POWER SYSTEM

FAULT CALCULATIONS
 POWER SYSTEM FAULT CALCULATIONS

Chapter

4
POWER SYSTEM FAULT CALCULATIONS
1. Introduction
A fault is said to occur on a power system when abnormally high currents flow due
to the partial or complete failure of the insulation at one or more points; the
complete failure of insulation is called a short circuit. Fault calculations involve
finding the voltage and current distribution throughout the system during the fault.
Abnormally high current may flow as a consequence of abnormally high voltage on
the system due to lighting or switching surges which puncture through, or cause
flashover across the surface of otherwise healthy installation, and the resulting
damage (or ionization of the surrounding insulation) causes a follow-through power
arc.

The phrase ‘abnormally high current’ must be interpreted relative to the current
which normally exists in the circuit. Usually there is little or no current leakage to
earth so that an earth-fault current could mean any current greater than zero. Also
it is possible for a fault current at times of light load—Sunday morning in
summer— to be less than peak-load current in winter. Light load infers few
generators far apart on the system so giving a high source impedance which limits
fault current.
Symmetrical three-phase faults will be dealt with first; then asymmetrical faults
such as earth-faults or line-to-line faults, and for these the method of symmetrical
components will be used.
When MVA is mentioned in connection with faults and circuit breakers, it must be
clearly understood that the voltage by which the current is multiplied to give the
volt-amperes, is the rated voltage of that part of the system and not the actual
voltage at the fault or at the circuit breaker. Since the rated voltage is always
known, MVA (in this connection) is an indirect way of giving the fault current.
Ch.#4

2
 POWER SYSTEM FAULT CALCULATIONS

When fault current passes through a two-winding transformer, it is changed in

magnitude, but fault MVA is unchanged.
When a fault occurs at a point in a power system, the corresponding fault MVA is
referred to as the fault level at that point, and, unless otherwise stated, will be
taken to refer to a 3-phase symmetrical short-circuit.
2. Symmetrical 3-phase Short-circuit
2.1 BASIC CALCULATIONS
This type of fault is defined as a simultaneous short-circuit across all three phases.
It occurs infrequently, as for example, when a line, which has been made safe for
maintenance by clamping all three phases together to earth, is accidentally made
alive; or when, due to slow fault clearance, an earth-fault spreads across to the
other two phases; or when a mechanical excavator cuts quickly through a whole
cable. It is an important type of fault in that it results in an easy calculation and a
pessimistic answer being almost the worst case with regard to fault-level. (An
exception to this statement is dealt with in section 4.9.)
The calculation is easy because, being a completely symmetrical circuit, it can be
treated per-phase-star, i.e. equipment is star- connected (or is replaced by its
star-equivalent if delta-connected, as shown in Appendix 2) so that calculations
can take place on one phase of the circuit. Calculations in this single-phase circuit
using phase (line-to-neutral) e.m.f.s. and phase impedances yield a phase current,
which defines the current in the other two phases, since they are equal to it in
magnitude and displaced from it by 120° and 240°. When the fault is
asymmetrical, this simple theory no longer applies and the more involved theory of
symmetrical components is needed, as shown in sections 3 and 4.
The estimation of the MVA rating required of a circuit breaker is usually made on
the assumption that it must clear a 3-phase fault because, as that is generally the
worst case, it is reasonable to assume that the circuit breaker can clear any other
fault. The circuit breaker rated breaking capacity in MVA must therefore be equal
to, or greater than, the 3-phase fault level MVA.
Since circuit breakers are manufactured in preferred standard sizes, e.g. 250, 500,
750 MVA, high precision is not necessary when calculating the 3-phase fault level
at a point in a power system. It is usual therefore to make the following
assumptions:
a) Immediately prior to the fault, the system is on no-load at rated
frequency and all currents are neglected except the fault current. Unless
Ch.#4

otherwise stated, it is usual to assume that all generators are running at

3
 POWER SYSTEM FAULT CALCULATIONS

their rated voltages, e.g. 11 and 33 kV, and are interconnected by

transformers operating at the appropriate ratio, e.g. 11/33 kV.
b) MI the generators have their e.m.f.s in phase. This means that the
system is not swinging, that is, that synchronising currents have pulled
all machines into step (or synchronism) prior to the fault.
The combined effect of (a) and (b) is that all generators can be replaced
by a single generator since all e.m.f.s are equal and in phase. The word
equal must be interpreted as meaning equal when expressed in per unit
terms or in volts on the same voltage level. This procedure is sometimes
called paralleling the sources on to a common e.m.f. busbar (shown
dotted in the diagrams because it does not physically exist). The
impedances of the alternators are then treated as part of the power
system impedance network.
c) System resistance is neglected and only the inductive reactance of the
system is allowed for. This gives minimum system impedance and
maximum fault current and a pessimistic answer; the great advantage
gained from this assumption is that all reactances, currents and voltages
can be treated arithmetically. The student should check that the
difference between Z and X when X = 5R is only about 2%, and this is
acceptable when it is remembered that the system impedances are
probably not known to an accuracy better than about 5% of their own
values. The X/R ratio for alternators is about 20/1, for transformers
about 10/1 and for high voltage lines about 8/1 to 3/1. The neglect of
resistance is not so acceptable for light overhead lines of 11 kV and less,
or for cables where K and P can often be nearly equal, but in many such
cases the system is a radial (series) circuit and Z = R +jX can be used
without great difficulty. The procedure is to reduce the system to a single
reactance, then to calculate the current at the fault and to retrace the
steps of the calculation to find the current and voltage distribution
throughout the system.
All the above assumptions are merely convenient when obtaining an estimate of 3-
phase fault level at a point in a power system. For more accurate work, the load
currents can be allowed for by using network theory, e.g. Thevenin (Helmholtz) or
superposition; and resistance can be included in the system impedances. By using
digital computers, the accurate methods of calculation are becoming economic
Ch.#4

propositions.

4
 POWER SYSTEM FAULT CALCULATIONS

Assuming a symmetrical 3-phase short-circuit on a system where the rated voltage

is E0 let E volts be the e.m.f./phase (of the common e.ntf. busbar) behind a
reactance of X/ phase representing the whole system reduced to a single
reactance. Then the phase fault current is
If = E/X amperes. (1)
E and X must relate to the same voltage level in systems involving transformers. E
is often taken to be the rated phase voltage Vr (see assumption (a), although it will
in general exceed it by an amount which is not large if the fault current is being
calculated for the first cycle or two after the short-circuit , as is commonly
required. The total 3-phase fault-level VA is therefore
(VA)f = 3 Vr . E/X volt-amperes. (2)
Changing to the per-unit notation (see Appendix 1) and using the subscript b for
the base value of any quantity, then

Appendix I shows that [1] may be written as

(3)
The base volt-amperes/phase = Vb,Ib so that the per-unit faui.t volt-amperes are
given by

It may be noted that the 3 in F [2] has cancelled out in the per unit volt-amperes.
If the base voltage is equal to the rated voltage, Le. V6 = V,

(4)
As discussed above
E is often taken as equal to Vr, and is thus equal to Vb, i.e. £ = 10p.u.
Equations [3 ] and [4] then reduce to
(5)

and (6)
Ch.#4

5
 POWER SYSTEM FAULT CALCULATIONS

Worked example 1
Fig. 1(a) shows four identical alternators in parallel each rated at 11 kV; 25 MVA,
and each having a (subtransient) reactance of 16 % on its (continuous thermal)
rating. Find the 3-phase fault level at one of the outgoing feeders

FIG. 1. Three-phase fault.

(a) Actual system (single line). (b) Equivalent circuit.
.
SOLUTION
Base 11 kV, 25 MVA. The fault MVA from one alternator is

The total fault MVA = 625 x 25 x 4 = 625 MVA. Alternatively, Fig. 1(b) shows the
four alternators paralleled on to a common e.m.f. busbar and the alternator
reactance’s treated as part of the system reactance, and the total reactance is
0.16/4 = 0.04 p.u. and the fault MVA is
(VA)fpm = l/0O4 = 25p.u.
(VA)f = 25 x 25 = 625 MVA.

2.2 CURRENT LIMITING REACTORS

In example [1] the fault level is 625 MV whereas the circuit-breaker standard
ratings are 500 and nearest and 750 MVA. If, for economic reasons, it is necessary
to use 500 MVA circuit breakers, then the fault current must be reduced by
increasing the system reactance. If the system had been in the design stage, the
student should check that specifying alternators with 20% reactance would have
been a solution.
If however the system exists then it is necessary to add separate reactance’s either
in series with each alternator, or in series with each feeder, or, as is more usual, to
Ch.#4

sectionalise the busbar and connect sections via a busbar reactor as shown in

6
 POWER SYSTEM FAULT CALCULATIONS

Fig.2.

