UNIVERSITY OF GONDAR

Institute of Technology
Department of Electrical and Computer Engineering
Computer Aided Power System Analysis (CAPA)
Chapter 3 Fault Studies

By Dawit Adane
 Fault Analysis
• During normal conditions, a power system operates under balanced conditions with all equipment’s
carrying normal load currents and the bus voltages are within the prescribed limits. This condition
can be disrupted due to a fault occurred in the system. A fault in a circuit is a failure that
interferes with the normal flow of current. A short circuit fault occurs when the insulation of the
system fails resulting in low impedance path either between phases or phase(s) to ground. This
causes excessively high currents to flow in the circuit, requiring the operation of protective
equipment’s to prevent damage to equipment. An open circuit fault occurs when one or two of the
circuit path breaks and interrupts the current flow in that paths. This causes the disruption of the
balanced nature of the system.
Faults in power system can be classified into two based on their balanced ness nature
1. Symmetrical(Balanced) faults
2. Unsymmetrical(Unbalanced) faults
• Majority of the faults occur on transmission lines as they are exposed to external elements.
Lightening strokes may cause line insulators to flashover, high velocity winds may cause tower
failure, ice loading and wind may result in mechanical failure of line or insulator and tree branches
may cause short circuit.
• Fault analysis is necessary for selecting proper circuit breaker rating and for relay settings and
coordination.
 Symmetrical Faults:
 A simultaneous short circuit across all three phases. It occurs infrequently (about 5%),
but it is the most severe type of fault encountered. Because the network is balanced, it
is solved on a per-phase basis. The other two phases carry identical currents except for
the phase shift of 120 degree.
This fault represents a structural network change equivalent with that caused by the
addition of an impedance (i.e., Zf) at the fault place. If Zf = 0 the fault is called a solid or a
bolted fault (this is usually the most severe fault that can occur in a system). These faults
can be of two types: (a) line to line to line to ground fault (LLLG fault) or (b) line to line
to line fault (LLL fault). The faulted network can be solved conveniently by the
Thevenin’s method and ZBUS impedance method.
 Unsymmetrical Faults:
 Faults in which the balanced state of the network is disturbed resulting unbalanced
system operation. The types of unbalanced faults are:
a) Single line to ground fault (LG Fault): - the most common type (Almost 60 to 75% of
faults in a system)
b) Line to line faults (LL faults): - about 5 to 15% of system fault
c) Double line to ground faults (LLG faults): - about 15 to 25% of system fault
d) Open Conductor (Single or Double line open)
 The unsymmetrical faults are analyzed using symmetrical components.
 ZBUS Formation Without Mutual Coupling Between Elements
• For a network with ‘m’ buses and a reference bus, a relation between bus currents and
bus voltages is:

• The above equation can also be written as

Where,
IBUS is (m× 1) bus current injection vector
VBUS is (m× 1) bus voltage vector
ZBUS is m×m bus impedance matrix and is given by,

• The voltage at the ith bus is:

• Then, Zij can be written as

Following points should be noted for the ZBUS matrix

• Zij is the off-diagonal element of ZBUS matrix and is called the ‘open-circuit transfer
impedance’ between ith and jth bus.
• Zii is the diagonal element of ZBUS matrix and is called the ‘open-circuit driving point
impedance’ of ith bus.
• If the YBUS matrix is symmetrical, then the matrix ZBUS is also symmetrical i.e. Zik = Zki.
• Since in a power network each bus is connected to very few other buses, the YBUS
matrix of the network has large number of zero elements and is therefore, sparse in
nature. The ZBUS matrix, on the other hand, is invariably a full matrix.
• Instead of inverting the YBUS matrix of the network to obtain the modified the ZBUS
matrix, a step-by-step ZBUS building algorithm overcomes the problems of every time if
there is a modification in the YBUS matrix due to network modification, avoids the
inversion process to obtain the modified the ZBUS matrix.
 Graph: - The graph of a network describes the geometrical structure of the network
showing the interconnections of network elements.
 The buses are represented by nodes or vertices and impedances by line segments
called elements or edges.
 Tree: - A tree of a graph is a connected sub graph that connects all the nodes without
forming a closed path or a loop. A graph can have a number of distinct trees.
 Branches: - The elements of a tree are called branches. The number of branches ‘b’ of
a tree with ‘n’ nodes, including reference, is given by
b=n–1
 Links: - The elements of a graph not included in the tree of the graph are called links.
Each link is associated with a loop. The collection of links are called a cotree. If ‘e’ is
the number of elements in a graph, then the number of links ‘l’ is given by
l=e–b=e–n+1
Figure below-1 is a single line diagram of a power system having 4 buses, bus (1) to bus
(4) and six elements element e1 to element e6 (bus (0) is taken as the reference bus). Figure
below-2 shows the graph of the network depicting the interconnection of the elements and
the reference node.
Figure 1: Single Line Diagram of a Power System
Figure 2: A graph of the Power system of Fig. 1
• In figure below-3, the branches and the links have been shown with solid lines and
dotted lines respectively.

