LECTURE NOTES

d
ON

l
or
COMPUTER METHODS IN
POWER SYSTEMS

W
B.Tech III-II semester
TU
JN
ll
A
 CONTENTS (CHAPTER-WISE)

d
Chapter 1 LINEAR GRAPH THEORY

PART A: INCIDENCE AND NETWORK MATRICES

l
or
• Definitions of important terms
• Incidence matrices: Element node incidence matrix and Bus
incidence matrix
• Primitive networks and primitive network matrices
• Performance of primitive networks

W
• Frames of reference
• Singular transformation analysis
• Formation of bus admittance matrix by rule of inspection and
singular transformation
• Examples

PART B: ZBUS BUILDING ALGORITHMS

TU
• Partial networks
• Addition of Branch and addition of Link
• Algorithms for formation of ZBUS for single phase systems
• Special cases without and with mutually coupled elements
• Deletion of an element, changing the impedance value of an
JN

element, etc.
• Examples

Chapter 2

LOAD FREQUENCY CONTROL

ll

• Turbine speed governing system

• Modeling and block diagram representation of single area
• Steady state and dynamic response
A

• Area control error

• Two area load frequency control
• Examples

2
 Chapter 3

LOAD FLOW STUDIES

d
PART A: REVIEW OF SOLUTION OF EQUATIONS

• Solution of linear, nonlinear and differential equations

l
• Importance of iterative methods
• Flow diagrams for the iterative methods

or
• Comparison of methods
• Examples

PART B: BASIC LOAD FLOW ANALYSIS

W
• Classification of buses
• Importance of slack bus
• Static load flow equations
• Gauss Siedel Method: for systems with PQ, PV and Constrained
PV buses
• Limitations of Gauss Siedel method
TU
• Importance of acceleration factor
• Newton Raphson method in polar coordinates
• Concept of Jacobian matrix
• Comparison of load flow methods
• Examples
JN

PART C: ADVANCED LOAD FLOW ANALYSIS

• Decoupled load flow

• Fast decoupled load flow analysis
• Representation of TCUL transformers
• Examples
ll

Chapter 4

ECONOMIC OPERATION OF POWER SYSTEMS

A

• Optimal distribution of loads within a plant

• Langrangian multipliers
• Transmission loss as a function of plant generation

3
 • Determination of loss coefficients
• Automatic economic load dispatch
• Examples

d
Chapter 5

TRANSIENT STABILITY STUDIES

l
• Basic Stability terms

or
• Swing Equation & Swing Curve
• Numerical solution of differential equations
• Modified Euler’s method
• Runge-Kutta IV order method
• Milne’s predictor corrector method

W
• Representation of synchronous machine, loads, etc. for TS studies
• Network performance equations
• Solution techniques and flow charts
• Examples
TU
EXPECTED PATTERN OF QUESTION PAPER

One question each on chapter 2 and 4, and Two questions each on chapters 1, 3 and 5.
Note: Five questions out of 8 are to be answered in full.
JN

TEXTS/ REFERENCES:

1. Stagg and El Abiad, Computer methods in power system analysis, MH.

2. MA Pai, Computer techniques in power Systems, TMH..
3. K Umarao, Computer Techniques and Models in Power Systems, IK
International Publishing House Pvt. Ltd., New Delhi.
4. RN Dhar, Computer techniques in power system operation and control.
ll

PREREQUISITE SUBJECTS:
A

1. Power System Analysis

2. Synchronous Machines
3. Numerical analysis

4
 ELECTRICAL POWER SYSTEMS-

THE STATE-OF-THE-ART:

d
To Begin With ……

l
No. TOPICS SUB-TOPICS

or
Representation of
- p.u. React./ Imp. Diagram
1. Power systems

Electric Power - Generation - Machines

- Transmission - Distribution - Utilization
– Tariffs

W
2. System

- Sym. Faults
- Sym. Components
3. Fault studies - Seq. Imps. / Networks
- Unsymmetrical Faults
TU
- SSS, TS, DS
- Angle Stability
4. System Stability
- Solution of Equations
- EAC, Clarke’s Diagram

- Incidence Matrices
JN

5. Linear Graph Theory - Singular/ NS Transformations

- Network Matrices
(Linear Equations)
- ZBUS Building

Power Flow Studies - Buses, YBUS Advs., Loads flow equations

- Iterative Methods
6.
ll

(NL Equations) - GS, NR, FDLF & DCLF

A

5
 Present Scenario ……

No.
TOPICS SUB-TOPICS

d
- Importance of VArs
- Compensation Devices,

l
Sizing, Placement,
7. Reactive Power Design, Optimality,

or
Management
- VAr Dispatch
- VAr Co-ordination

- Optimality

W
- Load Prediction: Short,
8. Gen. Expansion Medium and Long
Planning Term Forecasting

-EMS: EMC, SLDC,RLDC

Operation and -ALFC, Voltage Control
9. Control -Tie-line Power Control
TU
System - Requirements
10. Reliability - Methods

- Unit Commitment
JN

- Parallel Operation
11. Economic Operation - Optimal Load Dispatch
- Constraints

12. Instrumentation - CTs, PTs

- SCADA
- Bad Data Elimination
ll

13. State Estimation - Security/ Cont. Studies

….. and so on.

