to withstand the thermal and mechanical stresses induced
by those currents. Armed with the result of his fault calculations, the designer will specify
exactly what the various items of equipment are required to withstand.

Fault levels calculated as shown are expressed in MVA, which is usually sufficient for most
purposes, but if act+1 fault currents are needed, the MVA is converted to current (kA) by
dividing by,/3 times the voltage (kV). Thus:

67MVAat 6.6kV = -6.g3 = 5.9kA =5 900A

8.7MVA at 440V = 0.4;&3 = 11.4kA=ll 400A

15.4MVA at 440V = ,,&$, = 20.2kA = 20 200A

(Note that MVA +43kV gives the current in kA.)

These are all runs symmetrical currents.

From the above calculations two points should be noted:

- A single transformer between the source of supply and an LV switchboard

greatly ‘cushions’ the MVA fault level beyond it and reduces it drastically. In
this case an HV fault level of 67MVA is reduced by the transformer to 8.7MVA.

- Notwithstanding the great reduction of MVA by a transformer, the actual

fault current on the LV side is actually increased - in this case from 5 900A
to 11 400A. This is because the smaller MVA is obtained fror,l a still smaller


Whereas apple designs of switchgear exist to handle the fault levels to be found on the
largest offshore and onshore high-voltage systems, this is not the case on the low-voltage
boards. As will be seen from the above calculation, the LV fault currents are in general
higher than the HV - in some cases much higher indeed. On many offshore installations the
LV fault level with the LV section breaker closed (that is, with both transformers feeding in
parallel) exceeds the breaking capacity of the largest LV circuit-breaker available.’

It is therefore ~necessary to ensure that the switchgear fitted is not subjected to such a
possible fault. From the above calculation it can also be seen that the LV fault level with
one transformer feeding is about half of that with two. Therefore it is arranged in such cases
that the two transformers should no? be allowed to feed in parallel. The LV section breaker
is normally kept open when both transformers are feeding (the normal condition). The
section brea!;er may only be closed if one or other transformer supply breaker is open. This
is known as the ‘two-out-of-thl-ee’ method. Any two LV breakers out of the three (two
incomers and one section) may be closed at any one time, but not all three. Interlocks
.~p@vetit this.

However, this arrangement may be temporarily ‘cheated’ and the vulnerability accepted
when changing over from one transformer to the other. FOI- a short period the operation of
the interlock is delayed to avoid interrupting supplies to the board. If one of the three
breakers is not opened within that short time (typically 30 seconds) the section breaker will
trip automatically.


The fault currents as derived from the above calculation. usins the oercentage reactances of
the various items of plant, are all ‘rms symmetrical’.

rms Symmetrical


The actual currents which occur in the early part of a fault however are generally asymmetrical,
giving a greater heating rate. Moreover the highest mechanical forces will occur with the first
asymmetrical peak of current, as shown in Figure 2.4. This can, with 100% asymmetry, be
up to 2.55 times the rms symmetrical value, so the 67MVA at 6.6kV referred to above,
equivalent to 5.9kA rms symmetrical, can rise to 15kA asymmetrical peak. It is this latter
figure which determines the mechanical strength of.the busbars and other equipment.

When a fault current is quoted in kA it is always wise to add the words ‘rms symmetrical’ if
that is what is meant. This avoids confusion with ‘kA peak asymmetrical’ which is also often
quoted in addition. Fault levels quoted in MVA are always rms symmetrical.


In, some cases the actual fault current, even with a bolted short-circuit, may be limited by
the operation of the protection device itself - notably in the case of a current-limiting fuse
which can interrupt the current before it reaches its first peak (see Chapter 3, para. 3.3). In
that case the current never attains its calculated level. However for the purposes of calcula-
tion it is assumed that no such limiting effect occurs and that the current will reach its
calculated value. This value is called the ‘prospective fault current’, even though, in certain
given systems, the fault current will not reach that level.

Fuses are given the credit in their ratings for interrupting the full prospective fault current
notwithstanding that they do so by preventing it ever happening.


A factor that may have a significant effect on the fault level of a system which includes an .~
appreciable proportion of motor loads is that known as ‘motor contribution’. This refers to
short-circuit current which is generated for a very brief period by a short-circuited induction
motor as a result of magnetising currents still circulating in the rotor. (Synchronous motors
also generate short-circuit current but are not likely to be encountered in Shell installations.)
The magnitudes of such currents are not easy to determine with any accuracy, but they are
commonly roughly estimated on the basis of a percentage reactance in the motor assumed
to be about 30%. Thus a motor that presents a full load of 1 MVA is calculated to contribute
l/O.3 = 3.3MVA to the fault ievel of the circuit which supplies it, and it will contribute to
the fault level at any point in the system as if it were a IMVA generator with a subtransient
reactance of 30%. This should be added to the calculated system fault level.


A short-circuit between two phases results in a lower fault cukent than does a symmetrical
short-circuit, because it is driven through the impedance of two phases by the line voltage
which is only ,/3 times the phase .voltage. This condition therefore requires no further
consideration here.


Earth-fault currents, especially when iimited by earthing resistors,are dealt with in Chapter 5.
Havir;g only 1143 of the system voltage behind them, they are in general lower than the
short-circuit fault currents and will not therefore influence the fault level calculations.




Overcurrent protection is related primarily to heating effects in, and in some circumstances
to electromechanical forces on, electrical conductors and may cover both fault and overload

Overcurrent protection methods may be divided into three categories according to the type
of device used:.

- Overcurrent relay tripping a circuit-breaker or contactor (but see Chapter 1,

para. 1.3 for the limitations of contactors).

- Direct tripping device (‘release’) on a circuit-breaker or Contactor.

- Fuse.

A distinction should be made between~ the terms ‘overcurrent’ and ‘overload’. Whereas the
mechanical overloading of a machine will certainly cause overcurrents, OvercUrrents can
occur from causes other than overloading - for example, stalling or single-phasing of a
motor, short-circuits Or earth faults, none of which is an overload.

The term ‘overload’ should be reserved for mechanical loading, and the word ‘overcurrent’
should be used in Its literal sense. All the devicesdescribed in thischapterare trueovercurrent


Overcurrent devices, though all depend on an excess of current to operate them, are of
several different forms; these are described below. The official abbreviation (BS 3939) for
each type is also given.

3.2.1 instantaneous Overcurrent (OC)

An instantaneous overcurrent relay is shown pictorially in Figure 3.1. It consists of a simple

iron armature attracted by a coil carrying the current from a line current transformer and
restored to its rest position either by gravity or by a control spring. When the current in the
coil just exceeds a certain preset value, the pull on’the armature overcomes the spring or
gravity and causes it to close. In so doing it operates auxiliary contacts which initiate a
tripping circuit or other desired function.

Though termed ‘instantaneous’ this type of relay nevertheless requires ‘a small but finite
time to operate; this /s usually taken to be a maximum of 0.2 seconds, but it is often much
less. The current/time characteristic is thus a ‘square’ one as indicated in Figure 3.1 (c). The
value of the current required to operate the relay is set by the screw adjustment at the top.

In a single-phas6 system a current transformer in one line,is connected to the relay coil (see
Figure 3.1 (a)). In a 3-phase system a current transformer in each phase is connected to one
of three relay coils (see Figure 3.1(b)). The three relay elements may be enclosed in one
case or in separate cases. However, in a 3-phase, 3-wire, system any overcurrent in one line



OC R&v

-T--Fe s Trip

Current Setting



must be accompanied by an overcurrent in one or both of the return lines. Therefore, to

achieve complete overcurrent protection in a 3-wire system, it is only necessary to provide
overcurrent relay elements in two of the three phases (see also Figure 5.2).

3.2.2 Inverse Time Overcurrent (OCIT) ’

An inverse-time overcurrent relays is shown pictorially in Figure 3.2. It has an ‘induction-

type’ movement similar to that of a household meter. It consists of a rotating aluminium
disc driven by a shaded-pole magnet element which receives the driving current from the CT
in the circuit to be monitored. As in a household meter, the disc also revolves between the
poles of an eddy-current brake magnet; it is restrained by a light pre-tensioned control

The relay is used with current transformers in single-phase or 3-phase systems as described
for simple overcurrent type and as shown in Figures 3.2(a) and (b).

When normal current flows from the CT a driving torque is applied to the disc, but it is
prevented from rotating by the pre-tensioned spring. If the current exceeds a certain preset
value the disc begins to move and is driven, against the drag of the brake, right round until a
contact on the spindle touches a fixed contact. The greater the excess of current above this

Time Multiplier

I OCIT Relay

: 6-
-. Trip
Circuit -e -id Hieh Set
El&&t (as Fig 3.1)
be added


. . . CU~~~~t Setting


OCIT Relay


Operating Current



value, the greater the drive torque and the faster the disc tries to rotate. But the drag of the
eddy-current brake also increases with the speed of rotation, and itsslowingeffect isgreatest at
the highest currents.

The combined effect of this is to produce a time/current characteristic as shown in Figure

3.2(c). For currents less than a certain .min,imum the disc does not move at all. For currents
in excess of the ‘just move’ minimum the disc moves, and the operating time becomes
shorter with increasing current - that is, it has an ‘inverse-time’ characteristic. This is very
important when dealing with ‘discrimination’ (Chapter 4).

Two setting adjustments are possible with this relay: current and time. Current adjustments
are made by fixed taps on the driving coil. They are +sually set by moving a peg between a
number of holes on the front of~the relay face. Typically the range is from 50% to 200% of
the normal operating current (1 A or 5A depending on the CT used). The time adjustment is
made by moving the ‘fixed’ contact so as to increase or decrease the travel of the disc before
the contacts touch. The relay is fitted either with a time-scale marked in seconds, or more
usually with a ‘time multiplier’ adjustment, which is used in conjunction with curvessupplied
with the relay.

Relays could be custom-made to operate with any given CT and any given circuit data, but
in practice relays are manufactured t.o certain standard conditions, and adjustments are
provided to match this standard relay into a wide variety of circuit arrangements. This
results in a fairly complicated setting procedure which is described in detail below.

Movable Plug I I

OUT Relay
Operating Coil


3.2.3 Setting of an Inverse Time Overcurrent Relay

To understand the setting of the current and timing adjustments and the interaction between
them, consider the particular circuit shown in Figure 3.3. A standard inverse-timeovercurrent
relay designed to work on a nominal secondary current of IA is fed from a 400/1A current
transformer. Suppose the normal full-load line current is SOOA, then at full load the CT
secondary current will be 1.25A - higher than the relay’s designed working current. The
relay must. therefore be slightly desensitised. The operating coil has several taps, and a tap
which reduces the effective turns by 20% is chosen by inserting the plug in the 125% socket.
Thus 1.25A (the actual or ‘effective’ current) through 20% less turns has the same effect
(ampere-turns) as the designed IA through the full turns; that is to say, with the 1.25A
coming in, the relay will operate as designed for a IA input, and it will have the same
designed characteristic time/current curve. Thus the current plug setting compensates for
any deviation between the CT rated primary and the actual full-load current. If there is no
deviation the plug is set at 100%.

For cases where the CT primary current rating is greater than the full-load current, the relay
must be made more sensitive, and the tappings are extended below the nominal 1A (100%)
so as to increase the effective turns (an ‘auto-transformer’ effect). Hence there are 75% and
50% plug positions. It should always be remembered that settings below 100% make the
relay more sensitive, and settings above 100% make it less sensitive.

The plug settings 50% to 200% are seven discrete sockets, and no intermediate position is
possible. If a calculated setting (e.g. 1 loo/)D comes between two positions, the next higher
setting should be used.

Some relays, instead of having plugs marked in current percentages as already described, are
marked in CT secondary amperes --for example 0.5, 0.75, 1.0, 1.25, 1.5, 1.75 and 2.OA
instead of 50%, 75%, lOO%, 125%, 150%, 175% and 200% - but the purpose is the same.

Other relays are designed for use with 5A current transformer secondaries. If the plugs are
marked in current amperes, the markings would be 2.5, 3.75, 5.0, 6.25, 7.5, 8.75 and
1 O.OA.

If the calculated line fault current is, say, 5 OOOA - that is, IO times the normal full-load
cur-rent in the case of Figure 3.3 - the CT secondary will then give 12.5A. Note that this is
10 times the plug setting, not 10 times the nominal 1 A (= 100% setting). Consequently the
horizontal axis of the characteristic (set Figure 3.4) is scaled in ‘Current (Multiples of Plug
Setting 50% - 200%)‘, not simply in multiples of full-load current.

The purp~ose of this exercise is to determine what current and timing settings should be put
on the relay to achieve any desired time delay in its operation when subjected to a given
short-circuit current. (The desired time delay will come out of the discrimination caicula-
tions which~determine the delays required at various points of the network -see Chapter4.)
It therefore only remains, having determined the current operating plug setting for the
calculated short-circuit current at the point where the relay is installed, to choose the time
multiplier which will give the desired time delay.

aperating Time

(= Time Multiplier


_ Operating Line

1 30
Current$lultiples40f Plut S&:50% - 2:%)


Figure 3.4 is a set of time/current characteristic curves as provided with a typical OCIJ
relay. Both scales are logarithmic.

For the circuit of Figure 3.3 it has already been determined that a plug setting of 125% is
required, and that a short-circuit current of 5 OOOA will mean a current 10 times that of the
plug setting. If in Figure 3.4 a vertical line is drawn through the current multiplier point ‘IO’
on the horizontal axis, this is the ‘relay operating line’ for that short-circuit current.

Suppose the d~iscrimination calculation requires a time delay of 1.35 seconds at this point in
the network. Draw a horizontal line through the point 1.35s on the ‘Operating Time’ (that is
vertical) axis. Let it cut the vertical relay operating line at point P.

It will be seen from Figure 3.4 that P lies b~etween the Time Multiplier curves 0.4 and 0.5.
By interpolation it would be 0.44. Although such a setting would be possible, it is usual to

choose the next higher, namely 0.5 (point Q). This setting will actually give a time delay of
1.5 seconds at 5 OOOA - very slightly higher than the desired 1.35, which errs on the safe

Similarly, if a calculated current plug-setting came between two sockets, the next higher
plug-setting should be used. This too errs on the safe side in making the relay marginally less
sensitive, needing slightly more time to operate for a given fault current.

It is required to determine the current and timing settings on an OCIT relay to give a 1.35s
delay with a short-circuit current of 5 OOOA. Full-load current is 450A and the CT-ratio is
400/1A. (Note that in this example the full-load current is slightly different from before in
order to show how an ‘in-between’ figure should be interpolated.)

The setting sequence is shown in Figure 3.5 and uses the curves of Figure 3.4.

Full load
current of
Rated primary
current of

LuRRENT Short circuit
I 0, -9..__ --&:-^ /Y” =
(as a fraction) I (i.e. calculated
(If intermediate. I short circuit
next higherj current at
point of fault)


Effqctive Current OPERATING LINE, POINT ‘P’

Plugsetting=$$x 100=112.5%(use125%)
400/l ,,,=45OA ~ l&.jOOOA
Effective current = 400 x 1.25 = 500A

Relay operating line = E = lb

Desired time delay = 1.35s

I---- R Time multiplier setting at crossover (‘P’)

= 0.44 bf interpolation.
Use next higher 0.5 (‘Q’)
Desired Time Delay
1.35s Reading across, actual time delay achieved = 1.5s


It should be noted that the plug setting in this case comes out at 1 .125, or 112.5%. Th,ere is
no such plug position, so the next higher one, 125%, is chosen.

From here, the plug having been inserted in the 125% position, carry on as before. The
5 OOOA short-circuit current represents 10 times the chosen plug setting. Draw a vertical line
through ‘10’ on the Current Multiplier axis of Figure 3.4 and let itcut the horizontal through
the desired delay time of 1.35 seconds. The crossing point Pfalls between theTime Multiplier
curves 0.4 and 0.5, so the larger is chosen. This will give an actual time delay (horizontal
through Q) of 1.5 seconds, slightly longer than the 1.35 desired.

3.2.4 Combined inverse Time Overcurrent and High Set Instantaneous Relay (OCIT/OC)

An inverse-time relay may be equipped with an additional instantaneous element in the

same casing but operating at a ‘high-set’ current value. This gives it the feature ofa combined
‘inverse-time and high-set instantaneous’ relay, the instantaneous feature overriding the time
delay only on the most severe faults. An eiample of this additional feature is shown dotted
in the drawing of Figure 3.2.

This arrangement is particularly desirable where overcurrent protection is installed,near the

generator end of a network. It is at this end that discrimination requires the longest delays,
and a purely inverse-time relay would allow a severe short-circuit to persist until eventually
cleared. .An overriding high-set instantaneous overcurrent relay, fed through the same CTs
and actuating the same trip circuit, would clear such currents very quickly. It would however
operate only on severe faults and would take no part in fault currents below its own high
setting. This feature is further investigated in para. 3.2.6, ‘Busbar Protection’.

Usually a 3-element OCIT relay (one per phase) would be combined with two instantaneous
high-set elements in two phases only’- all in a single case (30CIT/ZOC).

3.2.5 Inverse and Definite Minimum Time (OCIDMT)

An inverse and definite minimum time overcurrent relay is shown pictorially in Figure 3.6.
The current transformer arrangements with single-phase or 3-phase systems are similar to
those for the simple overcurrent relay and are shown in Figures 3.6(a) and (b).