Figure 3: A tree of the graph of Fig. 2

• The total number of nodes (including reference node) is 5 (i.e. n = 5)
• The number of branches is b = n − 1 = 5 − 1 = 4, where e 1, e2, e5, e6, are such a set of
branches that form a tree of the graph.
• The total number of elements in the graph is e = 6.
• The number of links is = e − n + 1 = 6 − 5 + 1 = 2. The two links in the graph are e 3
and e4 shown with dotted lines in Figure above-3.
 Let us assume that the ZBUS matrix for a partial network with ‘m’ buses and a reference
bus ‘0’, as shown in Figure below-4, exists.
• The bus voltages and currents relationship for the
partial network is:

Where,
is m×1 bus voltage vector
is m×1 bus current injection vector
is m×m bus impedance matrix of the partial
network
• The four possible element additions to a partial network are:
a. Addition of a branch between a new node and the reference
b. Addition of a branch between a new node and an existing node
c. Addition of a link between an existing node and the reference
d. Addition of a link between two existing nodes
1) Addition of a branch between a new node and the reference node (case 1):
• Figure below shows the addition of a branch between a new node ‘q’ and the reference
‘0’. This increases the size of ZBUS to (m+1)×(m+1). Let the impedance of this branch
be zq0. The new network equation can be written as:

Figure: Addition of a branch between a new node and the reference

• It does not change the elements of the original matrix ZBUS but add elements to the new
row and column corresponding to qth bus.
• Since the elements of the power system network are linear and bilateral:
Zqi = Ziq, i = 1, 2, …, m.
• Since a current source of Iq = 1 p.u is connected to the qth bus, with all the others buses
open, and the voltage of qth bus (Vq) is computed as Vq = zq0 Iq, and

• For finding out Zqi, a current source Ii = 1 p.u. is connected between ith bus and the
reference bus with all other buses open circuited as shown in Figure below. Since V q =
0, then

Therefore all the off-diagonal

elements Zq1, Zq2, …, Zmq and
Z1q, Z2q, …, Zqm are equal to
zero.

Figures: For calculation of Zqq & Zqi for Case 1

• Hence, the modified ZBus matrix is given as

2) Addition of a branch between a new node and an existing node (Case 2):
• Let a branch with impedance zpq be connected between an existing node ‘p’ and a new
node ‘q’ as shown in figure below. The size of ZBus matrix increases by one to
(m+1)×(m+1). The modified network equations can be written as:
• The original matrix ZmBus remains unchanged.

Figure: Addition of a branch between an existing node ‘p’ and a new node ‘q’
• For calculating Zqq one can write

• For current source of Iq = 1 p.u is connected to the qth bus, with all the others buses
open circuited, and the voltage of qth bus Vq is computed, as shown in figure below.
• The voltages Vp and Vq can be related as Figure: Calculation of Zqq

• Since, ipq = - Iq = - 1 pu and Vp = Zpq and Vq = Zqq.

Thus,

Figure: Relation between Vp and Vq

• For a current source of Ii = 1 p.u is connected to the ith bus, with all the others buses
open circuited, and the bus voltage Vi is computed for all the buses, as shown in figure
below.

Figure: Calculation of Zqi for case 2

• Since Vq = Vp, the current in the branch p − q is zero. Hence,

Eg.; , , , … & , , , … and

• Hence, the modified ZBus matrix is given as,

3) Addition of a link between an existing node and the reference node (Case 3):
• Adding an element between an existing node and the reference creates a loop and the
size of modified ZBus matrix remains unchanged. However, all the elements are
modified.
• Let the added element, with an impedance of z qo, be connected between an existing
node ‘q’ and the reference node ‘0’ as shown in the Figure below.
Figure: Addition of a link between an existing node ‘q’ and the reference
The current Iℓ flowing through the link modifies the current injected into q th bus from Iq to
Iq − Iℓ. The modified network equations can be written as,

also,
Substituting Vq, results,
or

Together with the above equations, the modified network equations can be expressed as:

The above partitioned matrix can be written in compact form as:

where,
From the above equation, one can write
or

Substituting Iℓ results,

Hence,

Where,
4) Addition of a link between two existing nodes (Case 4):
Let an element with impedance zpq be connected between two existing nodes ‘p’ and ‘q’
and the current Iℓ flowing through the link changes the injected current at p th node from Ip
to (Ip − Iℓ), while at node qth from Iq to (Iq + Iℓ) as shown in figure below.

Figure: Addition of a link between two existing nodes ‘p’ and ‘q’
The modified network equations can be written as:

And
or

Substituting Vp and Vq into the above results;

Together with the above equations, the modified network equations can be expressed as:

Where,
;
;

 
For eliminating the link current Iℓ, the above partitioned matrix can be written in compact
form as:

Where,

From the above equation, one can write

or
Substituting Il results,

Hence,

Where,
Example–1
Modify the bus impedance matrix of a 4-bus power system as shown in the figure below
to account for the connection of a capacitor having a reactance of 5.0 per unit between bus
4 and the reference node of the circuit of the figure below.