FIG.: 2 Sectionalised busbar.

(a) Actual System (b) Equivalent circuit.

Worked example 2
Using the data of example 1 but connected as in Fig. 2, calculate the reactance of
the current limiting reactor X to limit the fault level on a feeder to 500 MVA.
SOLUTION
Base 11 kV, 25 MVA.
Two alternators will supply 625/2 = 312.5 MVA directly to the fault, so the other
section must be limited to 500—3l2.5 = 187.5 MVA = 187.5/25 p.u. = 1/Xt where
Xt is the per unit reactance of the total right hand section. Thus Xt = 25/187.5
=0.133 p.u. = 13.3%. Since two alternators in parallel have a combined reactance
of 16/2 = 8%, the busbar reactor must provide the remaining reactance, i.e.
13.3—8 = 5.3%.
The ohmic value of the reactance can be calculated as follows:

The specification of current-limiting reactors is given in B.S.S. 171 and some

salient points are now given. Rated current is the current the reactor, can carry
continuously without exceeding a specified temperature rise dependent on the type
of insulation used. In the example 2, the reactor might be rated to carry a current
Ch.#4

corresponding to 25 MVA at 11 kV, i.e. 1.31 kA, so that, should one alternator

7
 POWER SYSTEM FAULT CALCULATIONS

break down, up to 25 MVA could be taken from the other section. Notice that the
voltage used is the rated system voltage and not the voltage between the
terminals of one phase of the reactor.

= 187.5/25 = 7.5

Over-current time is the time in seconds that the reactor can carry the above fault
current without suffering damage.. A typical time would be 3 seconds. On the basis
that during a fault lasting at most only a few seconds, the transformer will absorb
all the I2Rt heating effect (and dissipate none), the relation between the fault
current and the permissible fault time is given approximately by I2t = constant.
This indicates, for example, that the reactor could carry half the above fault current
for about 12 seconds. The reactor must be designed to withstand the
electromagnetic forces corresponding to
2.55 x r.m.s. rated current x over-current factor
= 2.55 x r.m.s. symmetrical rated through-fault current.
Current-limiting reactors should have a reactance high enough to limit the fault
current, but not so high as to cause excessive voltage drop due to load current.
The use of busbar section reactors as shown in Fig. 2 is common at all voltage
levels. Reactors are occasionally inserted into sections of the transmission network
either to limit fault level, or to control the sharing of load current between lines in
parallel forming a closed ring between two stations.
If X is the reactance of the reactor in /phase, I the thermal rated (load) current
in amperes, and V the rated line-to-neutral voltage of the system, then at rated
current the voltage between terminals of one phase = IX volts
= IX/Vp.u. = Xpu
total 3-phase reactive volt-amperes = 3(IX)I = 3I2X VAr
= 3I2X/(3 VI) = Xpu
Thus in example 2, the reactor would have a through-fault rating of 187.5 MVA
(9.85 kA at 11 kV line), a through-load rating of 25 MVA (1.31 kA at 11 kV line), a

voltage between terminals of one phase of 53% of 11/ 3 kV=337V=

1310x0.258V, and an MVAr rating of 5.3% of 25 MVA = 1.323 MVAr = 3x 1.31 x
0.337 MVAr.
Ch.#4

8
 POWER SYSTEM FAULT CALCULATIONS

2.3 FAULT IN-FEEDS: SOURCE IMPEDANCE

Fig. 2 illustrates the layout of a power station prior to the mid 1930s. Each station
supplied its own local load independently of other power stations. In the mid-l930s
the 132-kV Grid was established to interconnect the larger power stations. Fig. 3
shows a typical arrangement. Each of the 11-kV busbar sections is connected to
the Grid via a 50-MVA, 12.5% reactance transformer. The remainder of the Grid
system is shown (dotted) as an equivalent alternator. If a 3-phase fault occurs on
the 132-ky busbar, the MVA fed to the fault by this equivalent alternator (but
excluding the power-station shown) is called the Grid in-feed and is an indirect way
of giving the source impedance of this equivalent alternator. Thus if the Grid in-
feed were given as 1500 MVA, then on a base of 100 MVA, the Grid in-feed MVA =
1500/100 = 15 p.u. and the Grid source reactance = 1/15 = 0.0667 p.u. Fig. 4
shows the system reduced to a common e.m.f. busbar and with all reactance’s
expressed in per unit on 100 MVA base. It is shown in Appendix I that machine
reactance’s are proportional to base MVA. The reduction of the network to a single
reactance will involve the use of the star-delta transformation—see Appendix 2.
Worked example .3
Calculate the fault level for a 3-phase symmetrical short circuit on an 11-ky feeder
in the system shown in Figs. 3 and 4.

Fig. 3. Grid in-feed to fault.

Ch.#4

9
 POWER SYSTEM FAULT CALCULATIONS

SOLUTION
Base 100 MVA.
Replacing the mesh-connected p.u. reactances connected to nodes 2, 3 and 4 by a
star-connected set with a star-point node numbered 5, then the latter are
X25 = (0.25x0.25)/0.712 = 0.0878p.u.
X35 = X45 = (0.25x0.212)/0.712 = 0.0743 p.u.

Then from Fig. 4(a) and (b), between nodes 1 and 5 there are two reactances in
parallel of (0.0667 +0.0878) and (0.32 + 0.0743) i.e. (0.1545 x 0.3943)/(0.0667
+0.0878 + 0.32+0.0743) = 0.1105 p.u.

The reactance between nodes 1 and 3 is then 0.12 in parallel with

(0.1105+0.0743) i.e. (0.12 x 0.1848)/0.5048 = 0.417 p.u.
The fault MVA level at the 11-kV feeder near the circuit breaker (i.e. node 3) is
thus 100/0.117 = 855 MVA.
This shows that the circuit breakers at the supply end of the 11-ky feeders would
need to be up-rated to the next standard size, viz. 1000 MVA.

Fia. 4. (a) Equivalent circuit for Fig. 8.3 on 100.MVA base.

(b) Simplified equivalent circuit.

During the period when the demand for electricity was doubling about every 8 to
10 years, new power stations and transmission lines were being added to the Grid
system. Thus the internal impedance of the Grid system was falling and its fault
level rising. Also, in any year, the fault level is highest when the Grid is running
Ch.#4

under maximum (peak) load conditions, usually in December or January. Under

10
 POWER SYSTEM FAULT CALCULATIONS

these conditions, the fault level at all major busbars on the Grid system is
calculated, and referred to as the maximum plant fault level. Although minimum
Grid impedance is associated with peak loading conditions, the same fault levels
would obtain if the same machines were all running but on no-load.

3. Symmetrical Components
3.1 GENERAL PRINCIPLES
The majority of faults in a power system are not symmetrical 3-phase short-
circuits, but occur between one line and earth or, less commonly, occur between
lines where they may also involve earth. Also there is usually a finite fault
impedance rather than a short- circuit. Thus instead of a completely balanced 3-
phase circuit which can be solved by calculating the current flowing in one of the
three phases, as in section 2, an unbalanced 3-phase circuit must now be solved.
This can most conveniently be done by using the symmetrical components of an
unbalanced system of currents or voltages, because this yields three (fictitious)
single-phase networks only one of which contains a driving voltage (e.m.f.). The
impedances in these three networks do include some which are different from
those used in the analysis of completely balanced 3-phase networks (which strictly
should all be prefaced by the words ‘positive sequence’ when used or defined). For
example, it is shown that a synchronous machine under balanced conditions may
be regarded as having effective reactances called synchronous, transient and sub-
transient at different times, and these are all positive-sequence reactances, but
that when the current is unbalanced other reactances also are needed.
The method of symmetrical components can be applied to any poly phase system
containing any number of phases, but the 3-phase system is the only one of
interest here. It can be applied to a set of unbalanced 3-phase curre.nts or
voltages and will be used for both in this chapter (section 4). though the equations
relating the original (real or physical) quantities to theft symmetrical components
will be stated in terms of currents. The symmetrical components of a line current
may be said to be fictitious, since metering the line current directly gives the real
current which may be unbalanced, i.e. unequal in magnitude to the other two line
currents and/or not at 120° with respect to each of them. On the other hand it is
possible to obtain a measure of any one of the symmetrical components of an
unbalanced current (or voltage), by connecting a metering circuit in which the
other components do not flow (and this has to be done for some circuits used in
Ch.#4

power systems, see section 5).