This relay is simply a variation of the inverse-time type shown in Figure 3.2, but here the
characteristic, instead of tending towards zero time for the highest fault currents, now tends
towards a definite and finite small value, as in Figure 3.6(c). This is built into the relay and
cannot be adjusted.

The relay is similar in construction to the normal inverse-time type shown in Figure 3.2.

The purpose of this variation is to render the relay settings more accurate. All characteristic
curves are subject to tolerance, and the separation of the sloping curves of Figure 3.2 at the
high-current end for different relays has to be enough to allow for such tolerances. Therefore
tripping delays would need to be longer than would be necessary with more accurate curves.
The definite minimum time feature at the highest currents, making the curves horizontal at
those currents, enables greater accuracy (that is, smaller tolerance) to be achieved, resulting
in less separation of the curves and consequently shorter tripping times.

An OCIDMT relay may be combined with instantaneous high-set overcurrent elements as

described for an OCIT relay in para. 3.24. It is shown in dotted outline in Figure 3.6.

Time Multiplier


Eldment (as Fig
may be added




IIlinimum +-‘----
Time :---..
* Minimum
Operating Current



3.2.6 Busbar Protection

A fault on a high-voltage busbar will produce severe overcurrents which should cause the
generator overcurrent relays - the only protection that the generator has upstream of the
busbars - to operate and clear it. However the generator inverse-time relays, being the last
on the discrimination ‘ladder: will have a comparatively long operating time, despite the
heavy fault current; it may well be of the order of 2 to 3 seconds in some networks.

During this time the short-circuit current will have passed from its initial large subtransient
value, thl-ough the transient value and probably well towards its final synchronous value
which, as already shown, could well be even less than normal full-load. In this state there is
no overcurrent at all, and the generator protection relays would probably not operate; there
would then be no protection against a busbar fault which, by its very nature, could be
highly damaging. There are several ways in which this situation can be dealt with, and these
are descl-ibed below.

(4 Instantaneous

Instantaneous protection is b y the addition of simple high-set instantaneous overcurrent

elements to the OCIT relay. With heavy faults these elements would cause an immediate trip
before the fault current had started to run down from its subtransient level and before the

inverse-time element had worked off its long d&y. This was briefly mentioned in para. 3.2.4
but not in the context of a busbar fault.

(b) Voltage Restrained

This type of protection is by using ‘voltage-restrained’ inverse-time overcurrent relays for

the generator.

Operating T:-- ,O





1 30
Current&ult?ples4af PluzSet8tinTSO% - 22d%)


The operation of this type of relay is explained in the manual ‘Electrical Central Devices’,
Chapter 2. It is a normal overcurrent inverse-time relay with an additional voltage-sensitive
element which, under normal voltage conditions, produces a counter-torque and so restrains
the drive of the overcurrent element. When voltage falls this restraint is lifted, allowing the
overcurrent element to operate more rapidly - that is, it becomes more sensitive as voltage
falls. The black curve in Figure 3.7 shows the normal (restrained) time/current characteristic
similar to Figure 3.4 with time multiplier 0.3, and the red curve shows the change of charac-
teristic (with the same time multiplier) as the voltage falls and the restraint is completely
removed. For example, an overcurrent of three times the plug setting will cause the relay to
operate in 1.8 seconds with normal voltage (restrained), but it will operate in 0.5 seconds
under low voltage with the restraint removed. This allows the generator breaker to clear a
busbar fault in a far shorter time than it would with normal OCIT protection (if indeed it
would at all).

The change of characteristic as the voltage falls is a gradual process, and the black and red
curves are the two limits - that is with full restraint and with no restraint. For partial fall of
voltage there would be a whole family of curves in between.

There is another type of voltage-restrained overcurrent relay where the change of characteristic
is not gradual but is abrupt. This type of relay is not strictly restrained, btit its sensitivity is
increased by a discrete step when the voltage falls to a specific percentage, either 60% or
30% of the nominal value. This is achieved not by use of a voltage element producing a
counter-torque as with the first type but by altering the main driving torque produced by
the current element. The sensitivity change is initiated by a voltage-sensitive changeover
relay which alters the pole-shading on the current element.

Figure 3.7 can be regal-ded as illustrating this type also if the black curve is the normal
characteristic and the red cut-ve the more sensitive characteristic when the voltage has fallen
and the changeover relay has operated. Its effect on speeding the operation of the relay
under short-circuit conditions is just the same as described above for the truly voltage-
restrained type of relay.

(c) Flame Leakage

A different way of dealing with a busbar fault is by the method known as ‘Frame Leakage’
protection, which is indicated ip. Figure 3.8.


It is assumed that a busbar arcing fault will rapidly go to earth on the switchboard frame.
The frame is lightly insulated from the deck and is connected to the true earth through a
current transformer-. If the switchboard consists of several sectionsasshown,each is insulated
both from the deck 2nd from its neighbours. Each section is separately earthed through a

If there is a busbar fault involving the frame, the frame becomes live and current flows to
earth through whichever CT is invqlved. This frame leakage current is used to provide an
instantaneous trip on all generators and section breakers feeding into that section.

Care must be taken not to allow any conducting material to lean against theswitchboard, as
this would shunt the CT connection.

(4 Zone Protection

This method provides each incomer, each feeder and each section breaker on theswitchboard
with a set of current transformers and applies differential protection (see Chapter 6) to the
entire group. Under heaithy conditions all the currents entering the switchboard from the
incomers and section breakers are exactly balanced by those leaving it through the feeders.
Any busbar fault upsets this balance and causes an instantaneous trip.

This method is known as ‘Zone Protection’. It is by its nature very costly and difficult to
set up. It is only likely to be found in major shore netw~orks.

In most normal installations busbar faults are considered so rare that special protection is
not provided. This is the case with Shell offshore installations, but busbar protection is
provided at certain onshore plants.

3.2.7 Relays - General

Most protective relays are fitted with flags which indicate when they have operated. They
show the operator, for example, which of the protective systems may have caused a turbo-
generator to have tripped out. Such relays are themselves narmally self-resetting - that is,
,they revert to their normal state as soon as the fault has been removed. This may occur
either because the circuit-breaker has tripped, so disconnecting the fault, or because the
fault itself has disappeared. The flag however remains showing until it has been reset by

Ian some protective systems, particularly for generators and transformers, all the protective
relays trip the breaker through an intervening hand-reset trip, or ‘lock-out’, relay (TH). It
too has a flag, but this relay, having once operated, does not reset itself automatically and so
prevents the breaker being reclosed until the relay has been deliberately reset by hand. This
prevents accidental reclosure onto a fault, and the breaker remains locked out until cleared
by the operatorresetting the lock-out relay.

Whenever an item of plant has tripped because one of the protective systems has operated, it
is most important that the operator should not reset the relay flags until he has carefully
noted down which. flags are showing. If this is not done, all evidence of the cause of the
malfunction will be lost. The lock-out relay must on no account be reset until it is safe to
operate the plant again.,

3.2.8 Electronic Relays

Those relays which have so far been described are of the ‘electromagnetic’ type, where an
electromagnet provides the driving force to a mechanical system of moving armature or
rotating disc and mechanical contacts.

Many of these relays are now being superseded on offshore, and numerous onshore, installa-
tions by electronic types which are entirely static except for their final output contacts.
Electronic circuits carry out the detection, processing and timing; only the output circuit is
passed through normal electromagnetic auxiliary contacts to the external trip circuits. This
also isolates the trip circuits proper from the electronics.

Though using different methods, electronic relays reproduce similar characteristics td’those
of the electromechanical types, and they have similar adjustments such as for current and.
time setting. Their use does not affect the principles of protection described in Chapter 1.

An electronic counterpart exists for almost every relay described in this and succeeding
chapters. To illustrate the principle of operation, a single-phase, electronic inverse-time and
instantaneous overcurrent relay (OClTjOC) is described here and shown in Figure 3.9.

The input from the line current transformer is fed through a small adapting transformer to a
low-pass filter Rl-Cl which suppresses transient voltage surgesA voltage proportional to
the input current is developed across the current-setting potentiometer R2. This voltage is
applied to the bridge rectifier.



Filter & Current
setting M

I nstantaneous Trip



The d.c. output voltage, which is proportional to the line current, is used to charge the
capacitor C2 through the potentiometer R5. Tbe setting of this potentiometer determines
the rate at which the voltage across C2 increases and hence the timing of the inverse-time
operating characteristic of the relay. When the’voltage across C2 reaches a predetermined
value, the detector circuit operates to switch the electromechanical relay RLA through the
output amplifier and power transistor T2.

instantaneous operation is obtained by applying the output voltage of the bridge rectifier
directly to the input of the amplifier through R4. Thus; for higher values of fault current,
the inverse-time delay circuit is bypassed.

The power supply for the solid-state circuits is applied through D3 and R6. It is stabilised by
zener diode DZl, and spike protection is affor-ded by R7 and C3. The diode D3 protects
against reversed polarity of the d.c. power supply.

By suitable choice of elements the electronic relay current/time characteristic can be made
to reproduce exactly that of the equivalent electromagnetic type. Having virtually no
moving parts, they are, in general, more robust, smaller and lighter. Current and time
settings in this case are applied through simple variable resistors.


3.3.1 The High Rupturing Capacity (HRC) Fuse

A fuse consists essentially of a length of metallic wire or strip carrying the cil-cuit current
which, if that current exceeds a certain stated value for a certain minimum time, will melt
and break the path of the current in that circuit. It has both a normal current rating corres-
ponding to its service current and a breaking current rating corresponding to the maximum
fault current of that part of the system in which it will be used.

Originally fuses consisted merely of a length of suitable wire stretched between the terminals
of, a holder, the holder being designed to plug into permanent fixed sockets. These had the
disadvantages of having much exposed live metal, and the melting open wire tended, under
some conditions, to give rise to severe arcing and risk of fire. The wire also tended to corrode
and weaken with the passage of time.

The open-wire fuse is no .longer used, having been superseded by the cartridge type. That
used on offshore and onshore installations consists of an outer ceramic tube in which there
is a silver fusible element completely surrounded by quartz powder, as.shown on the left of
Figure 3110.

Silver Wire Element Quartz Powder Filling

\ I
C‘eramic Body

Before Fusing After Fusing


If sufficient current flows through the silver element for sufficient time, the element melts
and vaporise:; it reacts chemically with the quartz, under the heat of the arc, to produce a
block of highly insulating material in the path of the arc, as shown on the right of Figure
3.10. This rapidly suppresses the arc and, unless the current is much in excess of the fuse
rating, it will break the fault current within a matter of milliseconds. Such fuses can break
very large prospective fault currents by simply preventing those currents from ever building
up. The fuses are consequently known as ‘High Rupturing Capacity’ (HRC) type.


Load Current jr\

Fault Current
Befpre Fault ; \

Tqtal Clearing Time



Figure 3.11 (a) shows a fault current passing through an open-wire fuse. The current may
continue for several cycles of arcing before it is eventually broken at a current zero.

When a moderate fault current passes through an HRC fuse the melting time will be com-
paratively slow, and the current may continue for several cycles before it is broken - in
fact the HRC fuse will behave just like the open-wire type in Figure 3.11 (a). If however the
fault current is very large, the melting time will be less than one-quarter of a cycle. The
ensuing arcing time is so short that the current is broken even before it reaches its first peak,
as shown in Figure 3.11 (b). Such a fuse is said to exhibit ‘cut-off’. If the current,wave is
wholly asymmetrical the first peak is not reached until half a cycle has elapsed, and cut-off
may occur if the melting time is somewhat longer - i.e. less than half a cycle.

If it had not been for this cut-off, the fault current would have risen to its full-fault peak
(called the ‘prospective current’ peak) before reaching its first zero. The fuse, by cutting off,
has protected the whole system from the effects of this severe peak. It is therefore given

credit for having interrupted the full prospective current, even though in fact the current
may never reach it because of cut-off. The fault rating of an HRC fuse is consequently very
h~igh for its size. In Chapter 4, ‘Discrimination ’ , it is shown how such fuses are used to back
up switchgear of lower fault rating capacity.

There is often confusion between the ‘normal’ and ‘breaking’ cxrent ratings of a fuse. The
normal rating is matched to the load and is the maximum value of current which the fuse
can carry continuously without melting or deteriorating. The breakingratingisthe maximum
prospective current which the fuse can safely interrupt at its rated voltage; it is usually
quoted in kiloamperes (kA) rms symmetrical and is related to the system fault level.

The energy needed~to melt Xfuse~ is the product of the rate of heat generatio+(in watts) due
to the fault the resistance of the element and of the total time during which such
heat is being generated. It is /‘R x t, where I is the rms current, R the resistance of the
element and t the total time. Since R is virtually fixed for.atiy given size of fuse, the energy
released is proportional to I* t.

A specific fuse element requires a given Izt to melt it. Therefore when I is very large, t
(the melting time) will be very small, as indicated in Figure 3.11(b):/‘t is often referred
to as the ‘let through’ energy.

Time (seconds)

Minihun Current (amps)



If the melting (or ‘pre-arcing’) time t is plotted against I (usually on log paper),~the curve of
Figure 3.12 is produced. Most of this is the familiar inverse-time curve which many relays
also have. There is of course a minimum current which will never melt the fuse however long
it is applied, but above this lower limit the fusin, - time varies inversely as the current. The
upper limit is set by the ability of a given make and size of fuse to absorbs the /2t energy and
to handle the mechanical forces involved.

As the fault current becomes higher, the melting (or pre-arcing) time becomes shorter until
the point is reached where it is less than one-quarter of a cycle (0.004 seconds at 60Hz), and
cut-off begins. From this point on the characteristic changes and becomes almost linear, as
shown on the extreme right of Figure 3.12 (this is because ‘rms’ no longer has any meaning).
With a fully asymmetrical current wave, cut-off may occur up to one-half of a cycle (0.008
seconds at 60Hz) after the onset of the fault.

3.3.2 Fusing Element

Figure 3.10 shows a silver wire as the fusing element. This is normal with small fuses, but
for larger ones a silver (or sometimes copper).strip is often used, as shown in Figure 3.13.

Copper~End Caps

/ Ceramic Tube

uartz Powder Filling

Constricted Silver


The strip has a number of constrictions which form hot spots and assist rapid melting under
short-circuit conditions. For the heaviest currents a number of such strips may be connected
in parallel within the common housing, or many separate fuse-links may be permanently
bonded in parallel to form a single multiple link.

3.3.3 Fuse Mountings

Within an equipment, especially high-voltage assemblies, fuses are often mounted without
individual enclosure on pillar insulators or directly on busbars. Reliance for safety is placed
on the metal enclosure of the HV compartment which houses them. Interlocks prevent the
compartment being opened until the circuit has been made safe.

On low-voltage distribution boards fuses are housed in a fuse assembly such as the typical
one shown in Figure 3.14. The replaceable ceramic cartridge with its metal terminal caps is
known as the ‘fuse-link’ and is held in an insulated ‘fuse-carrier’ which completely shrouds
all live metal. The carrier is supported on an insulated ‘fuse base’, where it is firmly fixed by
various mechanical means, amongst them tongue contacts, butt contacts held by insulated
screw pressure, or wedge contacts pressed in by insulated screws. A tongue-contact type is
shown in Figure 3.14.

1 Fuse Carrier


3.3.4 Fusing Factor

A fuse has a ‘normal current rating’, which is the current which it can carry continuously
without melting or deteriorating and without altering its characteristic. The current which,
under specified ambient temperature conditions, will just cause the fuse to melt after a
prolonged time (usually taken to be four hours) is termed the ‘minimum fusing current’.

The ratio minimum fusing current.

IS called the ‘fusing factor’ of that fuse.
normal rated current
Fuses are manufactured to different fusing factors for various applications. They are given
classification accordingly as follows:

Class Q2 Fusing factor greater than 2.0

Class Ql Fusing factor between 1.5 and 2.0
Class P Fusing factor between 1.2 and 1.5

In Figure 3.12 the time/current characteristic of the IOOA (normal rating) fuse is shown to
become almost vertical after IO seconds, at which point the minimum fusing current is
125A. After four hours it will still be only 125A, as the curve is vertical. The fusing factor
in this case is 125/100 or 1’.25, and its Class is therefore P.

Fuses which protect motor circuits often have a dual rating. Against steady overloads a
fusing factor of about 1.2 (Class P) is usual, but, to allow starting currents to flow, a short-
time overcurrent rating of the order of 1.6 times normal is given (this is inherent in their
time/current character&tics). Such a fuse with a normal 1 OOA rating would be termed
‘1 OOMI 60’. A marking of this type indicates that the fuse is specifically for motor protection.

Obviously the surrounding temperature, as well as the manufacturing tolerances, will affect
the precise current at which the fuse will melt; the higher the ambient temperature, the less
the current (and therefore the heat) needed to melt the metal. Consequently the fusing
factor, though easy enough to understand, is not a precise quantity and should therefore be
used with care.

To overcome this the International Standards, and even the latest British Standard No. 88,
no longer refer to fusing factor but use instead two other quantities: the ‘maximum non-
fusing current’ and the ‘minimum fusing current’, which represent the lower and upper
limits of the grey area in between. The maximum non-fusing current is that current which,
under any normal operating conditions, will never melt the fuse no matter how long it is
sustained. The minimum fusing current is that current which, under any normal operating
conditions, will be guaranteed to melt the fuse if sustained for the specified time (usually
stated to be four hours or one hour).

These two quantities are usually given, like the fusing factor, as a multiple (or percentage) of
the normal rated current: e.g. a maximum non-fusing current of 1.2 times rated current and
a minimum fusing current of 1.6 times. Such a dual rating would previously have been
referred to as a fusing factor of 1.4, being the average of the two.