The bus admittance matrix of the system w/o modification is:

YBUS =

The bus impedance matrix of the system w/o modification is:

j0.7313 j0.6914 j0.6132 j0.6368
ZBUS = j0.6914 j0.7197 j0.6082 j0.6418
j0.6132 j0.6082 j0.6989 j0.5511
j0.6368 j0.6418 j0.5511 j0.6989
Recognizing the system is modified by adding a link of capacitive reactance of –j5.0 per
unit element from the existing bus (node) of q = 4 to the reference node and Z orig = Zbus is
the 4 X 4 matrix;
Zqo = Z4o = – j5.0 per unit
To find the modified network equation as per rule 3:

The terms in the fifth row and column were obtained by repeating the fourth row and
column of Zorig with negative sign added and note that to obtain the value of element in the
fifth row and column as follow:
Zℓℓ = Zqq + zq0 = Z44 + z40 = j0.69890 – j5.0 = – j4.30110
Then, by eliminating the fifth row and column, we obtain for Z bus(new) as:
…

and other elements in a similar manner or/alternatively

to yield Zbus(new) as
Example–2: Formation of Bus Impedance Matrix
Construct the bus impedance matrix for the network shown in figure below (a). The single
line impedance diagram is shown in figure below (b). The graph, a selected tree and a
cotree for the network is shown in figure below (c) and (d).
The bus admittance and impedance matrix for the network is:
R
R
R
Ex

Exercise – 1
Modify the Zbus determined in above for the circuit of the above Figure by adding a new
node connected to bus 3 through an impedance of j0.5 per unit.
Exercise – 2
Modify the Zbus determined in above by adding a link of impedance j0.2 per unit between
buses 1 and 4 of the circuit of the above Figure.
Answer for Exercise–1
Recognizing the system is modified by adding a branch of inductive reactance of j0.5 per
unit element from the existing bus (node) of p = 3 to the new node q = 5 and Z orig = Zbus is
the 4 X 4 matrix;
zpq = z35 = j0.5 per unit
To find the modified network equation as per rule 2:

Z1q = Z15 = Z1p = Z13 = j0.53340; Z2q = Z25 = Z2p = Z23 = j0.64008; Z3q = Z35 = Z3p =
Z33 = j0.71660; Z4q = Z45 = Z4p = Z43 = j0.66951;
Zq1 = Z51 = Zp1 = Z31 = j0.53340; Zq2 = Z52 = Zp2 = Z32 = j0.64008; Zq1 = Z51 = Zp1 =
Z31 = j0.71660; Zq1 = Z51 = Zp1 = Z31 = j0.66951;
Zqq = Z55 = Zpq + zpq = Z35 + z35 = j0.71660 + j0.5 = j1.21660
Answer for Exercise–2
Recognizing the system is modified by adding a link of inductive reactance of j0.2 per
unit element between the existing bus (node) of p = 1 and the existing bus (node) q = 4
and Zorig = Zbus is the 4 X 4 matrix;
zpq = z14 = j0.2 per unit
To find the modified network equation as per rule 4:
to yield Zbus(new) as
 Symmetrical or Balanced three phase fault analysis:
– Balanced 3-ϕ fault can be solved using Thevenin’s equivalent network as seen from the
fault point. The bus impedance matrix is convenient to use for fault studies as its
diagonal elements are the Thevenin’s impedance of the network as seen from different
buses.
– The generators are represented by a constant voltage source behind a suitable reactance
which may be sub-transient, transient or normal d-axis reactance. The transmission lines
are represented by their π-models with all impedances referred to a common base.
– A balanced three phase fault, through a fault
impedance Zf is assumed to occur at kth
bus of the n-bus power system network
as shown in the figure below. A pre-fault
load flow provides the information about
the pre-fault bus voltage.

Figure: Fault at kth bus of a power system

• Let [VBUS(0)] be the prefault bus voltage vector = [V 1(0) . . . Vk(0) . . . Vn(0)]T p.u. The fault
at kth bus through an impedance Zf will cause a change in the voltage of all the buses
[∆VBUS] due to the flow of heavy currents through the transmission lines. This change can
be calculated by applying a voltage V k(0) at kth bus and short circuiting all other voltage
sources. The sources and loads are replaced by their equivalent impedances as shown in
figure below.