A

6
 Future Trends ……

No.
TOPICS SUB-TOPICS

d
14. Voltage Stability - Importance
- Angle/ Voltage stability

l
- Requirements

or
15. Power System - Control Blocks
Simulators - Data-Base Definition

16. Energy Auditing - Deregulation

W
- Time of Use Pricing
17. Demand Side
Management

18. Renewable Energy - The Paradigm

TU
- Sparsity: YBUS
- Ordering Schemes
Sparsity Oriented - LU- Factorization: Fills
19. Programming - Pivoting
- UD Table Storage
JN

- AI
- Expert Systems
Recent Computer - ANN,
20. Applications - Genetic Algorithms
- Fuzzy Logic, Etc.
ll

__________________-
A

7
 CHAPTER- 1-A

INCIDENCE AND NETWORK MATRICES

d
[CONTENTS: Definitions of important terms, Incidence matrices: Element node
incidence matrix and Bus incidence matrix, Primitive networks and matrices,
Performance of primitive networks, Frames of reference, Singular
transformation analysis, Formation of bus admittance matrix, examples]

l
or
INTRODUCTION

The solution of a given linear network problem requires the formation of a set of
equations describing the response of the network. The mathematical model so
derived, must describe the characteristics of the individual network components, as

W
well as the relationship which governs the interconnection of the individual
components. In the bus frame of reference the variables are the node voltages and
node currents.

The independent variables in any reference frame can be either currents or voltages.
TU
Correspondingly, the coefficient matrix relating the dependent variables and the
independent variables will be either an impedance or admittance matrix. The
formulation of the appropriate relationships between the independent and dependent
variables is an integral part of a digital computer program for the solution of power
system problems. The formulation of the network equations in different frames of
JN

reference requires the knowledge of graph theory. Elementary graph theory concepts
are presented here, followed by development of network equations in the bus frame
of reference.

ELEMENTARY LINEAR GRAPH THEORY: IMPORTANT TERMS

ll

The geometrical interconnection of the various branches of a network is called the

topology of the network. The connection of the network topology, shown by replacing
all its elements by lines is called a graph. A linear graph consists of a set of objects
A

called nodes and another set called elements such that each element is identified with an
ordered pair of nodes. An element is defined as any line segment of the graph
irrespective of the characteristics of the components involved. A graph in which a

8
 direction is assigned to each element is called an oriented graph or a directed graph.
It is to be noted that the directions of currents in various elements are arbitrarily
assigned and the network equations are derived, consistent with the assigned
directions. Elements are indicated by numbers and the nodes by encircled numbers.

d
The ground node is taken as the reference node. In electric networks the convention
is to use associated directions for the voltage drops. This means the voltage drop in a
branch is taken to be in the direction of the current through the branch. Hence, we

l
need not mark the voltage polarities in the oriented graph.

or
Connected Graph : This is a graph where at least one path (disregarding
orientation) exists between any two nodes of the graph. A representative power
system and its oriented graph are as shown in Fig 1, with:

W
e = number of elements = 6 l = number of links = e-b = 3
n = number of nodes = 4 Tree = T(1,2,3) and
b = number of branches = n-1 = 3 Co-tree = T(4,5,6)

Sub-graph : sG is a sub-graph of G if the following conditions are satisfied:

TU
 sG is itself a graph 

 Every node of sG is also a node of G 

 Every branch of sG is a branch of G 
For eg., sG(1,2,3), sG(1,4,6), sG(2), sG(4,5,6), sG(3,4),.. are all valid sub-graphs of
JN

the oriented graph of Fig.1c.

Loop : A sub-graph L of a graph G is a loop if

 L is a connected sub-graph of G 

 Precisely two and not more/less than two branches are incident on each node 
in L
ll

In Fig 1c, the set{1,2,4} forms a loop, while the set{1,2,3,4,5} is not a valid, although
the set(1,3,4,5) is a valid loop. The KVL (Kirchhoff’s Voltage Law) for the loop is
A

stated as follows: In any lumped network, the algebraic sum of the branch voltages
around any of the loops is zero.

9
 l d
or
Fig 1a. Single line diagram of a power system

W
TU

Fig 1b. Reactance diagram

JN
ll
A

Fig 1c. Oriented Graph

10
 Cutset : It is a set of branches of a connected graph G which satisfies the following
conditions :
 The removal of all branches of the cutset causes the remaining graph to have

d
two separate unconnected sub-graphs. 