In practice, ~with varying ambient temperature and with normal manufacturing tolerances,
the current which will just fuse the element when-sustained for the specified time will fall
somewhere between these two extremes. Thus a 1 OOA normal rated fuse may have a maximum
non-fusing current of 120A and a minimum fcsing current of 160A. We know then that it
will never fuse below 120A and that it is guaranteed to fuse above 160A if sustained for the
specified time. In practice it may fuse anywhere between.

Because of these uncertainties, particularly at low overcurrents, fuses are not particularly
effective in the low overcurrent region and are not generaily suitable for ove~rload as distinct
from shoyt-circuit protection. In such cases the contactor, if fitted, provides the protection
against overload.

3.3.5 Service of Fuses

Fuses may be used~ in various ways to protect equipment downstream of them. These
services may be grouped as follows:

General Application. The fuse protects all equipment and cable against the effects
of overcurrent, the degree of protection depending on the type and rating of the
fuse selected.

Close Excess Current Protection. The fuses used have a low fusing factor - that is,
the ratio of the minimum fusing current to the normal full-load current - which
enables the cables to be used to their fullest rating.

Motor ProtectionThe fuses have special time/current characteristics which enable

them to pass the repeated large, and sometimes prolonged, motor starting currents
without melting or deteriorating but will still protect the motor against steady
overloading (with small motors) and stalling. Motor fuses generally have to run well
below their rated full-load currents to allow for direct-on-line starting currents to
flow without melting them. Consequently they do not protect the motor against

normal overloads, although they do protect the motor and supply system against
short-circuits. Most motor fuses are specially designed to withstand repeated
starting surges without fatigue, which would otherwise shorten their life. Fuses
intended for application to motor supplies are given restricted continuous current
ratings (‘M’ ratings) ascompared with theirfusing characteristics. Thusa ‘TIA32M63’
fuse has a continuous rating of 32A but a rating of 63A for the period of starting.

Semiconductor Protection. Fuses are used for two purposes in semiconductor

equipment. In a large equipment which contains groups of diodes or thyristors in
parallel individual ceils are often fused to enable them to be disconnected from &he
circuitiftheyfail tosho~-circuit,soallowingtheequipmenttocontinueoperating. A
more demanding duty is the protection of semiconductor devices or assemblies
against external faults. Semiconductor devices have very limited overcurrent
capability, and it is generally necessary to use special fast-acting fuses carefully
matched to the semiconductor ratings and the circuit characteristics.

DC. Fuses. In a d.c. circuit there are not the periodic current zeros that occur in
a.c. circuits, so the mode of fuse operation described above in relation to moderate
overcurrents does not apply. For this reason fuse protection against low or moderate
overcurrents in d.c. circuits demands very careful consideration to avoid a dangerous
arcing situation within the fuse, particularly if the circuit is inductive.

Back-up Fuses. Fuses used to back up contactors or other switchgear used in areas
where the fault level exceeds the breaking capacity of the switch are discussed in
Chapter 4, ‘Discrimination’.

3.3.6 Specification of Fuses

In order to specify a fuse its properties must be given in four classes. These are:

- Application (e.g. industrial,-domestic)

- Electrical characteristics (e.g. voltage, normal current, breaking current, single-

phase or 3-phase system).

- Operating characteristics (e.g. general purpose, close-protection, motor

protection, semiconductor protection, time/current characteristic).

- Physical design (e.g. type of contacts, repiacement of links, degree of physical

protection, non-interchangeability).

3.3.7 Trigger Fuses

Certain fuses are fitted with a device which releases a trigger when the fuse blows. This may
be actuated by a spring which is held in tension until the element melts, or it may be operated
by a small explosive charge. The trigger, when released, may be used merely to indicate the
blown fuse or else to trip a circuit-breaker mechanically by a trip-bar, or it may close a
contact which trips it electrically.

Trigger fuses with tripping facilities are a protection against the effects of single-phasing. If
any one of the three fuses protectjng a motor blows, its trigger makes (or breaks) a contact
which trips the contactor and opens all three phases. Trigger-operated contacts are also used
to gjve a ‘Fuse Blown alarm.

There are special symbols (BS 3939, Nos. 3.10.5 and 3.10.6) which should be used on
diagrams to indicate trigger fuses.




In Chapter 1, ‘Principles of Protection’, it was stated that one of the main purposes of
protection was to remove a faulty equipment or circuit from the electrical system so that as
much as possible of the system could continue to function normally. It is therefore desirable
that any particular fault should be cleared by that protection device which will perform the
service with the least effect on healthy parts, and not by some device further upstream
which would disconnect an unnecessarily large section of the system. For example, if a fault
occurs on one of a number of circuhs fed from one transformer, it is better to isolate that
particular circuit by its own circuit-breaker or fuse than that the transformer should be
disconnected from the supply by its primary overcurrent protection or thegenerator tripped.
This preferential or selective operation of protection devices is known as ‘discrimination’.


Almost all switchgear is fitted with overcurrent protection of some sort. If a fault develops
low down in the system, fault current will flow right through the network from the supply
generator, through every intervening switch, down to the fault point itself. All these over-

Load Load


currents will be detected by the relays of each individual switch, and, if no steps were taken
to prevent it, all might trip together, so shutting down the whole system for what might
have been a purely local fault.

The overall protection system is therefore deveioped so that the breaker (or fuse) nearest
the fault operates first, thereby isolating only the fault itself. If this does not clear it, the
breaker nearest upstream of the fault operates next, thereby isolating only the minimum
number of consumers. If this one does not clear, the next upstream breaker operates, and
this continues until the generator breaker trips, but only as a last resort. Each Sreaker backs
up the one below it.

lt.has already been shown that most protective devices, such as overcurrent relays and fuses,
have an inverse-time characteristic as shown in the middle column of Figure 4.1. This causes
the tripping time to vary inversely as the magnitude of thee fault current. It has also been
shown that in relays the characteristic curve can be altered by adjustment of the relay
current and time settings. For fuses the characteristic cannot be altered, but a different
characteristic can be obtained by choosing a different fuse.

In Figure 4.1 it has been assumed that relay-settings have been chosen and applied:

- for the generator circuit-breaker (breaker C),

- for the HV feeder circuit-breaker (breaker B),

- for the LV feeder overcurrent device (breaker or fuse A),

as shown in the characteristic curves of the middle column. For the purposes of direct
comparison the three curves have been drawn to the same scales of time and current referred
to a common base voltage.

All three curves are superimposed on the right. If ,the settings have been properly chosen,
the curves should appear as shown, each clear of the other at ail points. Since these curves
are subject to tolerance (a relay accuracy of +7% is usual, and there will be other errors), the
curves should all be well clear of each other.

If a fault of current value F (adjusted to a common base voltage) appears at point P on the
network, the fault current flows through all the breakers A, B and C. Characteristics of A, B
and C show that this current would trip (or blow the fuse) A in time Tr , B in time Tz and C
in time T3. Provided that A does trip or blow in time. T,, the fault will be removed and B
and C will not trip at all and all the other consumers on both boards will remain in service.

Should A fail to trip or blow, or if the fault were at point Q higher in the network, the first
breaker to trip. would be B in time Tz, but C would remain closed. More consumers would
be lost, but the generator would remain on-line feeding all others. Only if both A and B
failed to clear would C trip and take the generator itself off-line.

It should be noted that the time delay increases as the tripping point moves nearer the
supply source (in this cdse the generator). For this reason generators and their HV switch-
gear have to have a 3.sec.ond through-fault rating under British and European rules, calling in
general for heavier copperwork, whereas distribution switchgearnormally hason!ya l-second
through-fault rating. (The 3-second rating does not apply in the US.)

Restricted earth-fault and differential protection, it should be noted, whichare instantaneous

and cover only faults within the protected zone, do not form part of a discriminating pro-
tective system. They may however be used together with one. If, for example, a fault
occurred within a transformer, the differential protection would deal with it instantly
without waiting for the time-delayed transformer HV breaker to trip.


To achieve adequate discrimination between two fuses of similar type, it is usual to give the
major fuse about three times the normal current ratingof the minorfuse. Between a moulded-
case circuit-breaker and a minor fuse the ratio can be reduced to about two.

Because Moulded Case Circuit-breakers (MCCBS) and Miniature Circuit-breakers (MCBS)

have instantaneous trips in addition to their normal thermal trips, they will not discriminate
with each other at the higher currents. For this reason it is bad practice to install two
MCCBs or MCBs in series, even though they may have different trip units.


Contactors, MCCBs and MCBs are all described in the manual ‘Electrical Distribution
Equipment, Part A’. Though they can all close onto a fault and carry it for a very short
time their breaking capacities are strictly limited and are far below those of conventional

When used to control equipment in networks, their breaking capacities are tz.ually much
lower than the fault levels of the sy<tem at the points where they are installed. For example,
a high-voltage contactor with a maximum breaking capacity of, say, 180MVA at 6.6kV is
often used on a high-voltage switchboard where the fault level may be SOOMVA. Similarly
low voltage contactors with a maximum breaking capacity of ISMVA at 44OV, or an MCCB
with a breaking capacity of 15MVA, may well be installed in an LV system where the fault
level is 31MVA, or even 50MVA. If any one of these was ever called upon to break such
fault currents, it would undoubtedly fail and probably cause a fire.


Load Fault Current for

MCCB or Contactor




To remove this risk contactors, MCCBs and MCBs are where necessary backed up by HRC
fuses in series. Such fuses would be chosen with a breaking current rating to suit the fault
levels of the system at the switchboard in which they ar e used. LV fuses used on offshore
or onshore installations have a maximum breaking current rating up to an equivalent of
61 MVA.

An LV back-up fuse and its contactor are shown (in single-line) in Figure 4.2(a). The fuse,
and the contactor (or MCCB) in series with it, both pass the same fault current. The character-
istics of most HRC fuses, which are thermal devices and therefore of the inverse-time form,
are generally of a somewhat different shape from those of the overcurrent relay protecting
the contactor or of the MCCB tripping device. Two typical characteristics, for the fuse and
for the contactor relay or MCCB, are shown in Figure 4.2(b).

The contactor relay or MCCB settings and the HRC t&e ratings are so chosen that their
characteristics cross just below the limiting breaking current (for example 20kA at 440V) of
the contactor or MCCB. Suppose the curves cross at point P, corresponding to the maximum
permissible fault current F for the contactor or MCCB, then for a fault current F, less than
F, the contactor or MCCB will be the first to open in time T,, and it will be well within its
rating. For a fault current Fz greater than F which could damage the contactor or MCCB,
the fuse will operate first in time Ta, so protecting the contactor or MCCB which will then
open on a ‘dead’ circuit. Fuses can even be used to back up a main circuit-breaker where the
fault level is near to, or exceeds, its rated breaking capacity. (This can happen, for example,
when the generating capacity of a network is extended after the switchgear has been installed.)

This use of fuses as a back-up for both HV and LV switchgear is very common on offshore
installation systems. Unlike circuit-breakers or contactors they cannot be reclosed but must
be physically replaced after blowing.

It should be noted ~that the back-up fuse selected is chosen solely for its characteristic curve
and not for its normal current rating. It is not intended as overload protection, which is
catered four by the contactor. It is there only to protect the contactor itself against heavy
short-circuits. The actual normal current rating of such a fuse may seem to bear little
relation to the load on the circuit in which it is used, and it must always be replaced by an
identical fuse, not one with a normal rating apparently more suited to the circuit. If this is
not done the whole back-up protection is lost.




Earth-fault protection, as applied to an earthed ax. distribution system, depends upon

connecting a relay to the system in such a way that it measures any earth currents that may
be flowing but is not affected by load currents. The relay is of the same type as used for
overcurrent protection: either instantaneous (attracted armature), symbol E, or inverse-
time (rotating disc), symbol EIT.

Earth-fault protection may be of two kinds, - ‘unrestricted’ (E) or ‘restricted’. The former,
as its name implies, is a basic scheme and provides protection almost anywhere the earth
fault may occur in the system, whereas restricted earth-fault protection (REF) operates only
for faults occurring within a limited and defined area called the ‘protected zone’.

Spill Current








Figure 5.1 shows three ways of connectin, - an earth-fault relay to provide unrestricted

In Figure 5.1 (a) the relay is connected to the ‘spill’ from the three line CTs. Under normal,
or even line-to-line fault, conditions, the currents in the three CTs produce no resultant,
since they are vectorially balanced, and there is no residue or spill to operate the earth-
fault relay. If there is an earth fault however, the three line currents are no longer balanced
and the resulting CT unbalance current spills through the earth re!ay. If it is sufficiently
large the relay will operate. It is essential that the three CTs used in this manner should be
identical, as otherwise a false spill current would be produced.

Figure 5.1 (b) is a simple method whereby a CT is insertedin the earthed neutral of agenerator
or transformer. A current flows in the neutral leg only when there is an earth fault or leakage
somewhere in the system. Earth-fault protection using this method is often employed as a
back-up to the more usual system of Figure 5.1(a), and when it is so used the setting is
somewhat higher. It is often then referred to as ‘Standby Earth-fault Protection’ (SBE).

If third harmonic currents are present in the system, they will return through the neutral.
If they are of sufficient magnitude they could cause the earth-fault relay to operate either
by spill current or through the neutral CT, even though no earth fault was present. Such
harmonic currents are distinguishable from true earth-fault currents by their frequency,
which is three times that of the system.

Special earth-fault relays are manufactured which are de-sensitised to the third harmonic
(150 or 180Hz) to avoid spurious operation.

Figure 5.1(c) is a totally different methods known as ‘Core Balance’, which may Abe used
where the main feeder consists of cables.

The feeder cables are grouped together and passed through the centre of an iron-cored
current transformer, whose secondary is wound as a toroid. Under nocinal current conditions,
or even with a phase-to-phase fault, the currents in the three cables are balanced, and there
is no resultant magnetic field in the core. If there is an earth fault the cable currents are no
longer balanced, and they give rise to a resultant field through the core, causing a current in
the CT secondary winding. This current actuates the earth relay.

Core balance protection is convenient and cheap in a !bw-voltage,system where cables are
readily available in the distribution panels.


It is sometimes convenient to house the earth-fault and overcurrent relays in the same case,
and the practice is encouraged by the fact that normally the re!ays for both functions are
fed from line current transformers and can in fact share a single group of CTs.

In a 3-wire system full overcurrent protection is given with protective elements in any two
of the three phases. Figure 5.2 shows a connection for such a &al-purpose ‘overcurrent-and-
earth’ relay fed from three line CTs. There are two overcurrent elements fed from two of
the three line CTs. The earth element takes the spill current from the three CT circuits,
being unaffected by theabsence of an overcurrent relay element in the third phase. All three
elements actuate parallel &rip contacts as the output of the relay. Each element also releases
its own flag to indicate which one has operated; it must be reset by hand. All three elements
are housed in a single case.

-w Trip
+ Circuit




The earth-fault element may be of the instantaneous or inverse-time type. The combined
relay is then referred to as ‘ZOCIT/E’ or ‘20CIT/EIT’ respectively.


Where an a.‘. system is unearthed there can be no such thing as an earth-fault current, since
there is no return path for such a current. However, small leakages to earth can occur; also a
fault can put one line solidly to earth. Although no fault current will then flow, small
earth-leakage currents are probable. They can be detested, by specially sensitive earth-leakage
relays, although these are not used in Shell systems, which are all earthed.

On some>unearthed systems a line which is faulted to earth can be detected by various

proprietary devices; an example is the ‘Vigilohm’ which continuously monitors the line-to-
earth resistances.

5.4.1 Earth Leakage Circuit Breakers

There are certain types of equipment where an!internal fault which produces even a small
earth-leakage current may present a danger to the operator - for example with portable
tools, welding equipment, etc. In other types it may prevent the equipment from functioning
correctly, as with some trace heating systems.

Such a small earth leakage would certainly not be sufficient to blow a fuse or operate a
normal overcurrent control device, and in these cases the circuit is often provided with an
‘Earth Leakage Circuit-breaker’ (ELCB). Such miniature circuit-breakers can be arranged to
trip with an earth-leakage current as low as 30mA.

The principle is shown in Figure 5.3. The ‘go’ and ‘return’ wires both pass through a magnetic
ring. As long as there is no leakage of current to eal-th both wires carry equal and opposite
currents, and there is no net magnctisation of the ring. However, if a small part of the
current in one wire leaks away to earth, the currcntj in the two wires are unbalanced,
and the ring becomes magnetised and acts as a ring-type current transformer. A toroidal
secondary winding on the ring develops a current which causes the circuit-breaker to release.

From Supply


To Circuit


This arrangement is self-powered and requires no separate trip supply. A difference in the
line currents of as little as 3OrilA is enough to cause an ‘instantaneous’ trip within about
30 milliseconds. After tripping t’le ELCB must be reset by hand.

The type of breaker shown in Figure 5.3 and described above is known as a ‘residual current
device’ (red), or ‘residual current circuit-breaker’ (rccb). The device works on the imbalance
of current in the wires which thread its core and functions independently of voltage. It can
be tested by a Fault Test pushbutton. This bypasses part of the outward (but not the return)
current thiough a resistance, ~thereby unbalancing the two wire currents inside the ring and
causing the ELCB to trip. The pushbutton latches in and must be released by a second
mechanical button.