11
 POWER SYSTEM FAULT CALCULATIONS

The theory of symmetrical components states that, by the Superposition Principle

in a linear network, any set of asymmetrical 3- phase phasors of sequence abc,
e.g. currents Ia,Ib and Ic can be replaced by the sum of three sets of currents:
(i) a positive-sequence set of three symmetrical 3-phase currents (i.e. all
numerically equal and all displaced from each other by 120°), having the same
sequence abc as the original set, and denoted by Ia1, Ib1, and Ic1
(ii) a negative-sequence set of three symmetrical 3-phase currents having the
sequence opposite to that of the original set, viz. acb and denoted by 1a2, Ib2 and
1c2.
(iii) a zero-sequence set of three currents all equal both in magnitude and in
phase, i.e. having no sequence, and denoted by Ia0, Ib0 and Ic0 and these all equal
I0.
Fig. 5 (a) illustrates these three sets of symmetrical components. It will be shown
later that the relative values of the three components of current in any one line Ia1,
Ia2 and Ia0 and the relative phase angles between them, will vary from one
asymmetrical fault condition to another. Fig. 5(a) can therefore illustrate only one
particular set of conditions. It must be noted that although the sequence acb of
negative-sequence is opposite to that of positive-sequence, the set of negative-
sequence phasors must still be regarded as rotating in the same direction
(anticlockwise) as the positive- (or zero-) sequence phasors.

(a) Sequence component (b) Addition of sequence component

Fig. 5. Symmetrical components
Ch.#4

12
 POWER SYSTEM FAULT CALCULATIONS

The negative-sequence currents when added to the positive- sequence currents,

will give a resultant which is a set of three asymmetrical 3-phase currents, but the
phasor sum of these three currents will be zero since the two components sets are
both symmetrical 3-phase sets, i.e. if

Addition gives
Ia+Ib+IC= 0+0.
Under unbalanced conditions, however, there will generally be a resultant total
current which is the sum of the three line currents, e.g. when there is an earth-
fault. The current flowing in the earth path which is the sum of the three line
currents must be the sum of the three zero-sequence currents flowing in the three
lines, since the positive- and negative-sequence components sum to zero. Thus
earth-path current
=Ia0+Ib0+IC0
3Ia0=3 I0 (7)
The three sets of sequence currents, each being symmetrical in themselves, can be
calculated on a single-phase (or per-phase-star) basis, with the actual current
flowing in any one of the phases given by the sum of its three sequence
components, i.e.
Ia=Ia1+Ia2+Ia0
Ib=Ib1+Ib2+Ib0 (8)
Ic=Ic1+Ic2+Ic0

Fig. 5(b) shows the particular unbalanced currents given by adding the
symmetrical components shown in Fig. 5(a).
By using the h operator discussed in Appendix 3, [8.8] is reduced to
Ia=Ia1+Ia2+Ia0
Ib=h2Ia1+hIa2+Ia0 (9)
Ic=hIa1+h2Ia2+Ia0
In the symmetrical component method the equations are simpler if only faults
which are symmetrical with respect to one phase are considered (this will be seen
to be true of all faults considered in this chapter) and this phase is taken as the
Ch.#4

reference phase and is called the a phase (whether it is 1?, Y or B depends on the

13
 POWER SYSTEM FAULT CALCULATIONS

actual system).
The three equations in [9] can now be solved for the three component currents in
terms of the three actual system currents, provided that the determinant

1 1 1
h2 h 10
h h2 1

The student should satisfy himself that the determinant has the value j3 3 and is

thus not zero. To solve for Ia1 the second equation in [9] may be multiplied by h,
and the third by h2, and these two equations may now be added to the first. The
result is
Ia1= (1/3)(Ia+ hlb+h2Ic)
Ia2= (1/3)(Ia+ hlb+h2Ic) [10]
Ia0= (1/3)(Ia+ lb+Ic)
Using [10] the student should show that if the actual currents form a symmetrical
3-phase set, they contain only positive-sequence components, and that there are
no zero-sequence components of current in a 3-phase, 3-wire system having no
leakage or charging currents to earth. (Students familiar with matrices should write
[9] in matrix form and by matrix inversion obtain [10].
The above three components of current in phase a can be considered to flow in
three fictitious single-phase circuits, which are coupled together at the point of
fault in a manner which depends upon the nature of the fault. If there were any
mutual coupling between these three networks, the analysis would be more
complicated. Before considering these networks therefore, we shall consider the
conditions under which no mutual coupling exists between them, and show that
these conditions are met in a power system. It is sufficient to consider the
components of current Ia1, Ia2, and I0 since the currents in phases b and c can be
determined from [9].

3.2 CONDITIONS FOR NO MUTUAL COUPLING BETWEEN THE THREE

SEQUENCE NETWORKS
Fig. 6 shows a 3-phase system having balanced self-inductive reactance XL/phase
and mutual-inductive reactance XM between pairs of phases. The line currents are
Ia, Ib, and Ic and they may be unbalanced.
Ch.#4

14
 POWER SYSTEM FAULT CALCULATIONS

where X=jX.
The positive-sequence component of the series voltage drop, with Vaa’ as the
reference phasor, is from [9]

Fig. 6. 3-phase system with balanced impedances.

The positive-sequence reactance/phase can be defined as

Similarly, the negative-sequence reactance/phase =

and the zero-sequence reactance/phase =

The equation for Vaa’ can be rewritten in terms of the sequence components of
current as

This result of [11] shows that, providing the system reactances are balanced from
Ch.#4

the points of generation right up to the fault, as they are in a power system, each

15
 POWER SYSTEM FAULT CALCULATIONS

sequence current causes a voltage drop of its own sequence only. There is
therefore no mutual coupling between the three sequence networks. The same
result may be shown more rigorously by matrix methods with impedance in the
neutral and mutual coupling between the three lines and neutral included.
This has been left to Volume 2 where other transformations of coordinate systems
will be dealt with by matrix methods.

3.3 THE THREE SEQUENCE NETWORKS

In section 3.2 it has been shown that positive-sequence currents do not give rise to
voltages of negative- or zero-sequence; that negative-sequence currents give
voltages of negative-sequence only, and that zero-sequence currents give voltages
of zero-sequence only. We can therefore regard each current as flowing within its
own network through impedances of its own sequence only, such as the reactances
X1, X2 and X0 referred to in section 3.2.
Consideration of a synchronous machine’s windings shows that a normal machine
without any fault can only generate balanced e.m.f.s of normal sequence, i.e.
positive-sequence abc. Thus an e.m.f. Ean1 only is generated in the positive-
sequence circuit of the a phase, and Ean2 and Ean0 are zero. Ean1 is the line-to-
neutral e.m.f. of the generator. Z1 in Fig. 7 is the single positive-sequence
impedance which appears between the generation and the point of fault, when all
the impedances of the system have been reduced as discussed in section 2 (since
the system is balanced Za1= Zb1= Zc1 = Z1 and similarly for the other two
sequences). The positive- sequence voltage of phase a to earth at the point of fault
Vae1 is thus given by

For the negative-sequence, since Ean2 = 0, the negative-sequence component of

the voltage of phase a to earth at the point of fault is

[13]

For the zero-sequence, if we consider the source to be a single alternator with its
star-point earthed through an impedance Zn then [7] shows that Zn carries a
current of 3I causing a voltage drop of 3Ia0Zn. To give the same voltage drop the
zero-sequence network of Fig. 7 must therefore contain an impedance 3Zn as
Ch.#4

16
 POWER SYSTEM FAULT CALCULATIONS

shown. Thus
[14]

and the p.d. between phase a and earth at the point of fault is Vae given by

[15]
Similarly Vbe and Vce may be determined from [9]

[16)
and

[17]
Already then one point of difference has emerged between the zero- and negative-
sequence networks. Other differences will emerge when we consider briefly in the
next section how the impedances Z1, Z2 and Z0 may be obtained for a power
system with many items of interconnected equipment.

Fig. 7. The three sequence networks.

Clearly no current will flow in the three sequence networks when they are isolated
as shown in Fig. 7, but in section .4 the interconnection of these networks will be
derived to represent various unbalanced fault conditions.
3.4 Z1, Z2 AND Z0 FOR AN INTERCONNECTED POWER SYSTEM
The process of reducing the positive-sequence impedances/phase of an
interconnected network to a single value between generation and fault, is exactly
that dealt with in section 2, where though it was not said then, all impedances
were in fact the positive-sequence impedances of the equipment.
All static equipment has the same impedance to negative-sequence current as to
Ch.#4

positive-sequence current.