Figure: Network representation for

calculating [∆VBUS]

Zi and Zk are the equivalent load impedances at bus i and k respectively, Z ik is the impedance of
line between ith and kth buses. xdi is the appropriate generator reactance, Z f is the fault
impedance, Ik(F) is the fault current and Vk(0) is the prefault voltage at kth bus.
From the superposition theorem, the bus voltages due to a fault can be obtained as the sum
of prefault bus voltages and the change in bus voltages due to fault, i.e.,

Where,
[VBUS(F)] = Vector of bus voltages during fault = [V1(F) . . . Vi(F) . . . Vn(F)]T
[VBUS(0)] = Vector of pre-fault bus voltages = [V1(0) . . . Vi(0) . . . Vn(0)]T
[∆VBUS] = Vector of change in bus voltages due to fault = [∆V1 … ∆Vk ... ∆Vn]T
The bus injected current [IBUS] can be expressed as,

With all the bus currents, except of the faulted bus k, equal to zero, the node equation for
the network can be written as
As the fault current Ik(F) is leaving the bus it is
taken as a negative current entering the bus.
Hence,
[∆VBUS] can be calculated as:

By Substituting the expression of [∆VBUS], we get

Expanding the above equation results,

The bus voltage of kth bus can be expressed as:

We also know,

Zkk is the Thevenin’s impedance or open circuit impedance of the network as seen from
the faulted bus k.
For a bolted fault Zf = 0 and hence, Vk(F) = 0. Thus the fault current Ik(F) for bolted fault
can be expressed as

For faulty with non-zero fault impedance Zf, the fault current can be calculated as:

The bus voltage after fault for the unfaulty or healthy buses can be written as:

Substituting Ik(F) results,

The fault current Iij(F) flowing in the line connecting ith & jth bus can be calculated as:

where zij is the series impedance of line connecting buses i and j.

If bus i is directly connected to the faulted bus k by a line of series impedance z ik, then the
current contributed from bus i to the current in the fault at bus k is simply:
D

The voltage at each bus during the fault is:

The current flow in line 3 – 1 is:

Fault currents contributed to bus 2 by the adjacent un-faulted buses are

Except for round-off errors, the sum of these current contributions equals I2(F).
Example–2: Power System Analysis – Hadi Saadat – Page 367
 Unsymmetrical or Unbalanced Fault Analysis:
• Unbalanced system operation can result.
• Analysis can be carried out on a three phase basis using symmetrical components
where the three-phase voltages (and currents) which is unbalanced are transformed
into three sets of balanced voltages (and currents) called symmetrical components.
• A transformation of various power system elements (synchronous generators,
transformers and lines) impedances to symmetrical components are decoupled from
each other resulting in independent system networks for each component (balanced
set).
Symmetrical Components
• Any unbalanced set of three phase voltage or current phasors can be replaced by three
balanced sets of three phase voltage or current phasors. These three balanced set of
voltage or current phasors are called symmetrical components of voltages or currents.
• A set of three balanced voltages(phasors) Va, Vb, Vc is characterized by equal
magnitude and interphase differences of 120o.
• A set is said to have a phase sequence abc (positive sequence) if V b lags Va by 120o and
Vc lags Vb by 120o. The three voltage phasors can be expressed in terms of the
reference phasor Va as:
Va = |Va|, Vb = a2|Va| and Vc = a|Va|
• If the phase sequence is acb (negative sequence), then
Va = |Va|, Vb = a|Va| and Vc = a2|Va|
• Where the complex number operator, a, is defined as:

• It has the following property

, and
and
and
, =, =,&
• Any phasor multiplied by ‘a’ undergoes a counter clockwise rotation of 120o without
any change in the magnitude.
• If suffix 1, 2, and 0 is used to indicate positive, negative and zero sequence respectively,
then
1. A set of positive sequence balanced phasors is written as:
Va1, Vb1 = a2Va1, Vc1 = aVa1
2. A set of negative sequence balanced phasors is written as:
Va2, Vb2 = aVa2, Vc2 = a2Va2
3. A set of zero sequence balanced phasors is written as:
Va0, Vb0 = Va0, Vc0 = Va0
• Note: A set of three voltages (phasors) equal in magnitude and having the same phase is
said to be zero sequence phasors.
• Consider a set of three voltages (phasors) V a, Vb, Vc which may be unbalanced.
According to Fortescue's theorem, the three phasors can be expressed as the sum of
positive, negative and zero sequence phasors. Thus,
Va = Va0 + Va1 + Va2 In Matrix form:
Vb = Vb0 + Vb1 + Vb2
Vc = Vc0 + Vc1 + Vc2
• The three phasor sequences (positive, negative and zero) are called the symmetrical
components of the original phasor set Va, Vb, Vc. This is shown by the following figure
below.
Where,
is the arbitrary set of three voltage phasors of phase voltages.
is the set of zero sequence components. The magnitudes of the three zero sequence
components are equal i.e. Va0 = Vb0 = Vc0 and they are co-phasors.
is the set of positive sequence components, with V a1 = |Va1|0o, Vb1 = |Vb1|-120o, and Vc1
= | Vc1|120o, with |Va1| = |Vb1| = |Vc1|.
is the set of negative sequence components, with V a2 = |Va2|0o, Vb2 = |Vb2|120o, and Vc2
= | Vc2|-120o, with |Va2| = |Vb2| = |Vc2|.
• Let us express the above matrix equation in terms of reference phasors V a0, Va1 and Va2.
Thus,

• Expressing the above equations in a compact form as:

• Where,
= vector of phase quantities
= vector of sequence quantities
is the symmetrical component transformation matrix.
• We can also get the symmetrical components from phase quantities as:

Where,

• By expanding
• Similarly, for current

• where, is the set of phase voltages, and is the set of sequence voltages.
Symmetrical Y and ∆ Circuits 
Relationships between the symmetrical components of Y and ∆ currents and voltages can
be established by referring to Figure below, which shows symmetrical impedances
connected in Y and ∆. Let us agree that the reference phase for ∆ quantities is branch a-b.