A different type of ELCB depends on voltage, not current. The voltage developing on the
earthed casing of a faulty piece 05 equipment is sensed and operates a shunt trip on the
breaker if it exceeds about 4OV, which is a voltage considered not to endanger life. The
ELCB behaves much as a very sensitive overvoltage relay and acts almost instantaneously.
This type is suitable where the earth resistance is too high to permit enough earth-leakage
current to flow, but elsewhere it is now being superseded by the residual current type.

ELCBs are being increasingly fitted in domestic and industrial installations. Indeed it is now
required in certain sittiations by the IEE Wiring Regulations (15th Edition), particularly
where sockets are used to feed portable apparatus. Their prime purpose is to protect personnel
against serious electric shock due to faulty apparatus.

If a person in good contact with earth accidentally touches a live conductor or the case of a
piece of apparatus made live by an internal fault, an earth-leakage current wil,l flow through
him. A current of 30mA, if sustained for even a short time, could prove fatal. An ELCB
which operates at 30mA or lessand breaks the supply within about 30ms, while not preventing
a shock, will not allow sufficient energy to pass to prove fatal or even dangerous to most

5.4.2 Earth Leakage Protection in D.C. Systems

Most d.c. battery-supported supplies (other than communications power supplies) are not
earthed directly or through a low resistance, as in the case of ax. systems. Consequently a
single earth fault does not give rise to any considerable unba!ance between go and return
currents, and the methods of earth-fault and earth-leakage detection used in ax. systems are
not suitable for d.c.

Earth Leakage .~.




The usual meth~od of earth-leakage detection in d.c. systems is based on a potential divider
consisting of a pair of equal highavalue resistors connected in series across the supply, as
shown in Figure 5.4. As long as there is no leakage to earth (Figure 5.4(a)) the supply
voltage is equally divided between the two resistors - that is, V, =~ V, = V/2. A voltage-
operated relay is placed between the mid-point and earth. Since normally no current flows
to earth (as there is no leakage), there is no potential difference across the relay, and it is
not affected. Both sides are at earth potential.

Suppose now a leakage to earth developed in the negative line (Figure 5.4(b)). ?he leakage
resistance is R, and the consequent leakage current I,. This leakage current will flow, as
indicated in red, from the positive line, through the left-hand resistor R, through the mid-
point connection, through the relay to earth and back through the leakage resistance R, to
the negative line. This leakage loop current can be considered as superimposed upon the
steady potentiometer current, which continues to be V/2R.

The total current in the left-hand resistor limb is therefore

‘V[2R + iE

It flows through the resistor R, giving a volt drop across the resistor of

(V/2R + /,)R

= v/2 t- I, .R.

This is therefore the new potential of the mid-point, which was previously at earth potential.
This voltage appears across the relay and, if sufficient, causes it to operate.

A similar situation would exist if the leakage had occurred on the positive line, except that
the currents~ would be in the opposite direction. If the relay were of the ‘2-way’ type it
could be made to distinguish between a nositive or a negative line leakage.


Restricted earth-fault protection is similar to that for a normal earth fault already described,
but it is arranged to provide protection against earth faults occurring within a limited zone
only. It can only be used on earthed systems.

There are two variants: one where the neutral is the switchboard (in a 4-wire
system), and the other where it is locally earthed at the source of supply (at the transformer
secondary star-point in the case of an LV system). Roth systems are found in Shell installations.

5.5.1 Switchboard Earthing

Figure 5.5 shows the connections as applied to the secondary of a star-connected transformer
which is earthed through the LV switchboard neutral busbar. Three CTs are arranged as for
normal earth-fault protection, and a fourth CT is provided in the neutral line between the
star-point and the switchboard earth and is paralleled with the other three. The earth relay
ins, as before, in parallel with the CTs and is operated by the resultant spill current from all

Under normal 4-wire conditions load currents, which may be unbalanced, flow in all four
CTs, but no resultant spill current flows in the earth relay. Even under phase-to-phase or
phase-to-neutral fault conditions the overall balance of the four CTs is not upset,~ and the
earth relay is not affected.

A ‘through-earth fault’ - that is, a fault downstream of the CTs- will produce the conditions
of Figure 5.5(a). The path of the primary fault current is shown in‘red; the earth path
between the fault point and the switchboard neutrai earth is shown dotted.

The fault current path passes in opposite directions through one line CT and the neutral
CT. The only CT secondaries affected are those shown in green, and~it will be seen that they
are equal and opposite. There is therefore no spill current due to the fault, and the earth
relay will ignore it.

If however the earth fault occurs upstream of the CTs as shown in Figure 5.5(b), the primary
fault current (red) passes only through the neutral CT and through none of the others. Its
secondary current therefore can only pass through the earth relay (green), which is thus
actuated and operates the trip contacts as shown.

No Spill





This system is therefore immune to earth faults downstream of the CTs (‘through faults’)
and also to line-to-line faults. It will respond only to earth faults upstream of the CTs: that
means to faults only in the transformer secondary windings, cables and cable boxes or
switchgear up to the CTs, but not beyond. This is called the ‘protected zone’ - hence the
name ‘Restricted Earth Fault’. It serves partly the same purposeas true differential protection
as described in Chapter 6 but should not be confused with it.

5.5.2 Local Earthing .

With this system there is a fifth CT in the earth connection, in parallel with the other four.
Figure 5.6(a) shows the primary fault current flow (red) with a through-earth fault, and the
two affected CTs secondaries are in green. Their currents cancel out, and there is no spill
current; the earth relay is not affected.

Figure 5.6(b) shows an earth fault within the protected zone - that is, upstream of the CTs.
Here only one CT, that in the earth connection, is affected, resulting in spill current and
operation of the earth relay.

Here again the arrangement ignores through-earth faults and responds only to earth faults
within the protected zone, but in this case five CTs are required.

I Zone
I ‘? Cable




As restricted earth-fault protection is insensitive to through faults and’does not take part in
the discrimination pattern (see Chapter 4), an REF relay is given a very low current setting.

5.5.3 Protected Zone

It should be noted that REF protection is the only means of protecting the transformer
secondary winding, cable box and cable. With unrestricted protection only faults downstream
of the CTs, which are usually in the LV switchboard, will be detected; any within the
transformer or between it and the CTs will not be sensed, and the whole of that zone wifi
remain unprotected. Only restricted protection will cover it.

Because REF protection is confined to the limited protected zone only and is not affected
by protection systems outside it, it can be made instantaneous in operation and need not
form a step in the discrimination ladder.

It is worth remarking that an earth fault occurring near.the earthed star-point of the trans-
former secondary winding may, due to the low voltage there, not develo,p sufficient fault
current to actuate the REF protection. Although a fault at such a point is unlikely, this is a
small.loophole in REF protection.



Differential protection depends on a method of fault detection based on the principle that
the total current flowing into one part of a system is equal to the total current flowing out
of it unless there is some unintended alternative path for it in between. This is just another
statement of Kirchoff’s Law.

This type of protection is used to gu~ard against faults arising only within the protected unit,
ignoring those occurring outside it. The unit itself then becomes the ‘protected zone’. It is in
some respects similar to restricted earth-fault protection but should not be cotifused with it.
REF guards only against earth faults in the protected zone, whereas differential protection
covers also phase-to-phase faults within the zone. It does not however deal with inter-turn
faults within one phase - say in a generator - since that will not c&e differing currents
at the two ends.

Differential protection is insensitive to through faults-that is, tofaultsoutside the protected

zone - because the same fault current flows through both ends of the zone. It may therefore
be used to provide relatively sensitive protection for the equipment inside the protected
zone without its being affected by the discrimination scheme of the whole network. The
advantage of this is particularly apparent in the case of generators and large bulk power
transformers, which may demand rapid and sensitive protection against internal faults but
which, because of their position at the high-level end of the power supply system, would be
among the last items to be tripped in the went of a through-fault.


emf emf
- -



The term ‘differential protection’ (symbol DIP) is used generally~throughout the offshore
installations, but elsewhere it may be known by the names ‘Merz Price’ (after the original
inventors), ‘Unit’, ‘Circulating Current’ or ‘Balanced Voltage’ protection. All these terms
may be met as well as such trade names as ‘Translay’ and ‘Solkor’ which introduce variations
into the basic scheme.

Differential protection is basically of two kinds, as shown in Figure 6.1. The two kinds are
described in principle below.

6.1.1 Circulating Current (cc) Principle

Figure 6.1(a) shows identical current transformers connected at each end of any piece of
electrical equipment - a generator, a motor, or even a‘length of cable - through which a
current is flowing. A single-phase circuit has been used for simplicity. The CT secondaries
are connected by ‘pilot cables’ in a loop as shown, and a voltage-sensitive relay is connected
across the pilots at about their mid-points. L

Current flowing through the electrical unit’causes a-secondary currentthrough both CTs to
circulate round the pilot circuit without producing any current in the relay. A fault within
the zone between the two CTs (the protected zone) will on the other hand cause secondary
currents of differing values in the two CTs, and their difference current will flow through
the relay. If this difference is sufficient, the relay will operate.

6.1.2 Balanced Voltage (bv) Principle

Figure 6.1 (b) shows another arrangement where the two current transformers are connected
in opposition and the relay is in series. With the same primary current flowing through both,
the secondary emfs oppose each other and no secondary current flows in the pilot circuit -
the voltages are balanced.

In the event of an internal fault causing differing primary currents in the CTs, the two
opposing secondary emfs will no longer be equal, and current will flow round the pilot
circuit, causing the series relay to operate.

It should be noted that in the balanced voitage system no CT secondary current flows
normally, and the CTs are effectively on open circuit, giving high voltages on the pilot lines.
Moreover thiscondition would cause theoverburdened CTs tosaturateand become inaccurate.
Special CTs are used having an air-gap or other non-magnetic gap to avoid saturation.


6.2.1 Voltage Distribution

The simpiified explanation of circulating current protection as given in para. 6.1.1 needs
some further attention in order to understand how it works in practice. In particular the
distribution of voltages round the secondary loop will be described.

If the potentials at all points round the secondary loop are plotted, beginning at 0 where
the potential is zero, the curve will be as shown red in Figure &2(a). From 0 to A the
potential will rise due to the emf in the CT; from A to B it will steadily fall due to the
resistance of the pilot leg AB; from B to C it will rise again within the CT; and from C to 0
it will fall once more to zero due to the resistance of the leg CO.


CT Secondary


/. Unit

CT Secondary



At a certain point P midway between the two CTs the potentials of the two secondary lines
(red) will be equal because of symmetry. A voltmeter applied across them there would read
zero. If a relay were connected across the lines at that point it would be unaffected.

If now a fault or leakage developed somewhere inside the equipment, part (or all) of the ‘go’
current would be shunted into the return line, so thar the currents I, and I2 on either side
of the equipment would be unequal. So, therefore, would be the CT secondary voltages, and
the potential curves would be distorted as shown red in Figure 6.2(b), the voltage gradients
on the faulty side being greater than on the other. They are no longer symmetrical, and
the crossover voltage-balance point has moved from P to some other point Q. At P there is
now, a voltage difference between the lines (PI--P,), and the relay (an attracted-armature
instactaneous type) inserted at that point would be energised. If the relay setting were
sufficient, it would operate to trip whichever circuit-breakers it was necessary to open. The
relay setting range is typically 5 to 20% of normal full-load current.

It has been shown that the relay must be connected at the point in the pilot lines where,
under normal conditions, the voltages are equal. In practice such a point is not easy to find.



What is done is to insert resistances intb the pilot circuit so that most of the voltage drop in
each lilie is concentrated in the resistors. The crossover point is then bound to besomewhere
in the resistors themselves,, so they are provided with tappings, which can be adjusted until
the balance point is found. By this means the crossover point, instead of being at some
unknown place far from ?he switchboards, is brought as a ‘resistance box’ right into the
switchboard where the relay itself is installed.

The resistances add to the burden on the CTs, but this is acceptable.

For satisfactory operation it is essential that the pairs of CTs be accurate and perfectly
matched. Therefore they are usually of the special class of accurdcy (Class X -see manual
‘Electrical Distribution Equipment, Part B’, Chapter 4) and are supplied as matched pairs.

Since differential protection operates only over a limited zone, it does not form a step m me
discrimination ladder. It is therefore instantaneous in operation and the relay can be given a
very low setting.

6.2.2 3-Phase Protection

Figures 6.1 to 6.3 show, for simplicity, a single-phase system, but the principle can be
applied -and usually is-to 3-phase systems.

Three carefully balanced pairs of CTs of high accuracy are inserted, one pair into each of the
three phases, and voltage balance is measured between each secondary line and neutral by a
3-element relay. A resistance box containing three tapped resistors is used as described
above. This is shown in Figure 6.4.

The 3-phase system requires four pilot lines between the sets of CTs, with further lines from
the relay contacts to trip the circuit-breaker. For long lines variations of the system such as
‘Translay’ and ‘Solkor’ operate over only two pilot lines and can initiate tripping simul-
taneously at both ends. It should be noted also that differential protection will operate
1 on both internal phase-to-phase and earth faults, and in this respect it issuperior to restricted
earth-fa Ilt protection.


1 1 1
6 &*-?iP



6.2.3 Differential Protection of a Transformer .j

The differential protection so far described, whether circulating current or balanced voltage,
depends on identical and matched current transformers at both ends of each phase. For
most electrical units the incoming and outgoing currents are, or should be, the same. This
applies to generators, motors and cables, but it is not true of transformers.

Th,e outgoing current in any phase of a transformer differs, ideally, from the incoming in
inverse proportion to the voltage ratio. For example a 2 OOOkVA, 6 600/44OV transformer
(ratio 15:l) has a primary current of 175A but asecondary current of 2 62SA (ratio 1:15).

Therefore, to achieve balance of the CT secondaries, the CT ratios must be inversely propor-
tional to the main transformer voltage ratio, as shown i,n Figure 6.5(a).

Most distribution transformers are delta/star connected, and this affects the line current
ratio in the individual phases by a factor ofJ3. If the main transformer is delta/star connected,
then the three CT secondaries must be connected in the opposite sense, namely star/delta.
This is shown in Figure 6.5(b).

1 n
pM -



1 i Relays
1 !




As stated in para. 6.1 the balanced voltage system is less used than the circulating current

One consequence of the high voltage on the pilot lines is that it can give rise to appreciable
shunt capacitive currents if the pilot cable is long; these can lead to inaccurate operation
unless special steps are taken to deal with them.

It is for these and other reasons that the circulating current type of protection is generally
preferred. In the US the balanced voltage system is referred to as ‘transactor’.




Undervoltage protection may be required for a variety of purposes. The principal ones are:

- to avoid risk or danger that could arise if power were restored to particular
parts of a system without due precautions after an,interruption,

- to shed load in order to avert a threatened collapse of the system in the event
of a severe drop in voltage resulting from a disturbance.

Under the first heading it is usually necessary, if power is lost for any reason, to clear the
system of major loads. This is to avoid the danger of motors restarting without warning
when power is restored and to avoid overloading of the system by many motors restarting
simultaneously. In addition, any interconnected switchboards must be disconnected.

Ordinary contactors will drop out on loss of power, thereby automatically disconnecting
their loads, but circuit-breakers (and some special contactors) are latched-in and will not trip
on loss of voltage unless they are given a positive trip signal. Such a signal is given, where
needed, by an undervoltage relay. This is an ordinary voltage-operated relay fed either
directly or through a voltage transformer from the supply side of the circuit to be protected.
The relay is calibrated to drop out when the voltage fails to a predetermined level (typically
55% of nominal), whereupon its contacts, which are open when the relay is energised, close
and initiate the trip signal. An undervoltage relay is usually fitted with a time delay, either
within the case or as a separate relay, to prevent operation due to momentary dips in the
system voltage.

This type of relay is sometimes incorrectly referred to as a ‘No-volt’ relay. This term should
not be used. The relay operates not only when no voltage is present but also when it falls to
a preset level; its function is ‘undervoltage’.


When a whole system undergoes a serious disturbance, with a consequent drop in voltage,
a complete collapse may result unless large blocks of consuming equipment are temporarily
removed~ from it to enable it to recover. To avoid this,, the reduction in system voltage is
detected by an undervoltage relay, typically set to release at 55% of nominal voltage (settings
vary from system to system), and loads are disconnected in groups. The items most likely to
prevent system recovery are the largest motors, and they must be disconnected. This mode
of protection, known as ‘System Undervoltage Trippihg’, is concerned specifically with the
eventual recovery of the system. It should be distinguished from load-shedding which is for
the purpose of keeping the total steady-state load on a system within the available generation

Figure 7.1 shows one form of system undervoltage tripping applied in an offshore installa-
tion. The largest motors in this case are 9 240kW and 8 600kW gas re-injection and booster
compression motors and six water injection motors, of 2 500kW each, all supplied at 6.6kV.
Undervoltage relays connected to the main 6.6kV busbars and the gas compression switch-
board trip the reciprocating compressor (the largest) instantaneously and the booster
compressor and the injection and other loads after a time delay if the system has not already
recovered sufficiently.

L Trip Trip
(3 set)

j -i
6.6kV Gas Compression
/ Switchboard
/ r


3 x 2500kW

i: water water
Injection Injection
Pumps Pumps M
9240kW Lean
Re-iniection Oil
Gas Pumps’
Compressor Compressor

Tl( T3 T4
( t

440V Gas Compression





The forms of protection described in Chapters 3 to 7 are the most common forms in offshore
and onshore installations. A number of other forms are used for specific purposes, based on
the particular requirements of individual items of equipment. They are in some cases simple
applications of relays and principles already referred to. Others entail special principles, and
these are described below. Forms of protection which are special to individual categories of
equipments are described in later chapters.