 The removal of all but one of the branches of the set, leaves the remaining
graph connected. 

l
or
Referring to Fig 1c, the set {3,5,6} constitutes a cutset since removal of them isolates
node 3 from rest of the network, thus dividing the graph into two unconnected sub-
graphs. However, the set(2,4,6) is not a valid cutset! The KCL (Kirchhoff’s Current
Law) for the cutset is stated as follows: In any lumped network, the algebraic sum of

W
all the branch currents traversing through the given cutset branches is zero.

Tree: It is a connected sub-graph containing all the nodes of the graph G, but
without any closed paths (loops). There is one and only one path between every pair
TU
of nodes in a tree. The elements of the tree are called twigs or branches. In a graph
with n nodes,
The number of branches: b = n-1 (1)
For the graph of Fig 1c, some of the possible trees could be T(1,2,3), T(1,4,6),
T(2,4,5), T(2,5,6), etc.
JN

Co-Tree : The set of branches of the original graph G, not included in the tree is
called the co-tree. The co-tree could be connected or non-connected, closed or open.
The branches of the co-tree are called links. By convention, the tree elements are
shown as solid lines while the co-tree elements are shown by dotted lines as shown
in Fig.1c for tree T(1,2,3). With e as the total number of elements,
ll

The number of links: l = e – b = e – n + 1 (2)

For the graph of Fig 1c, the co-tree graphs corresponding to the various tree graphs
A

are as shown in the table below:

Tree T(1,2,3) T(1,4,6) T(2,4,5) T(2,5,6)
Co-Tree T(4,5,6) T(2,3,5) T(1,3,6) T(1,3,4)

11
 Basic loops: When a link is added to a tree it forms a closed path or a loop. Addition
of each subsequent link forms the corresponding loop. A loop containing only one
link and remaining branches is called a basic loop or a fundamental loop. These
loops are defined for a particular tree. Since each link is associated with a basic

d
loop, the number of basic loops is equal to the number of links.
Basic cut-sets: Cut-sets which contain only one branch and remaining links are
called basic cutsets or fundamental cut-sets. The basic cut-sets are defined for a

l
particular tree. Since each branch is associated with a basic cut-set, the number of

or
basic cut-sets is equal to the number of branches.

Examples on Basics of LG Theory:

W
Example-1: Obtain the oriented graph for the system shown in Fig. E1. Select any
four possible trees. For a selected tree show the basic loops and basic cut-sets.
TU
JN

Fig. E1a. Single line diagram of Example System

ll
A

Fig. E1b. Oriented Graph of Fig. E1a.

12
 For the system given, the oriented graph is as shown in figure E1b. some of the valid
Tree graphs could be T(1,2,3,4), T(3,4,8,9), T(1,2,5,6), T(4,5,6,7), etc. The basic
cut-sets (A,B,C,D) and basic loops (E,F,G,H,I) corresponding to the oriented graph
of Fig.E1a and tree, T(1,2,3,4) are as shown in Figure E1c and Fig.E1d respectively.

l d
or
W
TU
Fig. E1c. Basic Cutsets of Fig. E1a.
JN
ll
A

Fig. E1d. Basic Loops of Fig. E1a.

13
 INCIDENCE MATRICES
ˆ
Element–node incidence matrix:
A

The incidence of branches to nodes in a connected graph is given by the element-node

ˆ ˆ

d
incidence matrix, A . An element aij of A is defined as under:
aij = 1 if the branch-i is incident to and oriented away from the node-j.
= -1 if the branch-i is incident to and oriented towards the node-j.

l
= 0 if the branch-i is not at all incident on the node-j.

or
ˆ
Thus the dimension of A is e  n, where e is the number of elements and n is the
number of nodes in the network. For example, consider again the sample system

W
with its oriented graph as in fig. 1c. the corresponding element-node incidence
matrix, is obtained as under:

Nodes
0 1 2 3
Elements
TU
1 1 -1
2 1 -1
ˆ 3 1 -1
A =
4 1 -1
5 1 -1
JN

6 1 -1

It is to be noted that the first column and first row are not part of the actual matrix and
they only indicate the element number node number respectively as shown. Further, the
sum of every row is found to be equal to zero always. Hence, the rank of the
ˆ
matrix is less than n. Thus in general, the matrix A satisfies the identity:
ll

∑ aij = 0
A

 i = 1,2,…..e. (3)
j=1

14
 Bus incidence matrix: A

By selecting any one of the nodes of the connected graph as the reference node, the
ˆ
corresponding column is deleted from A to obtain the bus incidence matrix, A. The

d
dimensions of A are e  (n-1) and the rank is n-1. In the above example, selecting
node-0 as reference node, the matrix A is obtained by deleting the column
corresponding to node-0, as under:

l
or
Buses
1 2 3
Elements
1 -1
A
2 -1 b Branches

W
A= 3 -1 =
4 1 -1
5 1 -1 Al Links
6 1 -1
TU
It may be observed that for a selected tree, say, T(1,2,3), the bus incidence matrix
can be so arranged that the branch elements occupy the top portion of the A-matrix
followed by the link elements. Then, the matrix-A can be partitioned into two sub
matrices Ab and Al as shown, where,

(i) Ab is of dimension (bxb) corresponding to the branches and

JN

(ii) Al is of dimension (lxb) corresponding to links.