Figure: Symmetrical impedances: (a) ∆-connected; (b) Y -connected.

• For currents in ∆ connection, we have

• Adding all the above three equations together, results zero-sequence line current as:

• Since line currents into a ∆-connected circuit have no zero-sequence currents.

• The phase ‘a’ line current is

• Note: Assuming zero value of circulating current Iab0 exists in the ∆ circuit.
• From symmetrical component relation, Ica1 = aIab1 and Ica2 = a2Iab2, we have:

• A similar equation for phase b is

• Expressing Ib1, Ib2, Ibc1, and Ibc2 in terms of Ia1, Ia2, Iab1, and Iab2 and solving the above two
equations simultaneously, results:
and
• In a similar manner, the line-to-line voltages can be written in terms of line-to-neutral
voltages of a Y-connected system,

• Adding together all the above three equations, results zero-sequence line voltage as:

• The line-to-line voltages in Y-connected circuit have no zero-sequence components.

However, the sum of phase voltages (line-to-neutral) may not be zero so that their zero
sequence component Van0 = Va0 may exist.
• Since the sum of the three line currents equals the current in the neutral wire, we have:
• The current in the neutral is three times the zero sequence line current. In the absence
of a neutral connection, the neutral current and therefore the zero sequence line current
is always zero.
• The line voltage Vab is:

• Solving using the same method as ∆-connection, we obtain:

• If the voltages to neutral are in per unit referred to the base voltage to neutral and the
line voltages are in per unit referred to the base voltage from line to line, the √3
multipliers must be omitted from the above equations. If both voltages are referred to
the same base, however, the equations are correct as given. Similar analogy works for
line ∆ currents.
• when there are no sources or mutual coupling inside the ∆ circuit:
By substitution,

So that,

• This shows that the ∆-connected impedances is three times the per phase or Y-
connected impedances when both are expressed in ohms or in per unit on the same
kilovolt-ampere and voltage bases.

Power Invariance
The sum of powers of the three symmetrical components equals to the three-phase power.
Total complex power in a three-phase circuit is given by
• If there is impedance in the neutral connection to ground, then the voltages V a, Vb and
Vc must be interpreted as voltages from the line to ground rather than to neutral.

Now,

Therefore,
 Sequence Networks and Sequence Impedances of Power System
Sequence Networks of a loaded Synchronous Generator:
• A three-phase synchronous generator, having a synchronous impedance of Z s per
phase, with its neutral grounded through an impedance Z n is shown in Figure below.
The generator is supplying a balanced three phase load. The generator induced
voltages Ea, Eb and Ec are balanced and hence treated as positive sequence set of
voltage phasors and can be expressed as:

Figure: Three phase synchronous generator supplying a balanced load

• As the generator is supplying a three-phase balanced load, the terminal voltages of the
generator can be obtained using KVL as:

And,

Substituting the neutral current, and writing the resulting equation in matrix form, we get:

The above matrix equation can be expressed in a compact form as:

Where,
is the vector of terminal phase voltages.
is the vector of terminal phase currents.
is the impedance matrix.
Replacing the phase quantities of the above equations by corresponding sequence
quantities, using the transformation equation results,

Pre-multiplying both sides of the above equation by [A] -1 and after simplifications one
gets:

Where,
is Generator Sequence Impedance Matrix and is defined as:

Since a synchronous machine is designed with symmetrical windings (balanced), it

induces EMFs of positive sequence only, i.e., no negative or zero sequence voltages are
induced in it. Thus,
Substituting the above result, we get:

where, Z1 = Zs is the positive sequence generator impedance, Z 2 = Zs is the negative

sequence generator impedance and Z0 = Zs + 3Zn is the zero sequence generator
impedance.
By expanding the above equation, the separate sequence equations are:

It is evident that the three sequence components are independent of each other. The current
of a particular sequence produces a voltage drop of that sequence only, hence the three
sequences are decoupled from each other. The three sequence networks of a synchronous
generator are shown in Figure below.
Sequence networks of a transmission line:
A fully transposed three-phase line is completely symmetrical and therefore the per phase
impedance offered by it is independent of the phase sequence of a balanced set of currents.
Therefore, for a static device such as a transmission line, the phase sequence of voltages
and currents have no effect on the impedance offered by the line as both positive and
negative phase sequences encounter identical line geometry. Hence, the positive and
negative sequence impedances offered by a line are identical i.e. Z 1 = Z2.