Overtemperature protection is often used to safeguard the windings of generators, motors

and transformers against excessive temperattme rise due to overloading or fault conditions.

Three main types.of temperature sensor are used:

- Thermocouple
- Resistance Temperature Device (RTD)
- Thermistor.

The principles of these methods are discussed, in the manual ‘Electrical Control Devices’.

Sensors are used for temperature-monitoring instruments in oil-cooled (or Askarel-cooled)

transformers. There~they ire typically suspended in the oil in a housing with a heating
element which carries a current (from a current transformer) proportional to the transformer
current. The housing is designed to reflect the thermal characteristics of the transformer,
and the~sensing element thus experiences something yery close to the actual winding tem-
perature; the technique is known as ‘thermal imaging’. The thermistor is connected into a
resistance bridge, whose output may operate indicating instruments as well as actuating
alarms and trips through an electronic detector circuit.

Whereas NTC thermistors can operate over a ranges of temperatures by adjustment of the
associated measuring circuits, a PTC thermistor is made to change its resistance at a particular
temperature, subject to a small tolerance. It is more suitable for detecting overtemperature
at particular locations in equipment - for example, at hot spots in generator or motor
windings into which they can be embedded during manufacture. As the PTC thermistor
passes through its critical temperature, the sudden change of resistance can be made to
actuate an alarm or even to give a trip signal.

PTC thermistors come in three classes, depending on their critical temperatures, as follows:

Class B 145°C
Class F 165°C
Class H 190°C


Arc detection offers a rapid means of protecting against arcing faults - in switchboards, for

x_., example - which can result in considerable damage if they are allowed to persist until they
are cleared by the normal protection. Protection is provided by light-sensitive cells which are
so located and of such a sensitivity that they respond to the ultra-violet light of arcs but not
to ambient light. A number of photocells can be linked to one control unit to pravide
protection over an area.


Protection against the reversal of power flow (such as~when a generator ‘motors’) is provided
by a relay which is sensitive to the direction of power flow.

Such protection is used mainly (but not exclusively) with generators and is discussed in
Chapter 9, ‘Generator Protection’.


Overpressure protection is used in liquid-filled transformer tanks and is discussed in Chapter
10, ‘Transformer Protection’.


Negative phase sequence protection is used mainly, but not exclusively, with motors and is
discussed in Chapter 11, ‘Motor Protection’.

It is sometimes also used in large generators to protect against overheating of the rotor. If a
generator suffers~ a seriously unbalanced load, a negative sequence current is present in then
stator windings, giving rise to a counter-rotating fie(d. This field cuts the forward-running
rotor at double the system frequency (i.e. at 100 or i20Hz) and can cause serious overheat-
ing of the rotor iron due to eddy currents and hysteresis.

When NPS protection is applied to a generator, it is set typically at 12% negative sequence


Other forms of protection, discussed in Chapters 9 and 10 for generators and transformers,
are as follows:

- Overfrequency (OF)
Underfrequency (UF)
- Overvoltage (OV)
- Field Failure (FF)
- Diode Failure (DF).

All types of protection, for which British Standard abbreviations have been given, may be
identified also by an international system of numbers, from 1 to 99, which are given in
BS 3939, Appendix II to the Guiding Principles.




All offshore installations have main generators driven by gas-turbines, and both offshore and
onshore installations have diesel-driven auxiliary or basic services generators. Each electrical
generator has its own protection system, which is usually more complex on the main generators
than on the smaller diesel sets. Many of the general protection measures described in Chapters
3 to 8 are applied to generators, but there are also some more specific ones. A typical gas-
turbine generator protection scheme is shown in Figure 9.1.

---- _______ +-----------,

Q. r g----------c ---------_

--. 7’, ’ I

c CT---,------‘------
I I i
ii ~_
i iii

All relays are located on the

generator panel of main switchboard

DIF DifFerential
FF Field Failure
OCiT Overcurrent (Inverse Time)
Reverse Power
SRSPE Standby Earth Fault
TH Trip, Hand Reset (Lockout)

1b The scheme of Figure 9.1 includes the following standard features:

- Overcurrent protection (30CIT) fed from protective CTs. The time and
current settings will be determined by the overall discrimination plan.

- Differential protection (DIF) fed from special (Class X) protective CTs.

- Standby earth-fault protection (SBE) fed from neutral CT.

- Lock-out hand-reset relay (TH). This also trips the turbine and suppresses
the generator field.

When a generator is subject to short-circuit the overcurrent in the initial subtransient stage is
at first high. The short-circuit current, if it is not cleared, quickly falls as the generator
passes through the transient stage until, by the synchronous stage, it may have fallen to less
than the normal full-load current due to armature reaction - that is to say,‘by then there
will be no overcurrent at all.

This tends to limit the effectiveness of any ov~ercurrent protection unless special devices,
.<-,,, such as a voltage-restrained overcurrent relay, are used. This renders the relay progressively
more sensitive as the generator voltage falls under the stiort-circuit.

The need for differential protection of a generator is important. It is explained in Chapter 4

that in a complete protection system there isa pattern of discrimination where the protective
relays nearest the fault operate in the shortest time, and the delay time becomes longer pro-
gressively further back from the fault. In this chain the generator is the last link and therefore
has the longest time delay. In really long chains this delay can amount to several seconds.
For faults well downstream of the generator this is perfectly satisfactory, but for serious
faults near the eenerator such a delay could be damaging.

A fault actually inside the generator would not be detected by the overcurrent CTs at all,
since the fault current does not pass through them. For protection against such an event
only differential protection can be used. It is sensitive only to internal faults in thegenerator,
its cable box and connections. If such a fault should occur, the differential relay trips the
generator breaker instantaneously and stops all other generators feeding into the faulty one.
It does not however prevent the faulty generator feeding its own fault, but the associated
field suppression will limit this (see para. 9.2.8).

Differential protection completely ignores external, or ‘through’, faults and allows the
normal discrimination to operate. Because of this the differential relay is given very light
settings and operates instantaneously as it does not have to be subject to the delay of the
normal overcurrent relays.


Figure 9.1 shows a typical protection system for a main generator and includes many of the
types already described. They are of a general nature used also on other items of plant.
In addition there are types of protection special to generators, and these are described
individually in the~following paragraphs.

9.2.1 Reverse Power (RP) Protection with Time Delay (TD)

Reverse power protection is provided by a wattmetric relay, often of the induction type,
which detects active power flowing only in the direction into the generator. When this
occurs it trips the generator breaker, so disconnecting the generator from the source of

reverse power. The relay is fed from the same protective CTs as the overcurrent, but its
voltage comes from the only VT, a ‘measurement’ type. A time delay prevents operation
with a momentary swing of power; it may be either built-in to the relay or provided as a
separate unit.

Figure 9.2(a) shows the RP relay connections with a typical balanced load such as occurs in
an offshore generator. It is a Z-element induction disc (or induction cup) type. This type of
relay develops its maximum operating torque when the applied current leads 30” on the
applied voltage.

The relay is connected as shown in Figure 9.2(a), with the voltage taken between red and
blue phases and the current in red phase. The primary voltages and currents appear as in
Figure 9.2(b) when the current is flowing in the normal (forward) direction and at unity
power factor (IR in phase with V,). The VT reverses the secondary voltage, and the CT
secondary is connected so that it does not reverse the current. The secondary voltage and
current applied to the relay are then as in Figure 9.2(c). Here the current lags on the applied
voltage by X0”, and the torque on the relay element is in the non-operating direction.

If the primary current direction is reversed but the voltage remains unaltered, a ‘reverse
power’ situation exists. The primary voltages and currents will then appear as in Figure
9.2(d), and the secondary as in Figure 9.2(e). The current in the relay now leads the applied
voltage by 30°, and the torque is maximum in the operating direction. The disc rotates until
a contact on it strikes a fixed contact and initiates the time-delay mechanism, and eventually
the trip signal.

The relay can be set to operate with a reversed in-phase component of current (that is,
reversed power) down to less than 3% of normal current.

Reverse-power protection is usually provided for all generators except where they only run
singly. It is necessary whenever two or more generators operate in parallel. If the prime
mover drive to one of them fails, the other will pump power into it so that it motors and
back-drives the (possibly faulty) prime mover. The RP relay detects this and disconnects the
faulty set.

9.2.2 Diode Failure (DF)

Diode failure protection senses a failure of one of the six excitation diodes which rotate
with the generator rotor. Because of AVR action, failure of a diode would not otherwise be
discernible in the output from the generator. The diode failure relay may give an alarm only;
but in the case of Avon generator sets it also shuts down the set. After the alarm has been
given the earliest opportunity should be taken to stop the set and replace the faulty element.

Diodes can fail due open-circuit or to short-circuit. The latter is the more serious,
as it can result in serious overloading of the exciter.

9.2.3 Overfrequency and Underfrequency (OF and UF)

Frequency is monitored by~relays through a voltage transformer, and an alarm or trip contact is
actuated when the frequency exceeds or falls below the preset values. Underfrequency can
be harmful, since it causes excessive current in most inductive apparatus such as relays,
solenoids, etc. The reactance of such items is given by 2rfL, where f is thefrequency in hertz
and L the inductance in henrys. If f falls, the reactance - and therefore the impedance -
falls also, causing a rise in current and consequent overheating. Operation of the OF or UF
relay will cause the generator breaker to trip.

Shading R:ing

Voltage Coil





9.2.4 ~‘Overvoltage (OV)

Overvoltage can occur due to AVR malfunction or loss of sensing, or to unskilled operation
of the manual voltage control. If too great it can clearly lead to excessive currents through-
out the network and to risk of widespread burnouts. It is most likely to occur when running
a single generator. Operation of the OV relay will cause the generator breaker to trip.

9.2.5 Field Failure (FF) with Time Delay (TD)

If the excitation of a generator should fall below a certain level, there is danger that the
machine will become unstable and fall out of step with other generators in parallel with it.
This situation is detected by a special ‘field failure’ relay fed with output current and
voltage. The time delay ensures that the condition must persist before the relay operates.
Field failure will also cause a large reactive circulating current to flow from the healthy and
to the faulty generator; this may cause both machines to trip on overcurrent.

9.2.6 Winding Temperature (WT)

The temperature of a generator winding is sensed by a number of embedded detectors

placed in the winding insulation, usually in the overhang.

Three main types of temperature sensor are used:

- Thermocouple
- Resistance Temperaiure Device (RTD)
- Thermistor.

The principles of these methods are discussed in the manual ‘Electrical Control Devices’.

The temperatures at various points of the windings can be scanned from the control board.
If any one of them should indicate a hot spot an alarm is given. The cause should be investi-
gated and corrected if possi,ble. If this cannot be done the set should be taken out of service
for examination and another substituted.

9.2.7 Overspeed (OSP)

Overspeed protection is provided by the overspeed device for the turbine. If it operates, it
.also trips the generator breaker_

9.2.8 ,Field Suppression

Field suppression protection is not a relay but operates whenever the generator breaker is
tripped through t~he lock-out relay. It operates by breaking the high-frequency supply from
the generator pilot exciter, so killing the AVR and main exciter and removing the~main field
drive. The main field will then decay at a rate determined by its own time constant, not

Although field suppression is a back-up protection when the generator breaker is tripped for
any reason other than normal manual opening, it is essential for differential protection. The
differential system is actuated by .a fault internal to the generator, and merely opening the
generator breaker will not remove it. Internal damage can only be prevented - or at least
minimised -by suppressing the field.

All malfunctions which trip the generator breaker through the lock-out relay also trip the
prime mover by shutting off fuel, so bringing the whole set to rest. They also actuate the
field suppression. It should be noted from Figure 9.1 that overcurrent is not one of these;
it trips the breaker direct and not through the lock-out relay, so that the turbine is not
stopped in this case, the argument being that overcurrent is not strictly a generator mal-
function. Notwithstanding this, in some installations the overcurrent does operate through
the lock-out relay thus also tripping the turbine: Similarly all mechanical malfunctions
which stop the prime mover also trip the generator breaker as soon as the trip signal is given
to the turbine.


It should be noted that Figure 9.1 is presented as a single-line diagram, as this is the clearest
way of showing protection circuits. The various relays are depicted in ‘boxes’ with identify-
ing letters or numbers, and they are fed often from common current and/or voltage trans-
formers, the current-operated relays in series and the voltage-operated relays in parallel. The
relays, though shown together in the diagram, may in fact be located on many different
panels or switchboards.


The usual practice on offshore installations, and indeed on most power station generators
onshore, is to earth the generator neutral point through a ‘Neutral Earthing Resistor’ (NER).
This limits the c’rrrent due to an earth fault on the system to a level which will not damage
the generator and which will reduce the energy liberated at the arcing earth-fault point.

Ideally the earth-fault current should be kept as low as possible, but not so low that it
would be insufficient to operate the earth-fault protective relay. A compromise is therefore
necessary, and the neutral resistance is chosen to limit the maximum earth-fault current to
something between one-quarter and normal full load of the generator. The earth-fault relays
are set accordingly; their settings will be lighter than those of the overcurrent relays.
If more than one generator is on load, each will contribute, in proportion to its size, to the
total current at the fault point. The minimum earth-fault current will occur when only one
generator (the smallest if their ratings are different) is running, and it is to this value that the
relay must be set.


Generator malfunction alarms are indicated, along with the turbine alarms, on local control
panels. Different makes of gas-turbine generator sets display different alarms in different
groupings, so no general description can be given. Instead the alarms associated with a
typical offshore installation using Avon generator sets are listed below as an example. They
are here grouped into two annunciator boards: an ‘Avon Alarm Unit’ (AAU) and a ‘General
Alarm Unit’ (GAU) on the local control panels. The AAU is confined to turbine alarms only
and is not reproduced here. The GAU includes both generator and other turbine alarms.
They are divided into two groups: those which cause alarm and trip, and those which cause
U alarm only.


Lame Lamp

1. Emergency manual trip 17. Generator stator temperature high

2. Generator protection operated 18. Generator cooler air/water leak
3. Avon fire protection operated 21. Generator vibration high
4. Generator gearbox bearing temperature 22. Gearbox vibration high
excessive 24. Excitation on manual
5. Lub oil supply fault 25. Lub oil pressure low
6.. Generator vibration excessive - standby pump selected
7. Gearbox vibration excessive 26. D.C. lub oil pump failure
8. Gas detected, high level 27. D.C. lub oily pump running
9. Generator stator temperature 28. Turbo-generator lob oil supply fault
excessive 29. Generator cooling system fault
10. Generator diode failure 30. Generator/gearbox bearing temperature
31. Avon fire-protection fault
32. Essential a.c. supply fault
33. Excitation fault
34. Discrepancy detected
35. Avon synchronisation failure
36. Gas detected, low level
37. Auxiliaries fault
40. Gas detector fault

Operation of any of the generator protection circuits which trip through the lock-out relay
will actuate Lamp No. 2 ‘Generator protection operated’. The specific cause of the trip
must be determined from individual relay flags.

Alarms are usually displayed as a set of annunciator windows on the control panel. An alarm
causes the appropriate window lamp to flash and an audible warning to sound. When the
operator presses an Accept button, the audible warning ceases and the lamp burns steadily.

After the fault has been cleared or isolated, pressing a Reset button extinguishes the lamp.
Until this has been done (and the lock-out relay reset by hand) a tripped generator cannot
be restarted.

A Lamp Test button enables all the lamps to be checked at any time.



All main transformers which transmit bulk power between thegeneratorsand the low-voltage
distribution system of an offshore installation, and between the Supply Authority’s system
and the low-voltage equipment in onshore installations, have their own individual protective
systems. This is to protect the transformer against damage due to electrical faults arising
both outside and inside it.

A typical transformer protection scheme is shown in Figure 10.1, which alsoshowsassociated

instrumentation. Many of the general protect& measures described in Chapters 3 to 8 are
applied also to transformers, but in addition there are some more specific~ones.


r’-- x D Fault Trio

a.-.-.-.-. ; I
-.-.-.-.-_-.-_ ;

A-MD Ammeter with Max. Demand Contacts

E Earth Fault
FG Flag Relay
oc Overcurrent (High Set)
OCIT Overcurrent (Inverse Time)
Q Qualitrol Device
REF Restricted Earth Fault
TH Trip, Hand Reset (Lockout)


Points worthy of note in Figure 10.1 include the following:

- Overcurrent protection is on the HV side only. It is provided by two inverse-

time elements combined with an earth-fault element (20CIT/E) together with
two instantaneous high-set overcurrent elements (2X), all in the same case.
The relay o~perates to trip the HV circu,it-breaker directly and both the HV
and the LV breakers through the lock-out relay (TH). The time and current
settings will be determined by the overall discrimination plan. Overcurrent on
the LV side causes corresponding overcurrent on the HV side, which th,erefore
takes care of both overloading and LV short-circuits.

- Restricted earth-fault protection is used on the secondary side (it is the only
secondary-side protection), with four protective-type CTs. The relay operates
instantaneously to trip both the HV and the LV breakers through the lock-out

- Lock-out hand-reset relay (TH).

- There is interlocking and intertripping from the HV to the LV circuit-breakers

(but not in reverse). This is explained in para. 10.4.3.

- Instrumentation includes a maximum-demand ammeter with an alarm contact.