17
 POWER SYSTEM FAULT CALCULATIONS

For synchronous machines, the positive-sequence reactance varies with time , and
the negative-sequence reactance is not necessarily equal to it, although the
difference may be small during the first few cycles after a fault. The lay-out of the
negative- sequence circuit is identical to that of the positive-sequence network
since both are balanced 3-phase systems, and they will only differ in the values of
reactances for synchronous machines and the lack of negative-sequence e.m.f.
(see Fig. .8(b) and (c). Z2 is thus found by the same process of reduction to a
single impedance by combining elements in series or parallel, and using delta-star
transformations.

Fig. 8. Unbalanced fault in inter-connected system.

(a) System (b) Positive-sequence network
(c) Negative-sequence network (d) Zero-sequence net work

In the zero-sequence network, however, an unearthed star-point of an alternator

or transformer winding, causes Z0 to be infinite for that part of the circuit since if
there is no path to earth Io = 0
Thus in the zero-sequence network of Fig. 8(d), the zero-sequence impedance of
alternator B does not appear because it is in series with an infinite impedance. The
zero-sequence network generally has different values of impedance, for every item
of equipment, from the positive-sequence network , though in some transformers
Ch.#4

the impedances may be taken as equal

18
 POWER SYSTEM FAULT CALCULATIONS

If a transformer has a delta-connected winding, then again an open-circuit is

introduced into the zero-sequence network since zero-sequence currents circulate
around a delta-connected winding without flowing in the 3-phase circuit connected
to it.
The alternator is the only equipment with a time-varying reactance to positive-
sequence current. The sub-transient reactance is used in the positive-sequence
network if for example the maximum mechanical forces during a fault are required,
and the d.c. component also is taken into account. The transient reactance is often.
used during the period while the protective system is operating, and a circuit
breaker is opening. It is often unnecessary to consider any variation in negative-
and zero-sequence reactances of a synchronous machine during the sub-transient,
transient and steady-state regions, so that one value for each is sufficient.
The process of calculating the unbalanced fault currents and voltages in any part of
a power system may now be summarised as follows:
(a) determine the three sequence networks from generation to point of fault
by inspection of the system diagram;
(b) insert the sequence impedances in each of the networks;
(c) reduce each network to a single impedance Z1, Z2 and Zo;
(d) connect these sequence networks together with any fault impedance in
the way shown in section 4 to apply to that particular unbalanced fault;
(e) calculate the sequence components of current in phase a at the point of
fault;
(f) calculate the sequence components of current in phase a in any branch of
the sequence networks, by working back from the point of fault;
(g) calculate the actual current in any phase in any branch of the circuit using
[9];
(h) calculate the sequence components of p.d. between phase a and earth at
any point in the system from [12], [8.13] and [14];
(i) calculate the actual p.d. between any phase and earth at any point in the
system from [15], [16) and [17].

3.5 PHASE-SHIFT THROUGH A DELTA/STAR TRANSFORMER

It can be shown that for a DY transformer, the equivalent primary line-to-neutral
voltage (i.e. the phase voltage of the star-connected alternator supplying the delta
primary) was 30o ahead of the corresponding secondary line-to-neutral voltage.
Ch.#4

19
 POWER SYSTEM FAULT CALCULATIONS

It will now be shown that positive-sequence currents in the primary lines are also
30o ahead of the corresponding secondary positive- sequence line currents. The
student should satisfy himself that if k equals the phase-to-phase turns ratio

(Ns/Np), the corresponding line voltage ratio is . 3 k.

The positive-sequence currents in the secondary are Ia1, Ib1 = h2Ia1 and Ic1 =
hIa1. Multiplying these by k will give the corresponding primary phase currents and
subtracting in pairs will give the line currents. Thus

This shift of positive-sequence voltage and current by the same amount in the
same sense is to be expected, since positive-sequence current represents load
current, and the load power factor is not changed by an ideal transformer.
The student should now show that for negative-sequence currents

Thus positive-, and negative-sequence currents are shifted 30° in opposite senses.
It should be noted that the numerical transformation factor is the line-voltage
ratio: this is to be expected since it is a secondary line to primary line current
transformation. (The student should now show that for a DY 11 transformer , the
same results hold except that the phase shift is now —30° and +30° respectively.)
The above results can be used when it is required to find the current distribution in
a complex network on the primary side of a delta/star transformer.
Another illustration of application is to show that, if Ia is an earth-fault current on
the secondary side, the primary fault current is k Ia (a result which is obvious from
the definition of k). For an earth-fault it will be shown in section 4.1 that Ia=Ia1
=Ia2=Ia0 and
Ia= Ia1 +Ia2+Ia0 (equation [ 8]). Since there are no zero-sequence currents in the
primary lines
Ch.#4

20
 POWER SYSTEM FAULT CALCULATIONS

4. Asymmetrical Faults
An asymmetrical fault is any fault other than a 3-phase symmetrical fault: e.g. an
earth-fault on one phase; a line-to-line fault (or phase-to-phase fault sometimes
contracted to phase-fault); and a double-earth-fault (or line-to-line-to-earth fault).
One notation is a-e for an earth-fault on phase a and b-c for a phase-fault across b
and c. Problems involving asymmetrical faults are solved by resolving the
asymmetrical fault currents into their symmetrical components. The power system
is then analysed three times, once for each of the three sequence components, and
the results added to give the actual fault current (or voltage).
During asymmetrical fault calculations it is usual (unless the contrary is explicitly
stated) to make the same simplifying assumptions that were made during
symmetrical 3-phase faults: namely, that all source e.m.f.s are equal (in per-unit)
and in phase so that they can all be replaced by a single source e.m.f. If all the
resistances can be neglected, the system reduces to a single reactance in the
positive-sequence network with a single e.m.f. behind it.
In the analyses that follow, it should be noted that the reference phase denoted by
a is always the symmetrical phase with respect to the fault: i.e. an earth-fault is on
phase a, while phases b and c are healthy; and a phase-fault is across phases b
and c while phase a is healthy. This is done to give the simplest results in
mathematical terms. In applying the analysis in terms of phases abc to a practical
problem in terms of phases RYB, the correct phase sequence must be maintained:
i.e. if phase a is B, then b is R and c is Y.

4.1 EARTH FAULT ON PHASE a

Fig. 9 shows a simplified unloaded power system with an earth- fault on phase a at
the point f. The source star-point is earthed via an impedance Zn (usually a
resistance) and the fault path is shown as having an impedance Z1 (usually a low
resistance). The e.m.f.s and impedances of the power system are always assumed
to be balanced up to but excluding the fault, so that there is no mutual coupling
between the sequence networks as shown in sections 3.2 and 3.3. The fault
current is Iae. Also shown in all three phases are the sequence components of
current.
The procedure in this and in the following analyses will be to write down by
inspection of the circuit diagram, the three equations for phases a, b and c which
define the fault condition in terms of the actual phase currents and voltages. These
Ch.#4

currents and voltages will then be transformed into their symmetrical component

21
 POWER SYSTEM FAULT CALCULATIONS

form.

Fig. 9. Earth fault on phase a

Using [9], these voltages and currents may be replaced by the sum of their
symmetrical components and [18], [19] and [20] become:

[21]
Va1, Va2 and Va0 are the symmetrical components of the p.d between the reference
phase and earth at the point of fault f.

.
Ch.#4

22
 POWER SYSTEM FAULT CALCULATIONS

Equation [24] shows that for this particular fault the three sequence currents in
phase a are all exactly equal in magnitude and in phase.
Re-arranging terms in [21]

The interconnection of the sequence networks needed to satisly [24] and [25] is
shown in Fig. 10 where
Ch.#4

23
 POWER SYSTEM FAULT CALCULATIONS

If the system star points are solidly earthed then Zn= 0, and if the fault is a short-
circuit (i.e. with negligible impedance) then
Zf = 0, and Fig. 10 reduces to a simple series connection of the three sequence
networks with a fault current given by

The voltages to earth of the un-faulted phases band cat the point of fault can be
found from [ 16] and [ 17].
Worked example 4
A 3-phase, 33-ky, 37-5-MVA alternator is connected to a 33-ky overhead line
which develops an earth fault on one conductor at the remote end. The positive-,
negatives, and zero-sequence reactances of the alternator are 18, 12 and 10% (on
rating), while those for the line are 6’3, 6-3 and 12-6 fl/conductor. Calculate the
fault current, and the phase voltages at the alternator terminals. Assume the
alternator star point is solidly earthed.
SOLUTION
Base 375 MVA (3-phase), 33-ky (line).
Base voltage = I p.u. = 33/.y/3 = 19-1 ky/phase
base current = I p.u. = 315/(.J3x33) = 0-655 kA
base reactance = I p.u. = 19-1/0-655 = 29-1 f2
so 6-3 ohms = 6-3/29-1 = 0-2165 p.u.
The total reactance of the equivalent network of the three sequence networks in
series (Fig. 8.10) is

and fault current

The voltages with respect to earth at the alternator terminals are calculated using
[12], [ 13] and [14] where the reactances are those from the source to the
Ch.#4

24
 POWER SYSTEM FAULT CALCULATIONS

alternator terminals.