Figure: The sequence networks of a transmission line

Sequence Networks of a Transformer:
The shunt magnetizing branch of transformer is neglected as the current through it is
negligible as compared to short circuit current. The transformer is, therefore, modelled
with an equivalent series leakage impedance. Since the transformer is also a static device
the positive and negative sequence impedances offered by a transformer are equal to the
series leakage impedance of the transformer.

The sequence networks for zero sequence depends on the winding connections and
whether or not the neutrals are grounded. Before considering the zero sequence networks
of various types of transformer connections, three important observations are made:
 Figure: The flow of zero sequence currents in a star connection

Figure: The flow of zero sequence currents in a

delta connection
(a) Star-Star connections with both neutrals grounded:
Since both the neutrals are grounded, the phasor sum of three unbalanced phase currents is equal
to three times the zero sequence current I a0. Hence, the zero sequence currents can flow in the
primary and secondary windings and the transformer, therefore, can be represented by the
equivalent zero-sequence leakage impedance as shown in figure below.

Figure: The zero-sequence equivalent circuit of a Star-Star transformer with both neutrals
grounded
(b) Star-Star connections with only one neutral grounded:
When the neutral of only one winding is grounded, the phase currents of the ungrounded winding
must add up to zero. This implies that the zero sequence currents cannot exist in the ungrounded
winding and hence the zero sequence currents cannot exist even in the transformer side with
neutral grounded.
Figure: The zero-sequence equivalent circuit of a Star-Star transformer with only one
neutral grounded
(c) Star-Star connections with only no neutral grounded:
The zero sequence currents can not exist on any winding in this case also.

Figure: The zero-sequence equivalent

circuit of a Star-Star transformer
with both neutrals ungrounded
(d) Star-delta connections with neutral grounded:
A zero sequence current on the grounded star winding will cause a circulating zero-
sequence current in the closed delta-winding. However, the zero-sequence current on the
delta-winding cannot exist on line side of the winding and is confined only to the closed
delta-winding. As a result, an open circuit exists between the star and the delta sides.

(e) Star-delta connections with ungrounded neutral:

Since the neutral is isolated, no zero sequence current can exist in the star side of the
transformer and as a consequence zero-sequence currents cannot exist in the delta side.
(f) Delta-delta connections with ungrounded neutral:
The zero-sequence currents can only circulate within the closed delta windings and cannot
exit on line sides of both the windings. Hence, an open circuit exists between the two
windings as far as zero-sequence currents are concerned.

Point to remember: If the neutral of a transformer is grounded through a grounding

impedance Zn, as shown in Figures above, then, the total zero-sequence equivalent
impedance to be used in the equivalent circuit is
 Unbalanced SC Fault Analysis Using Thevenin’s Equivalent Networks
1. Single Line to Ground Fault (LG)
Figure below shows a LG fault at F in a power system through a fault impedance Z f. The
phases are so labeled that the fault occurs on phase a.

Figure: Line-to-ground fault on phase a

At the fault point F, the currents out of the power system and the line to ground voltages
are constrained as follow:
and
Substituting Ib = Ic = 0, the symmetrical components of the fault currents are

From which we find that

Phase a voltage in terms of the symmetrical component is

Since all sequence currents are equal and the sum of sequence voltages equal 3Z fIa1, a
series connection of sequence networks through an impedance 3Z f results as shown in
figure below.
Substituting for Va0, Va1, and Va2 and noting Ia0 = Ia1 = Ia2, we get

Where, Z0 = Zs + 3Zn, then

The fault current Ia is given by;

Figure: Connection of sequence network for a single line-to-ground (LG) fault

The voltage of line b to ground under fault condition is

Substituting for Ia, and rearranging the above equation, we obtain:

Similarly, the line to ground voltage of phase c under fault condition is

2. Line-To-Line Fault
Figure below shows a three phase generator with a fault through an impedance Z f between
phases b and c. Assuming the generator is on no-load, the boundary conditions at the fault
point are:
By substituting, the symmetrical components of currents are
From the above equation, we find that

We note that

We have

Substituting for Va1 = Ea – Z1Ia1 and Va2 = – Z2Ia2 and noting Ia2 = –Ia1, we get

By substituting for , we obtain,

Solving for Ia1 results in

The fault current is

Thus,
&
The symmetrical components of voltages under fault condition are

By disintegrating and rearranging, we get

By subtracting the above two equations results

Substituting for Ib and simplifying, we get

Figure: Connection of sequence networks for line-to-line (LL) fault

The phase to ground voltage of each phases are:
3. Double Line-To-Ground (LLG) Fault
Figure below shows a three phase generator with a fault between phases b and c through
an impedance Zf to ground. Assuming the generator is on no-load, the boundary
conditions at the fault point are:

Figure: Double line-to-ground (LLG) fault

The phase voltages b and c are:

Since Vb = Vc, from the above equations we note that,

By substituting the symmetrical components of currents, we get

Substituting for Vb and Va2 from the above expressions, we have

For Ia2
Since Ia1 = -(Ia0 + Ia2) and substituting for Ia0 and Ia2, solving for Ia1;

Then, Ia0 and Ia2 is:

The fault current is obtained as:

The symmetrical components of voltages are given by

From which it follows that

From the above two equations,

The phase voltages are given by

Figure: Connection of sequence networks for a double line-to ground (LLG) fault
Example–4
B

The positive sequence impendence network is shown in Figure below.