It is explained in Chapter 9 that differential protection must be provided for generators

because an internal fault is self-fed and would not be cleared by the generator supply
breaker. Such differential protection, not forming part of the discrimination ladder, is
arranged to operate instantaneously;

In the case of transformers however there is a circuit-breaker upstream of the unit, and this
can clear an internal fault by removing~the supply that feeds it. If the upstream circuit-
breaker protection has an instantaneous ‘high-set’ relay (as here), the clearance can be

Therefore it is not usual practice to provide differential protection to offshore, or to smaller

onshore, transformers, but to rely on the HV protection to clear any internal primary or
‘through’ fault. Internal earth faults on the secondary side are within the protected zone and
are dealt with by the REF protection.

Nevertheless large onshore transformers are often provided with full differential protection
(not shown in Figure 10.1) using three primary side and three secondary side current trans-
formers, as described in Chapter 6. This gives the same benefits as restricted earth-fault
protection and, in addition, rapid protection against inter-phase faults in the transformer, as
well as earth faults on the primary (delta) winding. In these respects it is far superior to REF

The difference between the primary and secondary currents in a transformer because of its
turns ratio does not prevent the necessary balance in the differential,~relay circuits so long as
the current transformer ratios are in inverse proportion to that of the power transformer.
Where, as is usually the case, the power transformer has delta/star windings, which introduce
a phase shift between primary and secondary currents, a star/delta arrangement of the CT
secondary windings is necessary to achieve balance in the secondary circuit. Both these
conditions are explained in Chapter 6 and particularly in Figure 6.5.
Allowance has to be made, in differential protection schemes for transformers, for the
magnetising inrush currents which flow drily in the primary windings when the transformer
is switched onto the supply; they are not reflected in the secondary windings and therefore
appear similar to primary fault currents, which may falsely operate the differential protection.
The simplest solution is a short time delay in the relay - an induction disc relay may be
‘used - although there are more subtle solutions available in cases where a delayed response
is not desirable.


It should be noted that, although restricted earth-fault protection, as described in Chapter 5,

will operate satisfactorily for internal solid-earth faults on most parts of transformer secondary
windings, a high-impedance fault to earth may not give rise to sufficient fault current to
operate the relay, even though it is given a light setting.

of Fault
(% of Winding)

-r Relay Setting


Another point to be noted is that, if the fault occurs near the star-point, the voltage at that
point may not be sufficient to cause a fault current high enough to operate the relay. This
situation is shown in Figure 10.2. Thus, although restricted earth-fault protection is usually
installed for transformer secondaries, it cannot be regarded as one hundred per cent certain
to operate.


In addition to the protection listed above, whdse purposes have already been expldined,
there are the following additional features special to transformers:

10.4.1 ‘Qualitrol’ Protection (Q)

Qualitrol protection is fitted only on sealed transformers such as those used on offshore
installations. It is a proprietary device fitted at the top of the transformer. It detects over-
pressure within the transformer and, if it exceeds a certain preset level, trips both HV and
LV circuit-breakers simultaneously through a flag relay (FG) and the lock-out relay (TH).
The device has a spring-loaded discharge disc to relieve pressure immediately if it builds up
too quickly.

On large oil-filled grid and similar transformers internal pressure is normally relieved into
the conservator. Nevertheless it is customary to fit such transformers with a pressure relief
diaphragm on the tank top.

10.4.2 Buchholz Relay

Although termed a ‘relay’, this is in reality a mechanical device named after its inventor.

Gas Trap .

Mercury Switch

II --:T
v Counterweight

Drain Plug


The device is fitted in a horizontal section of the pipe running between the main tank and
the conservator in large oil-filled transformers.

It consists of two parts as shown typically in Figure 10.3, a gas trap and a surge section. If
an insulation weakness begins to develop under oil in any part of the transformer winding,
small discharge currents start and create tiny bubbles of gas. As the breakdown slowly
progresses, the rate at which gas is evolved increases. The bubbles rise slowly to the tank top
and pass on, through the connecting pipe, towards the conservator. On the way they pass
through the Buchholz relay and are caught in the gas trap. Over a period of time enough gas
is accumulated to cause the oil remaining there to have a free surface, and a float gradually
lowers until, on reaching a preset level, it actuates a mercury switch. This is usually arranged
to give an alarm, since the process is gradual and has not yet reached breakdown stage
calling for immediate disconnection.

The lower part is the surge section. Here a vane is suspended vertically across the flow of oil
between the tank and conservator and is held firmly against a stop by a counterweight.
Normally the oil flow is very slight, depending only on temperature changes in the trans-
former, and the vane does not move. But if there is a complete electrical breakdown in any
winding under the oil a power arc will develop inside the tank, causing an expanding, high-
pressure bubble of oil vapour round the arc. This will rapidly displace oil from the tank into
the conservator, causing a surge of oil past the vane, which will swing against the action of
the counterweight and actuate another mercury switch. Because an actual breakdown will
i,~ have occurred, this contact is always arranged to trip the supply side of the transformer.

The above describes the operation of a typical Buchholz relay in principle. Different manu.
facturers have added many refinements to this basic design.

10.4.3 Interlocks and Intertrips

interlocking and intertripping is provided between the HV and LV breakers. If the HV

oreaker opens for any reason, whether tripped by a fault or operated manually, the LV
breaker (if closed) trips in sympathy and cannot be reclosed until the HV breaker has been
closed first.

It will be seen from Figure 10.1 that a fault, whether on the HV or LV side, operates
through the lock-out relay and trips both the HV and the LV circuit-breakerssimultaneously.
This is to ensure that, after such a fault, not only is the transformer isolated from its normal
supply side but also that it cannot be back-fed from the LV side.

The intertrip acts as a back-up for this, but it is also needed to ensure sympathetic opening
.* of the LV breaker when the HV breaker is opened by hand, as distinct from by a fault.

10.4.4 Coolant Level

A sight-glass is provided to check the coolant level within the tank of a sealed transformer.
The level varies with temperature, and allowance must be made for this; level marks for
15°C and 45°C may be given.

Conservators of large oil-filled transformers usually have a sight-glass to indicate oil level.

10.4.5 Sealing Monitor

A centre-zero pressure/vacuum gauge may be provided to indicate pressure in the vapour

space over the liquid coolant of a sealed transformer, The transformer~is filled to a level
marked on the sight-glass and sealed at a specified temperature - say 45’C. In service any
variation above or below this temperature, due either to change of ambient temperature or
to transformer loading, causes the liquid level to fall or rise slightly and a consequent small
vacuum or pressure to be indicated on the gauge.
If the pressure shown by the gauge moves over a range less than its normal one, it may
indicate a failure of the tank sealing allowing air to be ‘breathed’ in and out. Such asituation
should be investigated.

10.4.6 Overtemperature~Protection

Whereas winding temperature can be monitored by normal temperature-sensing devices as

described in Chapter 8, a special arrangement is sometimes used in large liquid-filled trans-

In this application Negative-Temperature-Coefficient (NTC) thermistors are used in tempera-

ture-monitoring instruments. They are suspended in the oil in a housing with a heating
element and employ the technique of ‘thermal im~aging’ as described in Chapter 8, para. 8.2.


On all offshore and onshore installations the transformer secondary star-point is usually
solid-earthed either through a link or through the neutral bar of the LV switchboard which
it feeds as a 4-wire system. The earth connection can be isolated when desired (for example
when megger-testing-the secondary) by means of a link at the switchboard, or, where the
earth connection is made through a link in the 3-pole circuit-breaker, by withdrawing and
isolating the circuit-breaker unit itself.

Care ‘must be taken, after opening an earth link for any reason, to ensure that it is replaced
immediately after the test. The whole protection of the transformer may depend on it.

The safety earthing of the transformer tank, as distinct from its windings, is dealt with in
the manual ~‘Electrical Safety’.




A typical interconnector protective system is shown in Figure 11.1. It is for the HV inter-
connector between the 6.6kV Production Switchboard and the 6.6kV Main Switchboard on
an offshore installation. Normally the production generators are shut down and the main
generators feed both switchboards, the main switchboard direct and the production switch-
board through one or both of the two interconnectors - that is, the normal power flow is
from right to left.

6.6kV Production Switchboard 6.6kV Main Switchboard


1 I
L _____ ---------(---------------J


in the figure will be seen the fol,lowing features already described in Chapters 3 to 8:

- Overcurrent (2-element) inverse-time (20CIT) protection fed from a set of

protective CTs. The relay trips the production-end circuit-breaker through a
lock-out relay and thence the main-end breaker by intertripping (see para.
11.2). The time and current settings will be determined by the overall di5-
crimination plan. It will also trip the main end direct.

- Earth-fault inverse-time (EIT) protection fed from the same set of protective
CTs as above. The relay trips the production-end circuit-breaker through the
lock-out relay (TH) and then the main-end breaker direct, and by intertripping.

- Differential (DIF) protection between ‘both ends of the cable. ~Fed from two
sets of special Class X CTs, those at the main end being connected by pilot
wires to those at the production end. The relay instarttaneously trips the
breakers at both ends together through the lock-out relay.

- Undervoltage (UV) protection at the product~ion end. Fed by a p~rotective

VT from the production busbars, the relay trips the production-end inter-
connector breaker when production voltage fails; then the main-end breaker
by intertripping.

- Lock-out hand-reset relay (TH).

- Three ammeter instruments, one each at th,e production and main switchboards
and one in the Electrical Control Room. All fed from a single measurement

Differential protectio~n is provided in the case of an interconnector cable to ensure that, if

there is a fault in the cable itself, there will be instantaneous isolation at both ends even
before the overcurrent protection at eithe: end operates to disconnect other consumers. It
will also minimise any fire risk at the fault point, especially in hazardous areas, by removing
the power sources in the quickest possible time. Because differential protection is insensitive
to through-faults and takes no part in the discrimination pattern, the relay is given a very
light setting.

If the busbars of the production switchboard aredead,the undervoltage tripand the interlock
(see para. 11.2) would prevent the circuit-breakers at both ends of the interconnector from
being closed. To allow a dead production switchboard to be energised from the main switch-
board, a key-operated switch is provided on the Electrical Control Panel which defeats the
undervoltage trip while the interconnectors are being closed. By using this switch the
production-end interconnector breaker is first closed by hand; this lifts the interlock and
allows the main-end breaker to be closed. Power now flows from main to production
switchboard, and the latter becomes energi,sed and so lifts the undervoltage trip. The key-
operated ‘W inhibit’ switch must now be restored to its normal position and the key


An intertrip facility is provided so that, if the production-end breaker opens for any reason,
whether tripped by a fault or, through the undervoltage relay, by failure of power or opened
manually, the main-end breaker (if closed) trips in sympathy.

An interlock facility ensures that, if the production-end breaker is open for any reason, the
main-end breaker cannot, be closed until the production-end breaker has first been closed.

It will be seen from Figure 11 .l that a fault on the cable operates through the lock-out relay
and trips the breakers at both ends simultaneously. This is to ensure that after such a fault
the cable is isolated from both ends and cannot be energised from either switchboard.

The intertrip acts as a back-up for this, but it is also needed to ensure sympathetic opening
of the main-end breaker when the production-end breaker is opened by hand, as~distinct
from by a fault.

It should be noted that with the interlocking arrangement described, if it is desired to feed
power from the main to the production switchboard (i.e. from right to left in Figure 11 .l,
which will eventually be the normal arrangement), it is necessary first to close the produc-
tion-end breaker, using the ‘UV Inhibit’ switch as described, and then the main-end breaker.
Final control will thus always be from the main switchboard.

There will be minor variations of the system described for different interconnector applica-
tions, but the basic requirements remain the same, namely overcurrent, earth-fault and
differential protection, with intertripping and interlocking.


The above example is based on a typical offshore, high-voltage cable interconnector where
cable runs of a few hundred yards are involved. Cables in onshore installations, though
somewhat longer, would, if it were-required to protect them, be dealt with by similar means,
but perhaps with different detail.

By shore standards (other than in a limited installation)~platform cable rut&are very short.
In Supply Authorities’ distributio.? networks they may be many miles long, and beyond that
lies the still greater network of cables and overhead transmission lines. In these cases such
methods as differential protection, for example, involve long runs of pilot connections with
the special problems that they bring.

instead, very sophisticated protection methods are used which are beyond the scope of this
manual. They are mentioned only so that the reader~may be aware of this vastly greater field
of protection.





Virtually every offshore motor, whether operating on high or low voltage, is of the squirrel-
cage induction type, and all except the very largest are direct-on-line started. Except in the
living quarters all motors are totally enclosed, fan cooled (TEFC), the largest being supple-
mented by water cooling of the circulating air. Exceptionally, the variable-speed drives for
drilling, drawworks and mud pumping are d.c. motors.

6.6k V Switchboard

440/l 1ov 1 lo\: d.c.

-_ stop +$----
StOp c;



E Earth Fault
NPS Negative Phase Sequence
Trip, Hand Reset (Lockout)
TTlOC Thermal Overcurrent
uv Undervoltage


Most motors are switched on and off by contactors, and in most cases it is necessary to
provide back-up fuses lo protect the contactor itself, as well as the circuit, against~ high-level
faults such as short-circuits which are beyond the breaking capacity of the contactor. The
fuses have special ratings both toensure discrimination in relation tothe trippingcharacteristic
of the contactor and to prevent their being blown by the starting current of the motor.
They do not protect the motor against overloads; this is a function of the contactor.


Like any other piece of electrical plant a motor, whether HV or LV, must be protected
against overcurrent and usually against earth fault. In larger motors overtemperature may
also need to be monitored in both the cooling air and the windings themselves.

Overcurrent in a motor can be caused by any of four conditions:

- mechanical overloading
- short-circuit
- stalling
- single-phasing.

Other forms of abnormal motor fbnctioning include:

- earth fault or leakage

- undervoltage.

All the above will be considered in turn below,

Figure 12.1 shows the control and protection diagrams of motors which may be regarded as
typical of the types actually met in service. On the left is a LV motor and on the right a
HV type.


A motor may be overloaded mechanically by either overloading the driven end (e.g. pump
or compressor) beyond its rating or by some internal mechanical malfunction such as a stiff
bearing. Either may cause a rise of active current above the full-load rated current of the
motor. Mechanical overloading is probably the commonest cause of overcurrenrin a motor.

The motor is protected by an inverse-time overcurrent device which will cause the contactor
to trip if the overcurrent is sufficiently high and persists. The device usually takes the form
of a thermal element in each phase, either directly or CT-operated. It has an inverse-time
characteristic which is more nearly matched to the thermal behaviour of the motor itself
than that of the inverse-time electromagnetic overcurrent relay described in Chapter 3. It
must allow the large starting current (up to five times full-load current) to flow during the
run-up period without operating, but it must trip the motor if even a small overcurrent
persists for a longer period, A typical setting of such a device would be 120% full-load
continuous current with the appropriate time setting. F.or short starting times the inverse-
time characteristic must be such that the starting current and run-up time are taken into
account. In this respect it must be remembered that high-inertia loads such as a motor-
L generator set or a compressor take much longer to run up than, say, a centrifugal fan.

For all LV motors and a few HV the inverse-time device is thermal, as described above. For
the smaller LV motors it is in series with the motor itself (see Figure 12.1 (a)), but for the
others it is a separate relay operated through CTs. For most HV motors on the later platforms
however, the device is wholly electronic but with a similar characteristic; it too is a separate
relay, CT-operated (see Figure 12.1 (b)); Where these relays are separate the overcurrent
device is combined with certain other features into a single ‘Motor Protection Relay’ which
is further discussed in para. 12.10. Both types of overcurrent protection will be found in
shore installations.

A characteristic of inverse-time relays which is particularly noticeable in thermal relays, and

which has to be taken into account in allowing for starting current, is ‘overshoot’ (or ‘overrun’).
This means that if the relay is energised with something more than its minimum operating
current it may close its contacts even after the current has subsequently fallen below the
operating level. For example, a motor could be’ tripped after it had safely started and
reached~full speed, even though the relay had not operated during the starting period. This
can have a considerable effect on the discrimination that can be achieved between starting
and overload currents, unless complications are added to the protection scheme.

Whereas the contactor with its inverse-time overcurrent device (thermal or electronic)
provides overload protection for the motor, such contactors cannot in general clear a fault
of short-circuit proportions. For this they must be backed up by series HRC fuses.

When used with motors such fuses must have special characteristics. They must have a
continuous rating which will allow them to pass the full-load current of the motor con-
tinuously, and they must also allow the considerably greater starting current to pass for
the period of the run-up time without melting the fuses. These special motor fuses are
described in Chapter 3, para. 3.3.5, ‘Motor Protection’.


A short-circuit in a motor circuit will cause a severe overcurrenr. One of the more vulnerable
places to short-circuit in a motor is the incoming cable box where a too-small radius or
imperfect jointing could lead to weakness. As many process motors are located in hazardous
areas, it is essential to clear the circuit in the quickest possible time under these conditions.

The overcurrent produced by a short-circuit fault will operate the inverse-time device in a
relatively short time, but in general not short enough to cause a trip before the fuse blows.
Indeed it is important that the contactor should not operate as it is not rated to clear fault
currents. Where HV or LV motor feeders are provided with back-up fuses, it is these that
will blow first and clear the fault very quickly (see Chapter 4 ‘Discrimination’ and especially
Figure 4.2). Where very large motors are fed through circuit-breakers and there are no back-
up fuses, the inverse-time protection will be backed up by a high-set instantaneous overcurrent
relay element (see Chapter 3) to trip the circuit-breaker instantaneously without waiting for
the inverse-time elem%it to operate. Its setting will be swell above the ~motor overload and
starting current levels.