A is a rectangular matrix, hence it is singular. A b is a non-singular square matrix of

dimension-b. Since A gives the incidence of various elements on the nodes with
their direction of incidence, the KCL for the nodes can be written as
ll

T
A i =0 (4)
T
where A is the transpose of matrix A and i is the vector of branch currents. Similarly
A

for the branch voltages we can write,

v =A Ebus (5)

15
 Examples on Bus Incidence Matrix:
Example-2: For the sample network-oriented graph shown in Fig. E2, by selecting a
ˆ.
tree, T(1,2,3,4), obtain the incidence matrices A and A Also show the partitioned
form of the matrix-A.

l d
or
W
Fig. E2. Sample Network-Oriented Graph

nodes
TU
e \ n 0 1 2 3 4
 
1 1 1 0 0 0 
2 1 0 1 0 0 
 
3 1 0 0 0
ˆ  1

A = Elements  4 0 0 0 1 1 
 
5 0 0 1 1 0 
JN

6 0 1 1 0 0 
 
7 0 0 1 0  1
 
buses
e \ b 1 2 3 4
 
 1 1 0 0 0
 2 0 1 0 0
ll

 
 3 0 0 0 1

A = Elements  4 0 0 1 1
 
A

 5 0 1 1 0
 6 1 1 0 0
 
 7 0 1 0  1
 

16
 Corresponding to the Tree, T(1,2,3,4), matrix-A can be partitioned into two sub-
matrices as under:
buses
b \ b 1 2 3 4
 

d
 1 1 0 0 0 
Ab = branches  2 0 1 0 0 
 
 3 0 0 0 1

l


 
 4 0 0 1 1 

or
buses
l \ b 1 2 4 3
 
5 0 1 1 0 
Al = links 
 6 1 1 0 0 

W
 
7 0 1 0 1
 

Example-3: For the sample-system shown in Fig. E3, obtain an oriented graph. By
ˆ.
selecting a tree, T(1,2,3,4), obtain the incidence matrices A and A Also show the
TU
partitioned form of the matrix-A.
JN
ll
A

Fig. E3a. Sample Example network

Consider the oriented graph of the given system as shown in figure E3b, below.

17
 l d
or
Fig. E3b. Oriented Graph of system of Fig-E3a.

W
Corresponding to the oriented graph above and a Tree, T(1,2,3,4), the incidence
matrices Ậand A can be obtained as follows:

e\n 0 1 2 3 4 e\b 1 2 3 4
TU
1 1 -1 1 -1
2 1 -1 2 -1
ˆ = 3 1 -1 A= 3 -1
A

4 1 -1 4 -1
5 1 -1 5 1 -1
6 -1 1 6 -1 1
7 1 -1 7 1 -1
JN

8 -1 1 8 -1 1
9 -1 1 9 -1 1

Corresponding to the Tree, T(1,2,3,4), matrix-A can be partitioned into two sub-
matrices as under:
ll

e\b 1 2 3 4 e\b 1 2 3 4
1 -1 5 1 -1
Ab = 2 -1 Al = 6 -1 1
A

3 -1 7 1 -1
4 -1 8 -1 1
9 -1 1

18
 PRIMITIVE NETWORKS

So far, the matrices of the interconnected network have been defined. These
matrices contain complete information about the network connectivity, the

d
orientation of current, the loops and cutsets. However, these matrices contain no
information on the nature of the elements which form the interconnected network.
The complete behaviour of the network can be obtained from the knowledge of the

l
behaviour of the individual elements which make the network, along with the

or
incidence matrices. An element in an electrical network is completely characterized
by the relationship between the current through the element and the voltage across it.

General representation of a network element: In general, a network element may

W
contain active or passive components. Figure 2 represents the alternative impedance
and admittance forms of representation of a general network component.

Ep p Ep p
TU
ipq

epq
(ipq+ jpq)
jpq
y
JN

vpq = Ep - Eq pq
z
pq

ipq ipq
ll

Eq q Eq q
A

Fig.2 Representation of a primitive network element

(a) Impedance form (b) Admittance form

19
 The network performance can be represented by using either the impedance or the
admittance form of representation. With respect to the element, p-q, let,
vpq = voltage across the element p-q,
epq = source voltage in series with the element p-

d
q, ipq= current through the element p-q,
jpq= source current in shunt with the element p-
q, zpq= self impedance of the element p-q and

l
ypq= self admittance of the element p-q.

or
Performance equation: Each element p-q has two variables, vpq and ipq. The
performance of the given element p-q can be expressed by the performance
equations as under:

W
vpq + epq = zpqipq (in its impedance form)
ipq + jpq = ypqvpq (in its admittance form) (6)

Thus the parallel source current jpq in admittance form can be related to the series
TU
source voltage, epq in impedance form as per the identity:
jpq = - ypq epq (7)