G

Combining the parallel branches, the positive-sequence Thevenin impedance is:

This is shown in Figure below (a)

(a) +ve–sequence (b) –ve – sequence network

network
D

Combining the parallel branches, the zero-sequence Thevenin impedance is:

D
D
The sequence component of the fault currents are

The fault current is

 The positive- and negative-sequence components of the fault currents are:

The fault currents are:

The positive-sequence component of the fault current is:

The negative-sequence component of the fault current is:

The zero-sequence component of the fault current is:

The phase currents are:

The fault current is:

Unbalanced Fault Analysis Using Bus Impedance Matrix
• We have observed that for a fault at bus k, the diagonal element in the k-axis of the bus
impedance matrix, ZBUS, is the Thevenin impedance to the point of fault. In order to
obtain a solution for the unbalanced faults, the bus impedance matrix for each
sequence network is obtained separately, then the sequence impedances Z 0kk, Z1kk, and
Z2kk are connected together as described by the figures of corresponding unbalanced
fault types.
Single Line to Ground Fault Using ZBUS
• For a fault between phase a and ground through an impedance Z f at bus k, the
symmetrical components of fault current are:

• Where Z0kk, Z1kk, and Z2kk are the diagonal elements in the k-axis of the corresponding
bus impedance matrix of sequence network and Vk(0) is the pre-fault voltage at bus k.
• The fault phase currents are:
The fault current is:

Line-To-Line Fault Using ZBUS

For a fault between phase b and c through an impedance Z f at bus k, the symmetrical
components of fault current are:

The fault phase currents are:

The fault current is:

Double Line-To-Ground (LLG) Fault Using ZBUS

For a fault between phase b and c through an impedance Z f to ground at bus k, the
symmetrical components of fault current are:
The fault phase currents are:

The fault current is:

Bus Voltages and Line Currents During Fault

The symmetrical components of the ith bus voltages during the fault are obtained as:
Where Vi1 (0) = Vi (0) is the pre-fault voltage at bus i.
The phase voltages during fault are:

The symmetrical components of fault current in line i to j is given by;

Where , , and are the zero-, positive-, and negative-sequence components of the actual
line impedance between buses i and j.
The phase fault current in line i to j is:
Example–5

Using the bus impedance matrix formation technique, and are found for positive
sequence network and zero sequence network of corresponding Figures in
Example-4.

And the zero-sequence bus impedance matrix is

D
D
D
D
D
D
D
D
D
D
D
D
 Open Conductor Fault
When one or two phases of a balanced three-phase line opens it creates an unbalance in
the system and results in the flow of unbalanced currents. Open conductor fault occurs in
the system one or two conductors of a transmission line are broken due to storm or if
fuses, isolators or circuit breakers operate to break one or two conductors and leaving the
other connected.
An open conductor fault is in series with the line. Line currents and series voltages
between conductor broken ends of the conductors are required to be determined. Such
open conductor faults can also be analyzed with the help of [ZBus] matrices of sequence
networks.
Figure below (a) and (b) shows a section of a three phase system between buses i and j for
which one conductor open and two conductors open between points k and k' respectively.
Let [V0] be the vector of open-circuit bus voltages corresponding to the initial (pre-fault)
value of bus current vector [I0] injected in a network with bus impedance matrix [ZBUS].
We can then write

If the bus currents are changed to a new value, [I0 + ΔI], the new bus voltage [V] can be
expressed as:

where, [ΔV] represents the change in the values of the original bus voltage [V0].
Figure below represents a power system with buses i and j taken out along with the
reference node. The circuit is not energized so that [V0] and [I0] are zero. Currents [ΔIi]
and [ΔIj] are injected into the ith and jth buses respectively, through current sources
connected between the node and the reference node.
The changes in bus voltage [ΔV] can be calculated as:
The modified voltage at ith bus can be written as:

adding and subtracting ZijΔIi in the above equation, one obtains

Similarly, the modified voltage at jth bus can be written as:

Since the network is symmetrical, Zji = Zij.

The Thevenin’s Equivalent circuit of the network as seen from the ith and jth buses are
shown in figure below.
From the figure it can be observed that the Thevenin’s open circuit voltage between ith and
jth buses is (Vi0 − Vj0). Vi and Vj voltages are calculated as if the initial voltages Vi0 and Vj0
are set equal to zero and an ideal current source I is connected between the two busses.
Note that ΔIi = I and ΔIj = -I.