Earth faults, especially those occurring in solidly earthed systems, will also give rise to severe
overcurrents in the affected phases. They may be dealt with by an earth-fault relay (E)
which trips the contactor. With small motor starters where the thermal overcurrent devices
are direct-connected, the relay is usually energised by a core-balance CT through which the
three cables pass (see Chapter 5, para. 5.2). This arrangement is shown in Figure 12.1 (a).
With larger motors where there are CTs, these may be used to provide an earth-fault signal
by their spill current. The earth-fault relay may be separate or, as shown in Figure 12.1 (b),
may be part of the composite ‘Motor Protection Relay’, which is further described in para.

The same consideration will apply to earth faults in HV motors. Nearly all HV systems are
resistance-earthed, which limits earth-fault currents to a low level that will not actuate the
fuses or high-set instantaneous relays. Here also Bn earth-leakage relay operated through CTs
is necessary, but the fault current is then well within the breaking capacity.of the contactor.


Stalling can occur if the motor becomes~heavily overloaded -for example by a mechanical
seizure of a bearing or of the driven element, or it may be unable to start against an excessive
load. In all these cases the stalled motqr draws its ‘locked rotor’current (that is, its maximum
starting current) as long as the supply remains connected, and severe overheating results.
The condition is aggravated by the lack of ventilation while the rotor is stationary. There is
also a temptation to make repeated attempts to start if unsuccessful the first time.

More rarely, stalling can occur if the whole power network goes unstable and begins to run
down, causing all motors throughout the system to lose speed. If the system recovers while
the run-down is proceeding, all the motors in the system will find themselves running at a
large slip and all taking nea~rky their full starting currents. The combined effect on the
generators of all these simultaneous starting currents will be to depress the system voltage to
such a level that some of the larger motors may not develop sufficient torque to recover
against their loads. They will then continue to run down and st;.ll. This is sometimes called
the ‘Patrickson Effect’.

The long period of drawing ‘starting current’, though not actually starting, will appear as a
normal overcurrent and should, in principle, be dealt with by the motor thermal overcurrent
protection, but difficulty arises when this protection has to discriminate between normal
starting current (which is present during the run-up time but then disappears) and the stalled
motor current (which persists).

This problem is particularly difficult if the run-up time of the motor is of the same order as
the motor stall (or locked rotor) withstand time. It is still worse in the case of high-inertia
loads, where the run-up time may well be longer than the stall withstand time. For these
situations ‘stall relays’ are sometimes used, especially with HV motors.

Stall relays are of two types: one using electric sensing of the motor starting current, and
the other using detection of actual rotation. The former uses a current-sensitive element and
a timer. On a normal start the current-sensitive element energises, but the timer prevents its
causing a trip unless the normal run-up time has elapsed. On a genuine stall the timer will
trip the motor after run-up time has expired if the overcurrent is still present. This type of
relay is often fitted with a ‘thermal memory’ which prevents a restart until sufficie~nt cooling
time has elapsed. The stall relay is sometimes included with other elements in a combined
‘Motor Protection Relay’ (see para. 12.10).

The other type of relay uses a shaft rotation detector. This form is most suitable when the
motor safe stall withstand time is very close to the motor run-up time, or even less as in
the case of the very large gas t-e-injection motors on certain offshore installations. The
rotation method is a more accurate indication of a stall condition. However it does not
incorporate any ‘memory’ to protect against quick restarting, and it must be used in conjunc-
tion with some form of lock-out protection.


A problem special to 3-phase motors is single-phasing. Any such induction motor needs
three phases to produce its rotating field and to provide the necessary starting torque, but
once running, the removal of any one of the phases will not necessarily stop it. It will
however reduce the driving torque and will also increase the current in the two remaining
phases. If the motor were already well loaded it would eventually trip on sustained over-
current. If, however, the mechanical loading on the motor were not too high, there might
still be sufficient torque to drive the load. Also the current, although increased, might still
not be enough to actuate the inverse-time overcurrent relay set, typically, to 120% full-load
current. The situation could therefore pass unnoticed except for a high-pitched whine from
the motor, and no actual harm would result.

There is a much greater risk, however, when attempting to~start. If the single-phasing had
occurred while the motor was running partially loaded and had not been noticed, the motor
would have been stopped ~normally when the operation was complete, but still in its single-
phased state. If later an attempt were made to restart it, there would be excessive starting
current but no starting torque, and it would remain stationary. As the overcurrent relay
is set to allow adequate run-up time, this situation could persist for a dangerously long time
with no ventilation in the stationary motor. Still worse, the operator might make several
attempts to start, and each time the relay would reset and allow full run-up time afresh.
Eventually the motor would probably burn out before the overcurrent relay disconnected it.
Therefore-~.if a motor fails to start after two attempts, the operator must make no further
attempt to start it until the cause has been found and corrected.

Shell have laid down a general instruction that motors over 50kW are only suitab~le for two
successive starts (even if successful) when cold. Another one-start attempt is allowable after
a cooling period of 30 minutes atstandstill. No more than three starts may be attempted~in
any one, hour.

Damage to a motor by single-phasing is caused by overheating of the windings due to the

prolonged excessive currents in two of the phases. it is normally prevented by the protection
system. In the case of small motors provided with thermal overcurrent protection, the three
thermal elements are mounted in such a way that unequal heating produces a differential
movement which causes the contactor to trip.

With some HV and larger LV motors where protection is through current transformers and
where thermal overcurrent protection is used, th,e single-phasing protection is’ provided by
these same overcurrent elements where unequal heating produces the differential movement.

With motors which are protected by electronic relays the device includes a special element
which detects the single-phase condition, whether the motor is running or being started. It is
referred to as a ‘Single-Phasing’ or ‘Negative-Phase-Sequence’ (NPS) relay.

Single-phasing can be caused by the blowing of one of t,he three back-up fuses, by the
possible welding of contactor contacts or, less probably, by the open-circuiting of one line
due to damage or vibration. It will not occur however with HV fuses where they are fitted
with external striker-pin contacts which trip the contactor if any one of them blows.


Consider a balanced 3-phase system in which the phases rise in the normal sequence
R-Y-B-R. This is indicated by the vector diagram of Figure 12.2(a).

I 4





Consider next another balanced 3-phase system, but one of different magnitude and in
which the phases rise in the reverse order R-B-Y-R, as shown in Figure 12.2(b). The
phase relation of the second to the first system is quite random.

If voltages corresponding to both systems are applied simultaneously to a set of busbars, the
voltage of each bar_woul$ be the vector sum_of the two voltages applit& to>: that is, red
voltage would be RI + R z, yellow voltage Y, + Y, and blue voltage 5, + B2. This vector
addition is done in Figure 12.2(c), giving resultant voltages RT yand L7?

It will be seen that R, Y and 5 are now of different magnitudes and are no longer spaced
120” apart. This resultant system is still 3-phase, but it is now unbalanced.

The system R, Y, 5, in which the phases rise in the normal order is known asa ‘positive
phase.sequence’, and the system R2 Bz Y, in which the phases rise in the opposite order is a
‘negative phase sequence’. It should be particularly noted that, although the negative-
sequence vectors come up in the reverse order, they are still considered to rotate in the
conventional anti-clockwise direction, the same as the positive-sequence vectors. (if they did
not, it would not be possible to combine them as has been done in the lower diagram.)

It is not difficult to see how a pair of different, but balanced, positiveand negativesequences
combine to produce an u,nbalaticed system. It is far less easy to see the converse: that any
unbalanced system may be resolved into two balanced systems, positiveand negativesequence.
This can be proved mathematically, but is not included in this manual.

In Figure 12.2 the negative-sequence system was different in magnitude, and its angular
position relative to the positive-sequence was quite arbitrary. The negative system might
have been assumed to be in any other angular position, and the combined unbalanced set
would then be different in each case.

A special case is when both positive and negative systems have~the same magnitude but the
negative-sequence red phase is diametrically opposite (1 go”, or anti-phase) to the positive, as
shown in Figures 12.3(a) and (b).





If these two are combined in the same way as was used in Figure 12.2, it produces the
vector diagram of Figure 12.3(c). It will be seen that the two reds cancel out, and the two
blues and the two yellows combine to produce resultants at 9 o’clock and 3 o’clock respec-
tively - that is, equal and opposite to each other.

If these are considered as current vectors, this is just the situation which exists in a motor
whose red phase supply has become open-circuited (Figure 12.3(d)) but whose other two
phases are intact. In that case no current ~flows in red phase, but the current flow in yellow
phase returns via the star point through blue phase - that is to say, the yellow and blue
currents are equal and opposite, exactly as indicated in the vector diagram.

It can be inferred therefore that a single-phased condition, which is an extreme case of
unbalance, can be represented by two balanced systems, one positive and one negative, both
equal in magnitude and where in the negative system the-phase which has been open-
circuited is anti-phase with its positive colour.

It will be seen that the currents represented by the thick horizontal lines in Figure 12.3(c)
are greater than the phase currents of the original positive-sequence system by a factor of
d/3:1 - that is to say, when a running motor starts to single-phase, the line currents in the
two remaining phases increase by more than 70% over the original currents being taken just
before the event. In fact they increase even more than this. A single-phase condition also
causes some reduction in driving torque, and the motor will slow down slightly so that its
torque will rise again to meet that of the driven load. This increase of slip results also in
increased current, and in practice the currents in the sound phases will about double when a
running motor starts to single-phase.

This doubling of current, if the motor had been fully loaded at the time, would be more
than enough to operate the thermal overcurrent protection and trip it. If the motor had
been only half-loaded or less however, even the doubled current would not be sufficient to
operate its overcurrent protection. Therefore the normal protection cannot be relied’upon
in all cases to trip a motor if it should lose a phase while running.

Although the concept of negative phase sequence is purely mathematical, and one is tempted
to say that it does not really exist, its presence can nevertheless be detected by a relay and
used to indicate unbalance of currents in a 3-phase system. How this relay operates is
explained below. In the extreme ‘case a large negative-sequence component will indicate an
open-circuit condition in one phase. The relay is called ‘Negative Phase Sequence’ (NPS) and
is particularly used to protect motors against single-phasing. As already explained, this
condition is not serious if it occurs while a motor is running, but it could be very serious
indeed if an attempt were made to start a motor while in a single-phased state.

The negative-phase-sequence relay uses a simple network and a voltage-sensitive relay,

connected as shown in Figure 12.4(a).

Current transformers in R and B phases feed respectively a resistance RI and a resistance

R, with a reactance X in series. R, and X are chosen so that the circuit has a ower factor
cos @ = 0.5 (so that 6 = 60” lagging) and their combined impedancez (= d*) RZ has
the same ohmic value as RI

Then the current I, flowing in the red CT secondary passes through the resistance R,,
causing a voltage I, R, (or V,) across it in phase with II.

The current I, flowing in the blue CT secondary passes through the impedance Z, causing
a voltage /,Z (or V,) across it. Since the !a circuit is an inductive one, the current /a lags
60” on the voltage across Z, or, put another way, the voltage across Z (= /,Z) leads 60”
on the current I,.

If we look at the positive-sequence current vector diagram of Figure 12.4(b), the voltage I/,
(= I, R,) across R, is in phase with the red current, whereas the voltage V, (= I,Z) across Z
leads 60” on the blue current. Since R, and Z have the same ohmic value, thee two FItages
are equal and opposite. The voltage across the relay is the vector sum of V, and Va, and
these two cancel out, so the voltage across the relay is zero. Therefore the relay is insensitive
to positive-sequence currents.

Consider now the corresponding effect on the negative-sequence currents - the vector
diagram of Figure 12.4(c).


‘B Qh3



As before VI (= /,R,) is in phase with II, and 1/Z (= /,Z) leads 60” on the blue current /2.
The voltage vector Vz ic now at ‘2 o’clock and-no longer cancels out V, Since the voltage
across the relay is the vector sum of V, and V,, this resultant voltage is at ‘1 o’clock’ as
shown in Figure 12.4(c) and has a definite value. Therefore the relay is sensitive to n~egative-
sequence currents, and, if the resultant voltage is greater than ~the setting of the relay, the
relay will close and trip the cbntactor and lock it out.

Thus an NPS relay has been devised ,which will ignore the positive-sequence currents which
are normally present but will detect the presence of negative. In particular it will give
protection against a single-phase condition and will prevent, starting a motor while i” this

,.The foregoing description deals with the concept of balanced positive- and negative-sequence
currents as an aid to dealing-with single-phasing. In electrical engineering practice they have
a far wider application, which is beyond the scope of this manual.

There is a third type of balanced system called ‘zero-phase sequence’. It comes into being
when there is current in the neutral conductor of a 4.wire system, or with an earth fault on
an earthed 3-wire system (which effectively provides a fourth path). Zero sequence is not
dealt with in this manual but comes into the wider application referred to above.

The three ccmcepts positive-, negative- and zero-phase sequences are collectively referred to
as ‘Symmetrical Components’.


The thermal overcurrent, earth-fault and single-phase prote6tive~devices for HV and the
larger LV motors have been referred to previously as if they were separate elements. In
practice they are usually combined into a single ‘Motor Protection Relay’ which contains all
,these elements. Its form differs between manufacturers. One widely used type, made by P.B.
Engineering (commonly called ‘P & B Golds’), is shown in Figure 12.5; it uses flags to
indicate which element has operated. Another type is wholly electronic and indicates by
lamu. In this type the single-phasing condition is detected by an electronic relay.


In some installations a motor protection relay may also incorporate a stall relay element -
see para. 12.7.


Overtemperature protection is sometimes used, in addition to the thermal overcurrent pro-

tection afforded by the Motor Protection Relay, to safeguard the windings of a motor.

Three main types of temperature sensor are used:

- Thermocouple
- ,Resistance Temperature Device (RTD)
- Thermistor.

The principles of these methods are discussed in the manual ‘Electrical Control Devices’.

The sensing elements arc normally embedded in the winding insulation, usually in the over-


In any distribution system involving motors it is important that, if system voltage should be
lost, all motors should be disconnected so that they do not all restart together when the
system voltage returns. They must be carefully restarted in a controlled sequence.

All LV and some HV contactors are closed and held-in by their operating solenoids. It
follows that, if the operating voltage is lost or falls below a certain value, the contactor will
drop off and will not. retlose until given a positive closing signal. Such a contactor is said to
have an ‘inherent undervoltage’ facility and needs no other undervoltage protection.

Some HV contactors however are, like circuit-breakers, of th,e latching-in type and are
operated from a separate d.c. closing and tripping supply. If main voltage is lost they remain
latched and will therefore not trip. Such contactors are fitted with separate protection in
the form of an undervoltage reiay connected through a VT across the mains. When mains
voltage fails the relay trips the contactor through its tripping circuit, and it will not reclose
when mains voltage is restored until given a positive closing signal.

After motors have tripped on undervoltage certain installations which have been provided
with ‘re-acceleration units’ automatically restart their motors in a pm-arranged sequence
when voltage has been restored. Re-acceleration is discussed in the manual ‘Electrical
System Control’, Chapter 8.2.


A major problem in motor design and protection is to ensure that the starting current can
flow fur long enough to accelerate the motor without bringing out the motor’s over-current
protection, while at the same time not impairing the close protection given to the motor
while it is running.

For this purpose a time-quantity ‘t, ’ is considered. This is defined as the time taken for the
motor’s windings, while carrying the starting current I, continuously, to be further heated
from the maximum temperature reached in rated service and in a maximum ambient tem-
perature to the limiting allowable temperature. The te time of the motor should be given on
its nameplate.

20 min
10 min
8 min
4 min
2 min

‘42 -.
I II II I * la/IN ratio
1 2 3 4 6 8 10

The thermal relay for protecting a squirrel-cage motor should be selected so that the tripping
time read from the thermal relay’s time/current characteristic using the /,/I, ratio (ratio of
starting to normal rated current) is not larger than the motor’s stated tE time.


A motor has a rated current of 80A and a starting current of 480A. The /,/I, ratio is
therefore z = 6.0. The tE value on the nameplate isgiven as 16 seconds. Is thisacceptable?

The characteristic of the particular motor thermal protection relay is shown in Figure 12.6,
where the tripping time (t) is plotted against the /,/I, ratio. in this example the ratio is
6.0, so that the tripping time is 8 seconds. This is well within the stated tE time of 16 seconds
for the motor and is therefore acceptable. It will not trip when starting or restarting.

Had the tripping time been greater than tE, a relay with a different characteristic would
need to be chosen.

It must be realised that, if the motor had been at maximum temperature when first restarted,
it would be at a still higher temperature immediately after the start. A second or further
start would then not necessarily be permissible until after a cooling-down period. The closer
the run-up time is to the tE value, the more important this becomes.




1. Name five principal reasons for having a protection system.

2. Name four possible consequences of a fault condition.

3. What are the’particular dangers associated with arcs?

4. What are the three principal disconnection devices on which protection depends,
and what are the differences in how~they function?

5. What is meant by the ‘fault level’ at a point in a system, and in what units is it

6. What is the most important rating of a circuit-breaker in assessing its suitability for
a particular duty?

7. What is the effect of large motors on a system undergoing short-circuit? How are
they taken into account?

8. What is the impedance of an a:c. generator that determines the fault level at its
terminals, and why?

~9. If the subtransient reactance of a generator is 25%, how many times rated full-load
current will the initial short-circuit current be?

10. Calculate the rms symmetrical short-circuit current at a point in a 3-phase system
where th,e fault level is 12MVA and the normal voltage is 440V.