A set of non-connected elements of a given system is defined as a primitive Network

and an element in it is a fundamental element that is not connected to any other element.
JN

In the equations above, if the variables and parameters are replaced by the
corresponding vectors and matrices, referring to the complete set of elements present in
a given system, then, we get the performance equations of the primitive network in
the form as under:
v + e = [z] i
i + j = [y] v (8)
ll

Primitive network matrices:

A

A diagonal element in the matrices, [z] or [y] is the self impedance zpq-pq or self
admittance, ypq-pq. An off-diagonal element is the mutual impedance, zpq-rs or mutual
admittance, ypq-rs, the value present as a mutual coupling between the elements p-q and
r-s. The primitive network admittance matrix, [y] can be obtained also by

20
 inverting the primitive impedance matrix, [z]. Further, if there are no mutually
coupled elements in the given system, then both the matrices, [z] and [y] are
diagonal. In such cases, the self impedances are just equal to the reciprocal of the
corresponding values of self admittances, and vice-versa.

d
Examples on Primitive Networks:

l
or
Example-4: Given that the self impedances of the elements of a network referred by
the bus incidence matrix given below are equal to: Z1=Z2=0.2, Z3=0.25, Z4=Z5=0.1
and Z6=0.4 units, draw the corresponding oriented graph, and find the primitive
network matrices. Neglect mutual values between the elements.

A=
-1
0
0 W 0
-1
0
0
0
-1
TU
1 -1 0
0 1 -1
1 0 -1
JN

Solution:
ˆ
The element node incidence matrix, A can be obtained from the given A matrix, by
pre-augmenting to it an extra column corresponding to the reference node, as under.

1 -1 0 0
1 0 -1 0
ll

ˆ 1 0 0 -1
A =
0 1 -1 0
A

0 0 1 -1
0 1 0 -1

21
 ˆ
Based on the conventional definitions of the elements of A , the oriented graph can
be formed as under:

l d
or
W
Fig. E4 Oriented Graph

Thus the primitive network matrices are square, symmetric and diagonal matrices of
TU
order e=no. of elements = 6. They are obtained as follows.

0.2 0 0 0 0 0
0 0.2 0 0 0 0
[z] = 0 0 0.25 0 0 0
JN

0 0 0 0.1 0 0
0 0 0 0 0.1 0
0 0 0 0 0 0.4

And

5.0 0 0 0 0 0
ll

0 5.0 0 0 0 0
[y] = 0 0 4.0 0 0 0
A

0 0 0 10 0 0
0 0 0 0 10 0
0 0 0 0 0 2.5

22
 Example-5: Consider three passive elements whose data is given in Table E5 below.
Form the primitive network impedance matrix.

d
Table E5

l
Self impedance (zpq-pq) Mutual impedance, (zpq-rs)
Element

or
Bus-code, Impedance in Bus-code, Impedance in
number
(p-q) p.u. (r-s) p.u.

1 1-2 j 0.452

W
2 2-3 j 0.387 1-2 j 0.165

3 1-3 j 0.619 1-2 j 0.234

Solution:
TU
1-2 2-3 1-3

1-2 j 0.452 j 0.165 j 0.234

2-3
JN

[z] = j 0.165 j 0.387 0

1-3 j 0.234 0 j 0.619

Note:
 The size of [z] is e  e, where e= number of elements, 

 The diagonal elements are the self impedances of the elements 
ll


 The off-diagonal elements are mutual impedances between the corresponding
elements. 

A

 Matrices [z] and [y] are inter-invertible. 

23
 FORMATION OF YBUS AND ZBUS

The bus admittance matrix, YBUS plays a very important role in computer aided power
system analysis. It can be formed in practice by either of the methods as under:

d
1. Rule of Inspection
2. Singular Transformation
3. Non-Singular Transformation

l
4. ZBUS Building Algorithms, etc.

or
The performance equations of a given power system can be considered in three
different frames of reference as discussed below:

Frames of Reference:
Bus Frame of Reference: There are b independent equations (b = no. of buses)

W
relating the bus vectors of currents and voltages through the bus impedance matrix
and bus admittance matrix:
EBUS = ZBUS IBUS
IBUS = YBUS EBUS (9)
Branch Frame of Reference: There are b independent equations (b = no. of branches
of a selected Tree sub-graph of the system Graph) relating the branch vectors of
TU
currents and voltages through the branch impedance matrix and branch admittance
matrix:
EBR = ZBR IBR
IBR = YBR EBR (10)
Loop Frame of Reference: There are b independent equations (b = no. of branches of a
selected Tree sub-graph of the system Graph) relating the branch vectors of currents and
JN

voltages through the branch impedance matrix and branch admittance matrix:
ELOOP = ZLOOP ILOOP
ILOOP = YLOOP ELOOP (11)
Of the various network matrices refered above, the bus admittance matrix (Y BUS)
and the bus impedance matrix (ZBUS) are determined for a given power system by
the rule of inspection as explained next.
ll