The fault MVA = 1.555x 19.1 = 29.6 MVA.

The student should check that this last result can be obtained from [6], i.e. (VA)f =
1/Xp.u., providing that X is interpreted as Xt the total reactance of the fault
equivalent network.
In practice alternators are usually earthed via a resistor of value about I00% on
alternator rating. The corresponding value to be added to Xt is 300% = 3 p.u. Thus
the total impedance of the fault equivalent network becomes 3+j 1.266 p.u.

4.2 LINE-TO-LINE (PHASE-FAULT) ACROSS LINES b And c

Fig. 11 shows the system, and it should be noted that the conventional positive
direction of current flow is taken as being away from the source. There is no
voltage across Zn and I, is zero since there is no path in which current can flow to
the source star-point via earth.
Ch.#4

25
 POWER SYSTEM FAULT CALCULATIONS

Fig. 11. Line-to-line (phase) fault across b and c.

By inspection of Fig. 11 we may write

The student should show that by transforming the currents and voltages in these
three equations into their symmetrical components, they may be solved to yield
the results

by the interconnection of From [29] and [30] or by inspection of fig. (12) it follows
that

From [9], it is readily shown (though the student may wish to confirm this) that
the fault current is given by
Ch.#4

26
 POWER SYSTEM FAULT CALCULATIONS

Worked Example 5
Using the data of example 4 but with a short-circuit between lines b and c at point
f, calculate the fault current and the voltage of the healthy line to earth at the
point of fault.

Fig. 12. Interconnection of the sequence networks for a phase-faith

SOLUTION
The total reactance of the fault equivalent network, Fig. 12, is

Ch.#4

27
 POWER SYSTEM FAULT CALCULATIONS

The student should check that this last result can be obtained from [6] providing
that X is interpreted as the total reactance Xt of the fault equivalent circuit.

4.3 LINE-TO-LINE-TO-EARTH SHORT-C IRCUIT (DOUBLE-EARTHFAULT) ON

PHASEs b AND c

Fig. 13. Interconnection of the sequence networks for a

line-to- line-to-earth short-circuit on b and c.

Before proceeding to the general case of a double-earth-fault in section 4.4, the

simpler case of a dead short-circuit to earth on two phases b and c is outlined in
this section. For this fault, inspection of the circuit gives
Ia= 0, Vb = 0 and Vce = 0.
Transforming these three equations into their symmetrical components, gives (as
students should check)
Ch.#4

28
 POWER SYSTEM FAULT CALCULATIONS

It may be noted from [14] that lithe system star points are earthed solidly (i.e. Zn
= 0), then Z’o becomes Z0 which is the single effective zero-sequence impedance
corresponding to all the equipment connected between generation and fault. The
currents in the two faulty lines may be obtained from [35] and [9], and the fault
current in the earth path is Ib+Ic=3Ia0

4.4 GENERAL CASE or 000BLE-EARTH-FAULT ON PHASES b AND C

The general case of a double-earth fault which is symmetrical with respect to the
reference phase a, is a fault occurring between phases b and c with the mid-point
of the fault impedance 2Zf between them connected to earth by an impedance Zg.
It has already been noted that the method of symmetrical components is
considerably simplified if it is applied to faults which are symmetrical with respect
to the reference phase a. Again the system must be balanced from generation to
the point of fault.
The method used here follows exactly the procedure used in section 4.1. The three
equations which may be written for the actual voltages and currents at the point of
fault are :

Writing these currents and the potential differences between the line conductors
and earth at the point of fault, as the sum of their symmetrical components, gives
Ch.#4

29
 POWER SYSTEM FAULT CALCULATIONS

FIG. 14. General double-earth fault on phase b and c.

Collecting together positive-sequence terms on the left-hand side and negative-

sequence terms on the right-hand side gives
Ch.#4

30
 POWER SYSTEM FAULT CALCULATIONS

and substituting this result in [41] gives

and the last equation simplifies to

Equations [39] and [42] are satisfied by the interconnection of the sequence
networks shown in Fig. 15.
From this general shunt fault, the results for a number of simpler faults which are
more likely to occur in practice may be obtained, including those outlined in section
and in section 4.3 (ZF= 2Zf and Zg =, and section 4.3 (Zf=0 amd Zg=0)

Ch.#4

31
 POWER SYSTEM FAULT CALCULATIONS

4.5 GENERAL UNBALANCED THREE-PHASE-EARTH FAULT

Fig. 16 shows an earth fault on all three phases which is again symmetrical with
respect to the reference phase a. The equations which may be written by
inspection are:

By transforming these equations into symmetrical components the student can

deduce as an exercise that

Fig. 16. Unbalanced 3-phase earth fault.

Ch.#4

32
 POWER SYSTEM FAULT CALCULATIONS

Equations [43] and [44] are satisfied by the interconnection of sequence networks
in Fig- 17, from which the conditions for sections 4.2, 4.3 and 4.4, but not 4.1,
may be deduced.

4.6 SERIES FAULT ON PHASE a

If in one of the phase conductors a break occurs with an arc across the break, then
this is equivalent to an impedance Zr appearing in series with that phase with no
corresponding impedance in the other two phases, as shown in fig.18
Ch.#4

33
 POWER SYSTEM FAULT CALCULATIONS

Fig. I8. Series fault in phase a. For this condition, we can write by inspection

Fig. 19. Interconnection of the sequence networks for a series fault in phase a.

Transforming these equations into symmetrical component form gives

Ch.#4

34
 POWER SYSTEM FAULT CALCULATIONS

Equations [49] and [50] give Vaa’1 = Vaa’2 and substitution of this result back in
either equation gives
Vaa’1 = Vaa’2 = Vaa’0
Substitution of [51] in [8.48] gives
Vaa’1 = (Zr/3)(Iaa’1 +Iaa’2+laa’0). [52]

Equations [51] and [52] are satisfied by the interconnection of the sequence
networks as shown in Fig. 19. It should be noted that between terminals 1 and 1’,
2 and 2’, 0 and 0’, are connected the phase (line-to-neutral) positive-, and
negative-, and zero-sequence impedances respectively of the circuit to the right of
terminals a’, b’ and c’ in Fig. 18. If the arc at the broken conductor ends is
extinguished then Zr in Figs. 8.18 and 8.19 =.
4.7 SIMULTANEOUS SHUNT AND SERIES FAULTS ON PHASE a
If the a phase conductor breaks it may touch earth, on the source side, with an
impedance Zf to earth, as shown in Fig. 20.
For this condition [24], [25], [51] and [52] apply. In order to connect an
equivalent circuit for the three sequence networks which satisfies all four
equations, it is necessary to include ideal 1:1 isolating transformers as shown in
Fig. 21.
Fig. 20. a phase conductor broken and earthed.

Ch.#4

35
 POWER SYSTEM FAULT CALCULATIONS

Again, any power system connected to the right of a’, b’, c’ is connected to the
right of 11’, 22’, 00’. For example, if the right-hand system consists of live
generation, its positive-, negative-, and zero- sequence networks should be
connected across 11’, 22’ and 00’ respectively.

Fig. 21. Interconnection of the sequence networks for a p/wise

conductor broken and cart bed.
4.8 FAULT VOLT-AMPERES
The fault volt-amperes, usually termed the fault level and quoted in MVA, is the
product of the fault current and the rated voltage of the system (not the voltage at
the fault). In the case of an earth-fault, the voltage used is the line-to-neutral (or
phase) voltage, while for a phase-fault the voltage used is the line-to-line voltage.
In this context phase angles are ignored and only scalar values are considered.
It will now be shown that these volt-ampere fault-levels can be obtained directly
from the fault equivalent networks, lithe per-unit system is used. Let Zt be the
total impedance of the fault equivalent network so that for an earth-fault (Fig. 10)
= Z1+Z2+Z0+3Zn+3Zf
and for a phase-fault (Fig. 8.12)
Zt = Z1+Z2+ZF.
Let Vr be the rated phase voltage of that part of the system on which the fault
occurs. Then taking rated values as base values, the 3-phase base volt-amperes
are
Ch.#4

3VrIr = 3Vr2/Zr

36
 POWER SYSTEM FAULT CALCULATIONS

For an earth fault, [28] shows that

(fault VA) = 3Vr.Ean/Zf
and in per-unit

Similarly for a phase-fault, [ 31] and 32] show that (fault VA) = ,J3 V.. .J3 ECIZ

If the e.m.f. behind the fault is taken to be equal to rated voltage (which is not far
from the truth for the first few cycles after a fault), then as for a symmetrical 3-
phase fault (section 2), [53] and [54] reduce to 1/(Zf)p.u. These fault volt-
amperes are expressed relative to a 3-phase base volt-ampere value: this is logical
since it is usual to quote base volt-amperes as a 3-phase value.