11. The network opposite shows two

generators supplying a number of LV
motors totailing 800kVA through a
transformer. The subtransient re-
adtances of the generators ,and the
reactance of the transformer are
indicated. Calculate the fault level at
the LV board:

(a) ignoring the motor contribution,

(b) taking account of the motor


\ /
Total 800kVA

12. In the network of Q.11 the HV board operates at 6.6kV and the LV board at
440V. Would you expect the short-circuit (rms symmetrical) current at the LV
board to be less than, or greater than, the short-circuit current at the HV board?
Give the current figures In kA (rms sym,metrical).

13. What is the significance of the term ‘prospective symmetrical rms fault current’?

14. What is overcurrent protection primarily intended to protect?

15. What are the types of overcurrent protective devices most likely to be met?

16. What, approximately, is the operating time of a high-set ‘instantaneous’ overcurrent


17. A circuit is fitted with inverse-time overcurrent relays fed from current transformers
of ratio 500/1A. To protect the circuit against overloads the protection must
operate within 1.2 seconds when the current exceeds 3 OOOA. What plug setting
would you give to the relays:

(a) if the maximum continuous~load current is 300A,

(b) if the maximum continuous load current is increased to 600A without change
of CUT?

(Use the relay setting curves of Chapter 3, Figure 3.4.)

18. What is the purpose of a ‘trip, hand-reset’ or lock-out relay in a protective system?

19. What is the significance of ‘HRC’ in relation to a fuse, and how does its behaviour
make it particularly effective for short-circuit protection?

20. How does the operation of fuses with moderate overcurrents differ from that with
high prospective current? Why does fuse protection in d.c. ti.zuits require careful

21. What do you understands by ‘fusing factor’? What range of values does it have?

22. What is the purpose of a trigger fuse?

23. What is normally the purpose of discrimination in overcurrent protection?

24. What is the basis of discrimination between two fuses in series in a short-circuit,
and how, roughly, would discrimination be ensured?

25. Why should two MCCBs never be used in series with each other?

26. Can an MCCB and a fuse be used in series? If so, under what conditions?

27. Why is discrimination between a contactor and its back-up fuse important?

28. Sketch three methods of applying (unrestricted) earth-fault protection to a 3-phase


29. Make a sketch showing how earth-fault protection can be combined with over-
current protection in a single relay case and fed from the same CTs.

30. Describe an Earth Leakage Circuit Breaker (ELCB). What is its purpose, and where
is it principally used?

31. How is earth leakage in an unearthed d.c. system detected?

32. What is the principle of restricted earth-fault protection in a system which is

earthed at the switchboard neutral busbar?

33. What is the principal advantage of restricted earth-fault protection?

34. What is the advantage of full differential protection in comparison with restricted
earth-fault protection?

35. Name two possible reasons for using undervoltage protection.

36. What loads are most likely to prevent the recovery of a system from a temporary
drop in voltage?

37. What is the essential difference between NTC and PTC thermistors used for tem-
perature sensing?

38. Why is differential protection particularly useful for generators?

39. Against what malfunctions would you expect a large generator to be protected?

40. What kind of relay is used for reverse-power protection?

41. Why is a special means of detection necessary to detect excitation diode failure in a

42: What is the purpose of generator field suppression?

43. What is the purpose of a neutral earthing resistor associated with a generator?

44. How would a short-circuit on the LV busbars fed by a transformer be cleared?

45. How is a differential protection system balanced int,he case of a transformer?

46. Why is a short time delay (or some more complicated measure) necessary in differ-
ential protection for transformers?

47. What are the functions of:

(a) Qualitrol transformer protection,

(b) a Buchholz relay?

48. What interlocking and intertripping is provided between a transformer’s HV and

LV switchgear? What is their purpose?

49. Why is differential protection applied to high-voltage cables?

50. What type of motor is mostly used in Shell installations?

51. Against what types of malfunction would you expect a large motor to be protected?

52. What is the most likely cause of overcurrent in a motor?

53. Why do motor fuses have special ratings?

54. What is ‘overshoot’ in a thermal relay?

55. Why may stalling protection be necessary for a motor?

56. What is the purpose of negative-phase-sequence protection for an induction motor,

and in what respects is it superior to overcurrent protection?

57. How can current unbalance in a motor be detected in a simple way?

58. What results from the addition of positive-and negative-phase-sequencesymmetrical

components of current:

(a) if they are unequal,

(b) if they are equal in amplitude and in anti-phase in one phase?

59. A P & B Golds motor protection relay includes many protective elements. What
are they, and explain briefly how they function.

60. Why is separate undervoltage protection sometimes necessary to prevent the

restarting of motors upon restoration of an interrupted supply and sometimes not?


(Figures in brackets after each answer refer to the~relevant paragraphs in the text.)

1. (4 To maintain electrical supplies to as much of the system as possible after a

fault has been isolated.

(b) To protect generators and other plant against damage due to abnormal con-
ditions and fault.

(cl To protect the consumer equipment against damage due to abnormal con-
ditions (e.g. overload).

(4 To isolate faulted equipment to limit the risk of fire locally.

(4 To limit damage to the cable system resulting from a fault. (1.1)

2. (4 An arc.

(b) Overheating of conductors.

(4 Mechanical overstress on conductors.

(4 Disruption of operation. (1.1)

3; (4 Damage to eyes, and burns.

6’) Fire.

(4 Explosion in hazardous areas. (1.1)

4. (4 Circuit-breakers: normally rated to make, carry and break the maximum fault

U-4 Contactors: rated to make and carry maximum fault current but not normally
to break it.

(4 Fuses: can be rated for short-circuit protection but are expendable and have
to be replaced after~operation. This is costly. (1.3)

5. It is a measure of the energy that can be released at the point in a network where a
full ‘bolted’ 3-phase short-circuit may occur. It isJ3 times the product of system
line voltage and the calculated prospective short-circuit current at that point. It is
measured in volt-amperes, or more usually in MVA. (2.2)

6. Its fault level rating or breaking capacity, usually expressed ins MVA. (2.2.2)

7. For a very short period they act as generators and contribute, together with the
main generators, to the total fault current and so to the fault level at the point of
the’system in question. They are taken into account by considering them as extra
generators of the same kVA but with a reactance of 30%. (2.2.2)

8. The subtransient reactance, because itgoverns the most severe condition immediately
following a short-circuit, which is not greatly affected by the level of excitation.

9. 1 + 0.25 = 4; four times the normal full-load current. (2.2.2)

kVA = 12 000
10. I = 15700A
sc = J3kV J3 x 0.44
= 15.7kA (2.2.2)


Base 30MVA

Without Motor With Motor

Fault level 30
= ,.425

= 2lMVA = 24MVA

N.B. 800kVA of m~fors raised to 30MVA base

gives a factor of 30/0.8 = 37.5.
Motors assumed to have ~reactance 3ll%,
which multiplied by 37.5 gives 1125%.


12. From the first part of the answer to Q.ll above, the fault level at the HV board is
30/0.225 = 133MVA. The short-circuit current is therefore = 11.6kA.
d/3 x 6.6

The faulr level without motor at the LV board is 21MVA, giving a current of
= 27.6kA, and with motord3 Fo.44= 31SkA.
d/3 Z’o.44

Ptlthough the LV fault level is lower in MVA, the LV fault current is higher than
that of the HV. ~~_. ~(2.3)

13. It is the rms current that would flow in a bolted short-circuit, after any asymmetry
had disappeared, if not limited by the action of the protection device itself. (2.5)

‘v 14. Heating effect in conductors, and in some cases electromechanical forces between
conductors, due to fault currents. (3.1)

Instantaneous overcurrent relay.
Inverse-time overcurrent~relay.
Inverse and definite minimum time relay.
Electronic relays of all types.
Fuse. (3.2, 3.3)

Approximately 0.2 seconds (200ms), but often less. (3.2.1)

(a) Maximum load 300A

Plug setting = - = 0.6. Choose 0.75 (75%).

Effective current = 500 x 0.75 = 375A. :~

Multiple of plug setting at 3 OOOA = e = 8.

From curves (Figure 3.4), the horizontal time delay line for 1.2 seconds cuts
the vertical ‘8’ line between time multiplier setting curves 0.3 and 0.4.

Choose next higher TMS setting of 0.4, giving a trip in 1.35 seconds.

(b) Maximum load 600A

Plug setting = - = 1.2. Choose I:5 (150%) (1.25A is too close).

Effective current = 500 x 1.5 = 750A.

:lilultiple of plug setting at 3 OOOA = m&L 4.

From curves ~(Figure 3.4), the horizontal time delay line for 1.2 seconds cuts
the vertical ‘4’ line between time multiplier setting curves 0.2 and 0.3.

Choose next higher TMS setting of 0.3, giving a trip in 1.5 seconds. (3.2.3)

18. The trip signal for almost all protective relays trips the circuit-breaker through a
lock-out relay (TH). Unlike most other protective relays, it does not reset itself
when the cause of operation has been removed. When the lock-out relay has
operated it drops a flag, but the breaker which it has tripped cannot be reclosed,
either locally or remotely, until the lock-out relay has deliberately been reset by
hand. (3.2.7)

19. HRC (High Rupturing Capacity) signifiesan ability to interrupt very high prospective
fault currents safely. HRC fuses have a current-limiting action whereby a steeply
rising current can be interrupted very rapid,ly before it reaches its prospective peak
value. (3.3.1)

20. At a moderate current level a number of half cycles of current are needed to melt
the fuse, and the arc is extinguished at a natural current zero. With a high prospec-
tive current the fuse melts before the prospective peak is reache~d and interrupts the
current at a considerably lower value. In d.c. circuits there are no natural current
zeros, and dangerous arcing can occur within the fuse., (3.3.1)

minimum fusing current

21. Fusing factor is the ratio normal rated current

It varies between approximately 1.2 and 2.0. (3.3.4)

22. Some fuses are fitted with a device which releases a trigger when the fuse blows. It
either makes a contact which trips the contactor electrically or strikes a trip bar
which releases the contactor mechanically. Its purpose is to ensure the opening of
all three phases even if only one fuse blows. (3.3.7)

23. To ensure clearance of a fault with the minimum effect on healthy parts of the
system. (4.1)

24. It is necessary to ensure that the minor fuse blows before the major. This means
that the i*t ‘let-through’ energy of the minor fuse must be less than the melting or
pre-arcing /‘t of the major fuse. This is achieved when the normal current ratings
of the major and min~or fuses are in the ratio of at least 3:l. (4.3)

25. Although MCCBs with different trip units would discriminate between each other
with their thermal trips, their electromagnetic trips are instantaneous and would
not discriminate. Therefore there is no certainty that the major MCCB would not
trip before the minor. (4.3)

I&i. YesJhe normal current ratings of the MCCB and the fuses should have a ratio of
at least 2:l. (4.3)

27. The fuse protects the contactor from the danger of attempting to interrupt a
current beyond its breaking capacity (as well as protecting the circuit), and the
contactor prevents unnece+ry and possibly dangerous operation of the fuse on
low overcurrent. , (4.4)

28. As Chapter 5, Figures 5.1 (a), (b) and (c).

29. As Chapter 5, Figure 5.2.

30. In an earthed single-phase system the go any, return wires both pass through a
ring-type current transformer. Under normal conditions the currents through both
wires are equal, and there is no net magnetisation of the CT and hence no secondary
current. If either wire develops a leakage to earth the balance is upset, and the CT
secondary operates the ELCB trip release in under 30ms. Leakage currents as low
as 30mA can be detected.

ELCBs are used principally in domestic and industrial applications to protect

personnel from serious shock if they should come into contact with live metal.
Particularly effective in circuits to which portable apparatus can be connected.

31. By placing a high-resistance potential divider between positive and negative lines
and connecting a voltage-operated relay between the centre tap and earth. If there
is no earth leakage, no current flows through the relay, but if there is, the balance
is upset, the centre ~tap takes up a voltage above (or below) earth atid the relay
operates. (5.4.2)

32. The spill current from the three line CTs is balanced against the current from a
neutral CT. Only an earth fault on the supply side of the line CTs gives rise to a
neutral current that is not balanced by the spill current from the line CTs, and so
the earth-fault relay is energised. (5.5.1)

33. It can provide sensitive protection within a defined zone without having to fit into
the overall discrimination scheme. Without it the equipment in the defined zone
(i.e. upstream of the CTs located in the switchboard), including the transformer
secondary windings, would have no earth-fault protection at all. (5.5.3)

34. It protects against phase-to-phase as well as earth faults within the protected zone.

35. (a) To avoid uncontrolled starting of motors upon restoration of power after a
supply failure.

(b) To avert a threatened collapse of the system by shedding~large motor ,loads

following a disturbance. (7.1)

36. Large induction motors

37. The resistance of an NTC (Negative Temperature Coefficient) thermistor falis with
increasing temperature; that of a PTC (Positive Temperaturecoefficient) thermistor
rises sharply from a low to a high value~at a certain critical temperature. (8.2)

38. It provides instantaneous and sensitive protection against internal faults in spite of
a discrimination scheme in which the generator protection is subject tomthe longest
delay. (9.1)

39. Againstymost of the following:

- Overcurrent
- Earth fault
- Standby earth fault
Reverse power
- Differential
- Field failure
- Diode failure
- Winding over-temperature
- Overspeed

and sometimes

- Underfrequency
- Overvoltage
- Negative phase sequence. (9.1) 9.2)

40. A 2-w~inding induction disc power (wattmetric) relay with built-in or separate time
delay. (9.2.1)

41. A diode failure may not noticeably affect the operation of the excitation rectifier,
and the AVR will usually compensate for it. It is unlikely to be noticed unless a
special means of detectron is fitted. (9.2.2)

42. By suppressing the generated emf it protects the generator against excessive damage
from the internal fault continuing after the generator breaker has tripped on differ-
ential protection. (9.2.8)

43. It limits the earth-fault current in order to minimise damage to the generator and
the severity of any arcing while allowing enough fault current to flow to operate a
protective relay. (9.4)

44. By the overcurrent protection on the HV side. (10.1)

45. By using primary and secondary current transformers with ratios in inverse propor-
tion to the power transformer ratio and with reversed star/delta secondary connec-
tions as necessitated by the power transformer configuration. (10.2)

46. To allow for magnetising inrush current. (10.2)

47. (a) To sense, and if necessary relieve, overpressure in a sealed transformer.

(b) To sense evolved gas or a surge of oil in an unsealed transformer fitted with a
conservator. (10.4)

48. An interlock is provided to prevent the LV breaker being closed- unless the HV
breaker is already closed. An intertrip is provided which causes the LV breaker to
trip if ever the HV breaker trips or is opened manually. This ensures that a trans-
former can never be energised from the LV side. (10.4.3)

49. Because of their susceptibility to damage and the need to minimise the risk of fire,
particularly in hazardous areas. (11.1)

50. Squirrel-cage induction motors. (12.1)

51. Mechanical overloading.

Earth fault.
Overtemperature (sometimes). (12.3)

52. Mechanical overloading. (12.4)

53. To ensure discrimination with contactor characteristics and to avoid blowing due
~.o starting current. (12.4)

54. An effect whereby the contacts may operate some time after its energising current
has fallen below the operating level. (12.4)

55. With high-inertia loads it may be difficult to differentiate between normal starting
current and a damaging stalled condition purely by inverse-time overcurrent
protection. (12.7)

56. To detect the presence of negative-sequence current and in particular of a single-

phased state. In tripping the motor contactor instantly it prevents overheating the
windings in this condition and, most important, prevents any attempt to start.

It is superior to. normal thermal overcurrent protection because the line currents in
a single-phased state do not accurately reflect their heating effect in the motor, and
overcurrent protection alone cannot be relied upon to act quickly enough to save
the motor. (12.8)

57. By means of a thermal overcurrent relay in which tripping can be initiated by

differential movement of the elements, or by use of a negative-phase-sequence
relay. (12.8 and 12.9, Figures 12.4 and 12.5)

58. (a) Unbalanced 3.phase currents.

(b) A single-phase current. (12.9)

59. A P & B Golds motor protection relay contains:

(a) Thermal overcurrent elements in all phases. (These operate by heating of the
elements caused by excessive line current.)

(b) Earth-fault element. (This operates from spill current from the three phases
through a separate trip coil.)

(c) Single-phase ele.ment. (This is a mechanical linkage which is unbalay;,$

unequal heating of the current elements and operates the trip.)

60. Undervoltage protection is provided automatically by non-latching conta&ors but

not by latching contactors and circuit-breakers. These require a separate under-
vdltage~?elay to proCide~‘a positive~~shit’ht trip. (12rW)

Protective Relays Application Guide
- General Electric Co Measurements (GEC)

Protection of Industrial Power Systems

- T. Davies (Pergamon Press)

Protection Relaying for Power Systems

- S.H. Horowitz (John Wiley)

IEEE Recommended Practice for Protection-and Co-ordination of industrial and

,; .Commercial Power Systems
- IEEE (John Wiley)

Electric Power Systems

- B.M. Weedy (John Wiley)

El&tric Fuses
- A. Wright and P.G. Newberry (Peter Peregrinas)

Electrical Engineer’s Reference Book (13th Edition)

- M. G. Say (Butter-worth)

Advanced Electrical Technology

- H. Cotton, DSc (Pitman)

Alternating Current Electrical Engineering (10th Edition)

- Philip Kemp, MSc Tech, C Eng, MIEE (Macmillan)