Calculate the voltage difference ΔVij between ith and jth buses as:

Hence, the Thevenin’s open circuit impedance can be found as:

The opening of all the three phases is equivalent to the removal of the line i → j totally
from the network. If zij(0), zij(1) and zij(2) are the the three sequence impedance of the line i
→ j, then the removal of this line from the network can be simulated by adding −z ij(0),
−zij(1) and −zij(2) to the corresponding Thevenin’s equivalent network of the three sequence
networks of the original network as seen from ith and jth buses.
Let x represents the fractional length of the broken line i → j from ith bus to the break point
‘k’, where 0 ≤ x ≤ 1.
Figure below illustrates The positive sequence impedance of the conductor segment
between the ith bus and the point of break k is xZij(1), and the positive sequence impedance of
the remaining conductor from point k to j th bus is (1 − x)Zij(1). These two impedances are
then added to represent the broken conductor.

Figure: Positive sequence equivalent

network with line open between
buses k and k’

If Vkk’(a), Vkk’(b) and Vkk’(c) represent the phase component of voltage drops between points k
and k’, then Vkk’(0), Vkk’(1) and Vkk’(2) represent the sequence components of the voltage drops
between points k and k’.
Figure below shows simplification of the circuit by replacing a current source V kk’(1)/Zij(1)
and a parallel impedance Zij(1) from the voltage Vkk’(1) and the total series impedance [xZij(1)
+ (1 −x)Zij(1)] = Zij(1).

Figure: Thevenin’s Equivalent with transformed current source

The final simplified positive sequence impedance equivalent circuit is shown in figure
below as the parallel combination of Zij(1) and −Zij(1) is ∞ and hence, is replaced by an open
circuit.
Figure: Final positive sequence Thevenin’s Equivalent
circuit representing the opening of line I → j
between buses k and k’

Similarly, the negative sequence and zero sequence

equivalent networks are identical to the positive
sequence equivalent network but do not contain
any internal voltage sources as shown in figure below (a) and (b) respectively.

Figure: Final (a) negative sequence

(b) zero sequence Thevenin’s
Equivalent circuit representing
the opening of line i → j
between k and k’
If no conductor is open, then the sequence voltages are all zero and the current sources are
not present in the equivalent circuit. Further, the current sources can be regarded as current
injections into the buses i and j of the original sequence networks.
The current injections at the buses i and j can be tabulated as:

The sequence voltage drops Vn(0), Vn(1) and Vn(2) at any bus ‘n’ due to the current injections
at the buses ‘i’ and ‘j’ can be calculates as:
The positive sequence equivalent impedance Z kk’(1) as seen from the buses k and k’ is
found out as:

Similarly, the negative sequence and zero sequence Thevenin’s equivalent impedances can
be expressed as:

The open-circuit voltage from point k to k’ can be calculated as:

Substituting into the above equation, results;

Also prior to the occurrence of open-conductor fault on any conductor, the current I ij(1)
flowing in phase a is the positive sequence component and is given by:

By substituting the above equation, the final value of is obtained as:

The Thevenin’s equivalent network as seen from points k and k’ for the three sequence
networks are shown in figure below.
There are two types of open-circuit fault, i.e.,
a) One phase/conductor open
b) Two phases/conductors open
One Conductor Open
Consider that phase a conductor is open, hence phase a current Ia = 0. As a result:

where, Ia(1), Ia(2) and Ia(0) are the symmetrical components of phase a current. Since phases b
and c are closed, the voltage drops,

The symmetrical components of voltage drop across the fault point can be calculated as:

Hence,

It implies that open conductor in phase a causes equal voltages to appear across points k
and k’ of each sequence network. Hence, the three equivalent sequence networks can be
connected in parallel across points k and k’ as shown in Figure below.
Figure: Connection of Equivalent
sequence networks to
represent open conductor
a between k and k’

The current Ia(1) is given as:

Simplifying results;
The sequence voltage drops Vkk’(1), Vkk’(2) and Vkk’(0) can be calculated as:

Substituting Ia(1) and simplifying we get:

Then, the currents Ia(2) and Ia(2) are given as:

and
Steps:
1) , and are obtained from the impedance parameters of the sequence networks.
2) The pre-fault current or load current in phase a of the line i → j, Iij, is determined.
3) The equivalent injected currents Vkk’(1)/Zij(1), Vkk’(2)/Zij(2) and Vkk’(0)/Zij(0) are calculated.
4) ΔVi(0), ΔVi(1) and ΔVi(2) representing the changes in the symmetrical components of bus
voltage are calculated.
5) The bus voltages after fault are calculated using superposition principle as:

Two conductors open

When two conductors’ b and c are open, then the circuit conditions are:

The sequence components of line current are:

The above equation indicates that the three equivalent sequence networks are in series and
to ensure the circuit should be closed. The interconnection of the sequence networks is
shown in figure below.

Figure: Connection of Equivalent sequence networks to

represent open conductors’ b and c between k and k’
The sequence currents can be calculated as:

Where Iij is the pre-fault current in phase a.

The sequence voltage can be calculated as:

Remaining calculations are similar to those performed for single conductor open case.