Rule of Inspection
Consider the 3-node admittance network as shown in figure5. Using the basic branch
relation: I = (YV), for all the elemental currents and applying Kirchhoff’s Current
A

Law principle at the nodal points, we get the relations as under:

At node 1: I1 =Y1V1 + Y3 (V1-V3) + Y6 (V1 – V2)
At node 2: I2 =Y2V2 + Y5 (V2-V3) + Y6 (V2 – V1)
At node 3: 0 = Y3 (V3-V1) + Y4V3 + Y5 (V3 – V2) (12)

24
 l d
or
Fig. 3 Example System for finding YBUS

These are the performance equations of the given network in admittance form and

W
they can be represented in matrix form as:

I1 = (Y1+Y3 +Y6)-Y6 -Y3 V1

I2 = -Y6 (Y2+Y5 +Y6) -Y5 V2
0 = -Y3 -Y5 (Y3 +Y4+Y5) V3 (13)
TU
In other words, the relation of equation (9) can be represented in the form
IBUS = YBUS EBUS (14)
Where, YBUS is the bus admittance matrix, IBUS & EBUS are the bus current and bus
voltage vectors respectively.
By observing the elements of the bus admittance matrix, Y BUS of equation (13), it is
observed that the matrix elements can as well be obtained by a simple inspection of
JN

the given system diagram:

Diagonal elements: A diagonal element (Yii) of the bus admittance matrix,
YBUS, is equal to the sum total of the admittance values of all the elements
incident at the bus/node i,
Off Diagonal elements: An off-diagonal element (Yij) of the bus admittance
matrix, YBUS, is equal to the negative of the admittance value of the
connecting element present between the buses I and j, if any.
ll

This is the principle of the rule of inspection. Thus the algorithmic equations for the
rule of inspection are obtained as:
A

Yii =  yij (j = 1,2,…….n)

Y =-y
ij ij (j = 1,2,…….n) (15)

25
 For i = 1,2,….n, n = no. of buses of the given system, yij is the admittance of
element connected between buses i and j and yii is the admittance of element
connected between bus i and ground (reference bus).

Bus impedance matrix

In cases where, the bus impedance matrix is also required, it cannot be formed by

d
direct inspection of the given system diagram. However, the bus admittance matrix
determined by the rule of inspection following the steps explained above, can be
inverted to obtain the bus impedance matrix, since the two matrices are inter-
invertible.

l
or
Note: It is to be noted that the rule of inspection can be applied only to those power
systems that do not have any mutually coupled elements.

Examples on Rule of Inspection:

W
Example 6: Obtain the bus admittance matrix for the admittance network shown
aside by the rule of inspection

16 -8 -4 YBUS = j
-8 24 -8 -4 -8 16
TU

Example 7: Obtain YBUS for the impedance network shown aside by the rule of
JN

inspection. Also, determine YBUS for the reduced network after eliminating the
eligible unwanted node. Draw the resulting reduced system diagram.

-9.8 5 4
YBUS= j 5 -16 10
ll

4 10 -14
A

-1
ZBUS = YBUS

26
 l d
or
YBUS
New

W -1
= YA-YBYD YC
TU
YBUS = j -8.66 7.86
7.86 -8.66

SINGULAR TRANSFORMATIONS
The primitive network matrices are the most basic matrices and depend purely on
JN

the impedance or admittance of the individual elements. However, they do not

contain any information about the behaviour of the interconnected network
variables. Hence, it is necessary to transform the primitive matrices into more
meaningful matrices which can relate variables of the interconnected network.

Bus admittance matrix, YBUS and Bus impedance matrix, ZBUS

ll

In the bus frame of reference, the performance of the interconnected network is

described by n independent nodal equations, where n is the total number of buses (n+1
A

nodes are present, out of which one of them is designated as the reference node). For
example a 5-bus system will have 5 external buses and 1 ground/ ref. bus). The

27
 performance equation relating the bus voltages to bus current injections in bus frame
of reference in admittance form is given by
IBUS = YBUS EBUS (17)
Where EBUS = vector of bus voltages measured with respect to reference bus

d
IBUS = Vector of currents injected into the bus
YBUS = bus admittance matrix
The performance equation of the primitive network in admittance form is given by

l
i + j = [y] v

or
t
Pre-multiplying by A (transpose of A), we obtain
t t t
A i +A j = A [y] v (18)
However, as per equation (4),
t

W
A i =0,
since it indicates a vector whose elements are the algebraic sum of element currents
t
incident at a bus, which by Kirchhoff’s law is zero. Similarly, A j gives the
algebraic sum of all source currents incident at each bus and this is nothing but the
total current injected at the bus. Hence,
TU
t
A j = IBUS (19)
t
Thus from (18) we have, IBUS = A [y] v (20)
However, from (5), we have
v =A EBUS
And hence substituting in (20) we get,
JN

t
IBUS = A [y] A EBUS (21)
Comparing (21) with (17) we obtain,
t
YBUS = A [y] A (22)
The bus incidence matrix is rectangular and hence singular. Hence, (22) gives a
singular transformation of the primitive admittance matrix [y]. The bus impedance
matrix is given by ,
ll

-1

ZBUS = YBUS (23)

Note: This transformation can be derived using the concept of power invariance,
A

however, since the transformations are based purely on KCL and KVL, the
transformation will obviously be power invariant.