4.9 COMPARISON OF FAULT LEVELS

The relative magnitudes of fault currents and fault volt-amperes will now be
compared, taking the 3-phase fault as the standard for comparison (1.0 p.u.). For
simplicity it will be assumed that the system resistance is negligible and that X1 =
X2: i.e. for alternators, the sub-transient reactance is taken for X1. Also, it will be
assumed that the fault impedance is negligible. Since the main point of the
exercise is to show that an earth-fault near a solidly-earthed source neutral-point
can exceed that of a 3-phase fault, it will be assumed that the system is solidly
earthed.
(a) 3-phase fault :
Fault current = Ean/X1
Fault VA = 3VrEan/X1
(b) Earth-fault :
Fault current = 3Ean/(X1 + X2 + X0)
and expressed in per-unit of the 3-phase fault value
fault current = 3X1/(2X1 + X0) p.u.
If X0 = X1 then earth fault current = 1.0 pu.
If X0< X1 then earth fault current >1.0 p.u.
Ch.#4

Fault VA = 3VrEan/(X1+X2+Xo),

37
 POWER SYSTEM FAULT CALCULATIONS

and again in per-unit of the 3-phase fault value we have fault VA =

X1/(2X1+X0)p.u.
This fault will be cleared on one pole of a 3-phase circuit breaker having three
identical poles. Thus the 3-phase breaking VA rating of the circuit breaker will be
3X1/(2X1+X0) p.u.
Thus the conclusions are the same as for earth-fault current, namely, that if X0 =
X1 the same circuit breaker will clear a 3-phase fault or an earth-fault, but if X0<
X1 an earth fault requires a circuit breaker of larger rating. Since the zero-
sequence network only extends from the point of fault back to the nearest
transformer the primary of which is unearthed star-, or delta-connected, while the
positive-sequence network extends right back to the source alternators, it can
often happen that X0<X1.
(c) Line-to-line short-circuit (phase-fault)

Fault current = 3 Ean/2X1

= 3
= 0867 p.u.
2

Fault VA= 3 3 = V Ean / 2 X 1 = 0.5 p.u.

r

and since this fault is cleared on two poles of the circuit breaker, the corresponding
3-phase rating of the circuit breaker is
(3/2)0.5 = 0.75 p.u.
This simple exercise illustrates why a phase-fault is relatively unimportant (in this
context).

4.10 ASYMMETRICAL FAULT WHILST SYSTEM IS ON LOAD

Most elementary fault calculations assume that the fault occurs whilst the system
is on no load with the alternators excited to the rated voltage of the system. If
however the system is on load when the fault occurs at point f in Fig. 22, the fault
current can be found by using the Thevenin (Helmholtz) theorem. The current in
the fault path is found by open-circuiting the fault path and dividing the line- to-
neutral voltage across the open-circuit (this will be Vf the (r.m.s.) voltage at the
point of fault immediately prior to the fault) by the impedance of the passive
network obtained by looking into the open-circuit (this will include the impedance
of the fault path itself). The appropriate alternator reactance will be used—
Ch.#4

subtransient or transient. The polarity of Vi which creates the fault current should

38
 POWER SYSTEM FAULT CALCULATIONS

be carefully noted. Vi is superimposed upon the pre-fault Vf in order to reduce the

total voltage at the fault to zero. Thus the conventional direction of the fault
current is through the fault path from f to e. Assuming that the system was
balanced prior to the fault, Vf will be a positive-sequence voltage only.
FIG. 22. Application of Thevenin theorem at point of fault.

The fault currents can be calculated by using the equations appropriate to the
fault, e.g. [26] and [27] for an earth fault on phase a, [31] and [32] for a phase-
fault, etc, with Vf in place of Ean. This follows since Thevenin’s theorem shows that
for a fault on a linear system with balanced load, all the interconnections of the
sequence networks (Figs. 10, 15, etc.) apply to their particular fault when Vf
replaces Ean as the positive-sequence driving voltage. The resulting currents into
the fault itself and through the system supplying it, may then be added by
superposition to the original load currents in order to give the actual system
currents during the fault period (subtransient or transient) under consideration.

Worked example 8.6

In Fig. .23a G is an exporting grid area (generator) having a 3- phase fault level of
20 000 MVA at the 400-kV busbars S. The corresponding data for the grid
importing area M (motor) is 10 000 MVA at R. The short 400-kV line (L) has a
series reactance of 40
ohms/phase. The load transfer is 150 MW, 0.8 power factor lagging, 360 kV at R. If
a short-circuit to earth occurs on one phase at S, calculate the fault currents. For
both areas during the subtransient period the negative-, and zero-sequence
reactances may be assumed to be 100% and 50% respectively of the positive-
sequence values
Ch.#4

39
 POWER SYSTEM FAULT CALCULATIONS

FIG. 23. (a) Positive-sequence network before the fault. (b) The venin
equivalent for positive-sequence. (c) Sequence interconnection.

for the grid areas, and 100% and 200% for the line. Resistances may be
neglected. Both G and M are solidly earthed.
SOLUTION

This is the (balanced) voltage at the fault prior to the fault and is the value to be
used for Vf in section 4.10.
Fig. 23(b) shows the Thevenin equivalent at the point of fault for the positive-
Ch.#4

sequence network of Fig.23(a). The reactance’s of the positive-, negative-, and

40
 POWER SYSTEM FAULT CALCULATIONS

zero-sequence networks are j7, j7 and j 3.83 respectively, and their

interconnection is shown in Fig. 23(c). Thus the fault current flowing from the a
phase conductors into the fault at S is given by [28], with Ean replaced by Vf as
discussed in section 4.10, as

The positive-sequence current in the a phase fed into the fault from the generator
is

whilst that from the motor is

The negative-sequence currents are the same, while the zero-sequence current in
the a phase fed into the fault from G is

and that from M is

To these fault currents must now be added the load current, which is wholly
positive-sequence, in order to give the actual current in any part of the system.
The positive-sequence current in phase a flowing from the generator, is
(0.472—j 1.56)+(0.24—j 0.18) = (0.712—j 1.74) kA.
The positive-sequence current in phase a, flowing from the motor, is
(0.67—j 1.51)—(0.24—j 0.18) = —0.173—j 1.33 kA.
Thus the actual current in phase a, flowing from the generator is
(0.712- j 10.74)+(0.472—j 10.56) +(0.517—j 11.57) = 1.701 —j 32.87 kA.

The actual current in phase a, flowing from the motor is

(— 0.173—j 1 .33)+(0.061—j 1.51)+ (6.023—j 0.527) = —0.083—j 3.367 kA
(or 0.083+j 3.367 kA to the motor).
Ch.#4

The sum of these actual currents in phase a is

41
 POWER SYSTEM FAULT CALCULATIONS

(1.701 —j 32.87)+(—0.083.—j 3.37) = 1.618—j 36.24 kA which is the actual fault

current Iae already found.
The currents in the b and c phases of the system can be calculated using [9].
An alternative solution, which gives the actual currents (load plus fault) directly,
can be obtained by calculating the e.m.f.s of the generator and the motor, prior to
the fault, and by assuming that these e.m.f.s remain unchanged during the fault.
The e.m.f.s of the generator and motor are
EG= (208+j 0)+(0.24—j 0.18)(j48) = 216.6+j 11.52 ky/phase
EM = 205.1 —j 3.84 kV/phase.
The sequence-interconnection network for the earth fault is similar to that of Fig.
23(c) except that there are now two separate sources of positive-sequence e.m.f.
EG and EM, each feeding into their own section of the positive-sequence network. If
the whole network is reduced to an equivalent delta network, the student can
check that the reactance between the e.m.f. sources isj 105.1, the reactance
across the generator e.m.f. is j 20.39  and the reactance across the motor e.m.f.
is j 14.28 . The current from the generator in phase a, is
(216.6+j 11.52)/j 20.39 = 0.566—j 10.6 kA
and that from the motor is
(205,1 —j 3.84)/j 14.28 = —0.0268—j 1.44 kA.
The sum of these two currents is clearly the zero-sequence current flowing in Fig.
23(c), and three times this current equals the fault current already found.