28
 Examples on Singular Transformation:

Example 8: For the network of Fig E8, form the primitive matrices [z] & [y] and
obtain the bus admittance matrix by singular transformation. Choose a Tree

d
T(1,2,3). The data is given in Table E8.

l
or
W
TU
Fig E8 System for Example-8

Table E8: Data for Example-8

JN

Elements Self impedance Mutual impedance

1 j 0.6 -
2 j 0.5 j 0.1(with element 1)
3 j 0.5 -
4 j 0.4 j 0.2 (with element 1)
5 j 0.2 -
ll
A

Solution:

The bus incidence matrix is formed taking node 1 as the reference bus.

29
  1 0 0
 
 0 1 0 
A=  0 1  1

d
 
 1 0 0 
 1 0 
1

l


or
The primitive incidence matrix is given by,

 j0.6 j0.1 0.0 j0.2 0.0 

 
 j0.1 j0.5 0.0 0.0 0.0 

W
[z]=  0.0 0.0 j0.5 0.0 0.0 
 
 j0.2 0.0 0.0 j0.4 0.0 
 
 0.0 0.0 0.0 0.0 j0.2 

-1
The primitive admittance matrix [y] = [z] and given by,
TU
 j2.0833 j0.4167 0.0 j1.0417 0.0 
 
 j0.4167  j2.0833 0.0  j0.2083 0.0 
[y]=  0.0 0.0  j2.0 0.0 0.0 
 
 j1.0417  j0.2083 0.0  j3.0208 0.0 
 
JN

 0.0 0.0 0.0 0.0  j5.0 

The bus admittance matrix by singular transformation is obtained as

 j8.0208 j0.2083 j5.0 
t
YBUS = A [y] A =  j0.2083  j4.0833 j2.0 
 

 j5.0 j2.0  j7.0

 
ll

 j0.2713 j0.1264 j0.2299

-1 
=  j0.1264
A

ZBUS = YBUS j0.3437 j0.1885

 

 j0.2299 j0.1885 j0.3609

 

30
 SUMMARY
The formulation of the mathematical model is the first step in obtaining the solution
of any electrical network. The independent variables can be either currents or
voltages. Correspondingly, the elements of the coefficient matrix will be impedances

d
or admittances.

Network equations can be formulated for solution of the network using graph

l
theory, independent of the nature of elements. In the graph of a network, the tree-

or
branches and links are distinctly identified. The complete information about the
interconnection of the network, with the directions of the currents is contained in the
bus incidence matrix.

W
The information on the nature of the elements which form the interconnected
network is contained in the primitive impedance matrix. A primitive element can be
represented in impedance form or admittance form. In the bus frame of reference,
the performance of the interconnected system is described by (n-1) nodal equations,
TU
where n is the number of nodes. The bus admittance matrix and the bus impedance
matrix relate the bus voltages and currents. These matrices can be obtained from the
primitive impedance and admittance matrices.
JN
ll
A

31
 Chapter-1-B

FORMATION OF BUS IMPEDANCE MATRIX

[CONTENTS: Node elimination by matrix algebra, generalized algorithms for ZBUS

d
building, addition of BRANCH, addition of LINK, special cases of analysis,
removal of elements, changing the impedance value of an element, examples]

l
NODE ELIMINATION BY MATRIX ALGEBRA

or
Nodes can be eliminated by the matrix manipulation of the standard node equations.
However, only those nodes at which current does not enter or leave the network can be
considered for such elimination. Such nodes can be eliminated either in one group or by

W
taking the eligible nodes one after the other for elimination, as discussed next.

CASE-A: Simultaneous Elimination of Nodes:

Consider the performance equation of the given network in bus frame of reference
in admittance form for a n-bus system, given by:
TU
I =Y E
BUS BUS BUS (1)
Where IBUS and EBUS are n-vectors of injected bus current and bus voltages and
YBUS is the square, symmetric, coefficient bus admittance matrix of order n.
JN

Now, of the n buses present in the system, let p buses be considered for node-
elimination so that the reduced system after elimination of p nodes would be retained
with m (= n-p) nodes only. Hence the corresponding performance equation would be
similar to (1) except that the coefficient matrix would be of order m now, i.e.,

new
I =Y E
ll

BUS BUS BUS (2)

new

Where YBUS is the bus admittance matrix of the reduced network and the vectors
A

IBUS and EBUS are of order m. It is assumed in (1) that IBUS and EBUS are obtained
with their elements arranged such that the elements associated with p nodes to be
eliminated are in the lower portion of the vectors. Then the elements of Y BUS also
get located accordingly so that (1) after matrix partitioning yields,