4.11 Symmetrical Component Filter Networks

These networks are used in 3-phase power systems to measure symmetrical
components of current or voltage. A zero-sequence current usually indicates an
earth fault on the system but it could also be due to unbalanced leakage currents
over the surface of dirty insulators or to unbalanced line-to-earth charging
currents. Zero- sequence voltage is used to polarize directional earth-fault relays.
Negative-sequence currents Causes heating of alternator rotors. positive- sequence
voltage is sometimes used to operate the automatic voltage regulator (AVR)
controlling the excitation of an alternation
5.1 SEQUENCE CURRENT
Zero-sequence current can be measured by connecting an ammeter (or relay)
across the three identical line current transformers (C.T.s) connected in parallel
(residually connected) (see equation [10 ]). If the primary circuit is a 3-core cable,
Ch.#4

42
 POWER SYSTEM FAULT CALCULATIONS

only one C.T. is necessary and is fitted over the cable, but the cable sheath and
armour must be cut and the ends bonded together outside the C.T. These
residually connected C.T.s measure residual current which equals three times the
zero-sequence current.
A circuit for measuring negative-sequence current is shown in Fig.24. Thus

Since (Za+Zc) cannot be zero the zero-sequence current is suppressed by using

auxiliary, or interposing, C.T.s. A simple method is to insert two C.T.s (all four
C.T.s being identical) in the b line and connect each in parallel, but with reversed
polarity, across one of the existing C.T.s. The student should now analyse this
modified circuit and show that the ammeter current is independent of the positive-
sequence current if Za and Zc are numerically equal and their impedance angles
differ by 120°.
This condition must be satisfied without involving a negative resistance. One
possible solution is
Za = Z’/—90° and Zc = Z’/+30° ohms
where Z’ is any ohmic value less than the rated ohmic burden of the C.T.s. If the
ammeter impedance Z is made very much less than Z’, the student should show
that the negative-sequence current is one- third of the ammeter reading: he
should also consider how to modify this circuit so that the ammeter gives a reading
proportional to the positive-sequence current, independent of the negative-, and
zero-sequence currents. (In Fig..24 the C.T.s were placed arbitrarily in the a and c
lines with the polarities shown. Is there a better choice, especially one for which
one of the two impedances is a pure resistance?). The action of the above circuit
should be checked by assuming that there is an earth fault on line a.
Ch.#4

43
 POWER SYSTEM FAULT CALCULATIONS

Fig. 24. Circuit to measure negative-sequence current.

5.2 SEQUENCE VOLTAGE

Zero-sequence voltage can be measured by using a 3-phase voltage transformer
(V.T.) with a solidly-earthed, star-connected primary and an open (or broken)
delta-connected secondary across which is connected a voltmeter. The voltmeter
will measure the phasor sum of the three line-to-earth voltages (residual voltage)
which equals three times the zero-sequence voltage. The V.T. must not be of the
3-limb, core-type construction.
A circuit for measuring positive-sequence voltage is shown in Fig. 25. The two
V.T.s are identical. Thus

Solving these equations, the current in the ammeter is given by

The three line voltages Vab, Vbc, Vca, have no zero-sequence component. Taking
Vab as the reference phasor

This equation is independent of the negative-sequence voltage when

Ch.#4

Zb= Za /120°, i.e. the two impedances are numerically equal and their impedance

44
 POWER SYSTEM FAULT CALCULATIONS

angles differ by 120°. A possible solution is

Za = Z’/—90° and Zb = Z’/30°. Assuming that the ammeter impedance can be
neglected compared with Z’ (which, being the burden on a V.T. will have a high
ohmic value), the student should show that the ammeter current is Vab1/Z’. The
action of the above circuit should be checked by assuming that the lint voltages
form a balanced 3-phase system, and that the ammeter impedance is zero.

Fig. 25. Circuit to measure positive-sequence voltaje

REFERENCES:
LACKEY, C. H. W. 1951. Fault calculations. Oliver and Boyd, Edinburgh.
MORTLOCK, 3. R. & HUMPHREY DAVIES, lvi. W. 1952. Power system analysis.
Chapman and Mali, London.
WAGNER, C. F. & EVANS, R. B. 1933. Symmetrical components. McGraw-Hill,
New York.
WESTINGHOUSE (Ed.) 1964. Electrical transmission and distribution reference
book. Westinghouse Electric Corporation, Pennsylvania, U.S.A.
Ch.#4

45
 POWER SYSTEM FAULT CALCULATIONS

APPENDICES OF CH. 4
POWER SYSTEM FAULT CALCULATIONS

Appendix 1
THE PER-UNLT SYSTEM
The per-unit value of any quantity is defined as the ratio of that quantity to an
arbitrarily chosen value having the same dimensions and called the base value.
Thus a per-unit value is dimensionless. Base values will be indicated by a subscript
b, and per-unit values by a subscript pu.
If a single-phase supply of voltage V feeds a current I amperes to a load of
impedance Z and a base of Vb volts is chosen, the per-unit value of the supply
voltage is Vpu= V/Vb. Similarly the supply current could be given as Ipu= I/Ipu.
There is a base impedance
Zb= Vb/lb . Applying the per-unit system to impedances,

Thus the per-unit system obeys Ohm’s law (and hence also all the other network
laws).
1.0 per-unit value of any quantity equals the base value of that quantity, so that
per-unit values are relative to 1. Multiplying per-unit values by 100 gives the
percentage (%) values which are relative to 100. The student should now show
that Ohm’s law is not true in the percentage system of units. Hence any power
system data given in the percentage system is best changed to the per-unit system
before being used in any calculations.
The choice of base values is arbitrary, so clearly the choice cannot affect the
ultimate answer to any given problem. Since Zb= Vb/lb  only two of these three
can be arbitrarily chosen. Further, the choice also fixes the base volt-amperes Sb =
VbIb. Volt-amperes (and hence also power and reactive power) can be manipulated
in the per-unit system since

The reason for using the conjugate of current in [A 1.2] is given in Appendix 4.
The original quantities V’ I, Z, S can all be represented by complex numbers
having phase-angles. It is essential that the per-unit system should preserve these
Ch.#4

phase-angles. Hence it must be clearly understood that all base values are scalars

46
 POWER SYSTEM FAULT CALCULATIONS

having no phase-angles. Thus the phase-angle of a per-unit quantity is the phase-

angle of the original quantity in its numerator. The per-unit system obeys the rules
of complex algebra.
If Z= R±j X (equivalent series circuit) then Zpu = Rpu±j X. If Y = G±jB mhos
(equivalent parallel circuit) then Ypu = Gpu±jB . Also Yb = 1/Zb, and Ypu = Y/Yb
There is nothing strange about the per-unit system: it is in everyday use and
makes certain numerical data much more meaningful than it otherwise would be. If
the current in a circuit is given as 30 A, that information by itself is not very
informative unless the rated current is also given, as only then do we know
whether the circuit is lightly loaded or overloaded. If the rated current were given
as 40 A, then the circuit is on ¾ rated load and the current is 75% or 0.75 per-unit
of rated current. Similarly, the student is quite familiar with efficiency given as a
percentage value, power factor as a per-unit value and slip as either a percentage
or per-unit value. Conversely, the per-unit value of a current is meaningless unless
the base current is also given. The distinction between base values and rated
values is dealt with in the discussion which follows. The (thermal) rated current of
any circuit in the appropriate British Standard Specification (B.S.S.).
Formulae which are true for the usual system of units, are also true in the per-unit
system: e.g. take a series circuit of total impedance Zt and components Z1, Z2...
then
Zt= Z1+Z2+.... Dividing through by Zb changes the formula to the per-unit
system while leaving it still true. Thus
Ztp.u = Z1pu+Z2pu+....
Clearly the initial statement above must be qualified by adding ‘provided that only
one (common) base value is used’. This condition may become clearer by
substituting Zb = Vb/Ib, giving

This equation shows that ohmic impedances must be given at a common voltage
level, Vb (this point is important in connection with transformers and will be dealt
with later in this Appendix), and also that a common base current Ib must be used.
In general, different items of plant in a power system will have different rated
currents and voltages, so it is necessary, when using the per-unit system, to
change to a common or base value of current and of voltage. Hence the use of
base value rather than rated value.
Ch.#4

It follows from [A1.3] that the formula relating the ohmic value of an impedance Z

47
 POWER SYSTEM FAULT CALCULATIONS

to its per-unit value is

Thus the per-unit impedance of equipment at a certain MVA, i.e. a certain current
and at rated voltage, is the fraction given by dividing the phase impedance
voltage, when that current is flowing, by the normal rated phase voltage.
Worked example A.1
Two impedances, 1+j2 and 2-f-j2, are connected in series across a 250-V supply.
Calculate the total impedance and load current using the per-unit system.
SOLUTION
Base 200 V, 20 A (the choice is arbitrary)

The impedance voltage (IZ)/phase of a machine is defined as the product of its

rated phase current in amperes and its impedance in ohms/phase. If the
impedance voltage, in volts/phase, is divided by the rated phase voltage, it will be
seen from [A1.4] that the quotient is an impedance in per-unit. Thus if an
alternator has a (synchronous) impedance of 200% or 2 p.u., then on full-load
current, its impedance voltage IZs is twice its rated phase voltage.
A point which often causes confusion is whether a machine’s per-unit impedance
refers to its rated current or to some arbitrarily chosen base current. Either can be
used, but which it is should be clearly stated: if it is not stated, rated current will
usually be assumed since this is the known quantity.
An advantage of the per-unit system is that if a particular type of machine is
taken—say, a non-salient, 2-pole alternator—its impedance in /phase varies over
a wide range depending on its rating. But if this same impedance is expressed in
per-unit based on its thermal rating, it varies over a much smaller range and is
often a fixed (standardised) value.
Ch.#4

48
 POWER SYSTEM FAULT CALCULATIONS

The per-unit system of units can also be applied to a 3-phase circuit. The simplest
procedure, for balanced 3-phase circuits, is to use equivalent per-phase-star (line-
to-neutral) values for all original and base quantities. When dealing with balanced
3-phase circuits, it will always be assumed, unless otherwise stated, that the circuit
is star-connected: i.e. the impedances take line current at line-to neutral voltage.
The per-unit value of the line-to-line voltage equals the per-unit value of the line-
to-neutral voltage and the per-unit value of the 3-phase volt-amperes equals the
per-unit value of the volt-amperes! phase, provided always that the base values

form a balanced 3-phase system, i.e. base line voltage equals 3 x base
voltage/phase and base 3-phase volt-amperes equals 3 x base volt-
amperes/phase. In the per-unit system, the volt-amperes (whether 3-phase or per
phase) equal the product of the voltage (whether line or phase) and the current

and there are no factors such as 3 or 3.

The per-unit system is a completely consistent system of units which satisfies all
the usual electrical circuit formulae. Thus any two (but only two) base quantities
can be arbitrarily chosen, as all the other base quantities can be derived from the
two specified. 3-phase MVA and line kV are the two quantities specified (as these
are usually given on the rating plate of any machine), but the base quantities
chosen are the corresponding phase values, or 3-phase MVA and phase voltage.
Thus if the base values are taken as 100 MVA (3-phase) at 132kV (line) then

I p.u. voltage = 132/ 3 = 76.1 kV/phase

I pu. current 100/( 3 x 132) = 0.4373 kA

I p.u. impedance = 76.1/0.4373 = 174.3 /phase (star).
Equation [A1.4] may be re-written as

This formula relates an ohmic impedance to its per-unit value, in terms of the Lwo
normal base values. A given alternator, line or cable (transformers will be dealt
with later) has a fixed ohmic impedance. Hence it follows from [A 1.5 ]that
Ch.#4

49
 POWER SYSTEM FAULT CALCULATIONS

Thus if a machine’s per-unit impedance is given relative to its thermal rating, IL

can be changed to any base value using the above proportionality.

Worked example A.2

A 3-phase alternator rated at 50 MVA, 12-8 kV has a (synchronous) reactance
phase of 20 pu. on its Thermal rating. It supplies an 11-kV overhead line 5 km long
each phase of which has an impedance of 0.22 +j0.51  km. Calculate the total
impedance on a base of
100 MVA (3-phase) and 11 kV.
SOLUTION

The student is strongly recommended to avoid memorising equations other than

the minimum of basic relationships, (which tends also to reduce the risk of error)
and for this reason [A 1.4]has been used in Example A.2 rather than [A 1.5]. It
should be noted however that engineers accustomed to making particular
calculations, are able to reduce the amount of work by using other relationships.
An example of this is that [A1.5] gives the alternator reactance
j2.0x(100/50)X(l2.8/11)2 = j5.42 pu. From [41.4] the student should be able to
demonstrate that if a machine has a resistance of I % on its thermal rating, then
its full-load I2R loss is 1% of its VA rating.
A major advantage of using the per-unit method in power system analysis, arises
when the system contains transformers. An analysis using ohmic impedances
Ch.#4

would involve transferring impedances through the transformers, using the square

50
 POWER SYSTEM FAULT CALCULATIONS

of the turns-ratio, until all impedances had been referred to a common voltage
level. Having calculated the current at that voltage level (for any given conditions)
that current would then need to be transferred through the transformers, using the
inverse turns-ratio, to find the current at any other voltage level.
It will now be shown that the per-unit value of an impedance is the same on both
sides of a two-winding transformer, and similarly so is the per-unit value of a
current, provided that the base values on both sides of the transformer are
properly related by transformer action. If subscript p refers to the primary side and
subscript s to the secondary side, then given the base values Vb and Ibp on the
primary side, the base values on the secondary side are (assuming for si rnplicity a
single-phase transformer)

If Zp, is the ohmic impedance of the primary winding, its per-unit value is

When the secondary impedance is referred to the primary winding, the total
transformer ohmic impedance is Zb+Zs(Np/Ns)2, and the per-unit impedance of
the whole transformer on the primary side

Thus the total per-unit impedance of the transformer, for any base MVA and
voltage, is the sum of the per-unit impedances of the primary and secondary
windings, and is therefore the same on both sides of the transformer. Clearly the
impedance of any equipment connected to the primary and at the same voltage
may be included with Zp and the impedance of any equipment connected to the
secondary may be included with Zs. The base values could equally well have been
Ch.#4

in volt-amperes and volts. Clearly base volt-amperes is the same on both sides of a

51
 POWER SYSTEM FAULT CALCULATIONS

two-winding transformer. The problem of 3-winding transformers, where the

ratings of the three windings are not necessarily all equal, is dealt with in section
6.9.
Thus in power system analysis using the per-unit method, a transformer can be
regarded as equivalent to a series impedance and the transformation of current
and voltage qan be disregarded. The ultimate answers must be given in the usual
System of units and are obtained from the per-unit values using the base values
corresponding to the voltage level at-which they are required.
The above analysis is valid for a single-phase transformer or for a star/star
transformer where the symbols refer to the per-phase- star (line-to-neutral)
values. It will now be shown that the per-unit value of a phase impedance is the
same in star or in delta. Consider a delta winding having an impedance of Z
/phase, connected to a 3-phase supply of base line voltage VbL with a

corresponding base phase voltage Vb=VbL/ 3 . The base current in one phase of
the winding is Ib with a corresponding base line current

IbL = 3 Ib and

The equivalent star impedance is Z/3 (see Appendix 2) and its per-unit value is

Thus if a transformer impedance is specified in per-unit, it is not necessary to know

whether the windings are in star or delta, except in the final analysis when the
current distribution within the transformer is required. In the case of a delta/star
transformer, the 30° phase shift must also be accounted for (see sections 6.3 and
8.3.5).
When using [A 1.4] or [A 1.5] it is essential that all quantities on the right-hand
side are specified on the same side of the transformer.
Ch.#4

52
 POWER SYSTEM FAULT CALCULATIONS

APPENDICES OF CH. 4 Cont.

POVER SYSTEM FAULT CALCULATIONS
Appendix 2
STAR-DELTA TRANSFORMATION
Given three (and only three) terminals (Fig. A2.1) in any circuit, a star-connection
of three impedances can be replaced by a delta- connection of three impedances,
without affecting the remainder of the circuit, and the converse is true. Such
impedances are referred to as 5quivalent impedances. As with most network
theorems this refers to linear impedances.

Fig. A.2. 1. Star-delta transformation.

The usual proof calculates the impedances between the three pairs of terminals in
both circuits and equates them: this gives three simultaneous equations for three
unknowns. Thus, given the three delta impedances, it can be shown that

By visualizing the two circuits superimposed, this formula can be generalised as

Similarly, given the star impedances

Ch.#4

53
 POWER SYSTEM FAULT CALCULATIONS

If the three impedances are identical the formulae reduce to

It has been shown in Appendix I that the formulae above are true for per-unit
impedances; but the per-unit impedance of 3-phase equipment is the same
whether it is star or delta connected.

Fig. A2.2. Current distribution.

(b) Cc.) Cd)

The star-delta transformation is used to simplify complex networks, and often the
current distribution in the circuit is required. Given the delta currents in Fig.
A2.2(a), the current distribution in the equivalent star (b) is obvious. Given the
current distribution in the star circuit (c), the delta currents can be obtained from
relations such as
pZ12 = aZ1—bZ2.
Ch.#4

54