32
 m p
I m Y Y E
BUS-m A B BUS-m
=
I p Y Y E
BUS-p C D BUS-p

d
(3)
Where the self and mutual values of YA and YD are those identified only with the
t

l
nodes to be retained and removed respectively and YC=YB is composed of only the
corresponding mutual admittance values, that are common to the nodes m and p.

or
Now, for the p nodes to be eliminated, it is necessary that, each element of the
vector IBUS-p should be zero. Thus we have from (3):

W
I =Y E +Y E
BUS-m A BUS-m B BUS-p
I =Y E +Y E =0
BUS-p C BUS-m D BUS-p (4)
-1
Solving, EBUS-p = - YD YC EBUS-m (5)
TU
Thus, by simplification, we obtain an expression similar to (2) as,
-1
IBUS-m = {YA - YBYD YC} EBUS-m (6)

Thus by comparing (2) and (6), we get an expression for the new bus admittance
JN

matrix in terms of the sub-matrices of the original bus admittance matrix as:
new -1
Y
BUS = {YA - YBYD YC} (7)
This expression enables us to construct the given network with only the necessary
nodes retained and all the unwanted nodes/buses eliminated. However, it can be
observed from (7) that the expression involves finding the inverse of the sub-matrix
ll

YD (of order p). This would be computationally very tedious if p, the nodes to be
eliminated is very large, especially for real practical systems. In such cases, it is
A

more advantageous to eliminate the unwanted nodes from the given network by
considering one node only at a time for elimination, as discussed next.

33
 CASE-B: Separate Elimination of Nodes:
Here again, the system buses are to be renumbered, if necessary, such that the node
to be removed always happens to be the last numbered one. The sub-matrix YD then
would be a single element matrix and hence it inverse would be just equal to its own

d
reciprocal value. Thus the generalized algorithmic equation for finding the elements
of the new bus admittance matrix can be obtained from (6) as,

l
new old
Yij = Yij – Yin Ynj / Ynni,j = 1,2,…… n. (8)

or
Each element of the original matrix must therefore be modified as per (7). Further,
this procedure of eliminating the last numbered node from the given system of n
nodes is to be iteratively repeated p times, so as to eliminate all the unnecessary p
nodes from the original system.

Examples on Node elimination:

W
Example-1: Obtain YBUS for the impedance network shown below by the rule of
inspection. Also, determine YBUS for the reduced network after eliminating the
TU
eligible unwanted node. Draw the resulting reduced system diagram.
JN

The admittance equivalent network is as follows:

ll
A

34
 The bus admittance matrix is obtained by RoI as:
-9.8 5 4
YBUS= j 5 -16 10
4 10 -14

d
The reduced matrix after elimination of node 3 from the given system is
determined as per the equation:

l
New -1

YBUS = YA-YBYD YC

or
n/n 1 2
new 1 -j8.66 j7.86
Y
BUS =
2 j7.86 -j8.66

W
Alternatively,
new old
Yij = Yij – Yi3 Y3j / Y33 i,j = 1,2.

Y11 = Y11-Y13Y31/ Y33 = -j8.66

Y22 = Y22 – Y23Y32/ Y33 = -j8.66
TU
Y12 = Y21 = Y12 – Y13Y32/Y33 = j7.86

Thus the reduced network can be obtained again by the rule of inspection as shown
be low.
JN
ll
A

35
 Example-2: Obtain YBUS for the admittance network shown below by the rule of
inspection. Also, determine YBUS for the reduced network after eliminating the
eligible unwanted node. Draw the resulting reduced system diagram.

l d
or
n/n 1
1 -j50
W
2
0
3
j20
4
j10
TU
YBUS=2 0 -j60 0 j72 = YA YB
3 j20 0 -j72 j50 YC YD
4 j10 j72 j50 -j81
JN

New -1
Y
BUS = YA-YBYD YC
n/n 1 2
new 1 -j32.12 j10.32
Y
BUS =
ll

2 j10.32 -j51.36
A

Thus the reduced system of two nodes can be drawn by the rule of inspection as
under:

36
 l d
or
ZBUS building

FORMATION OF BUS IMPEDANCE MATRIX

W
The bus impedance matrix is the inverse of the bus admittance matrix. An alternative
method is possible, based on an algorithm to form the bus impedance matrix directly
from system parameters and the coded bus numbers. The bus impedance matrix is
formed adding one element at a time to a partial network of the given system. The
TU
performance equation of the network in bus frame of reference in impedance form using
the currents as independent variables is given in matrix form by

 Z I
E bus bus
bus (9)
When expanded so as to refer to a n bus system, (9) will be of the form
JN

E1  Z11 I1  Z12 I 2  .......  .Z1k I k ...  Z1n I n

Z I Z I
Ek k1 1 k2 2  ......  Z kk I k  ....  Z kn I n

E
ll

n  Z n1 I1  Z n 2 I 2  .........  Z nk I k  .....  Z nn I n (10)

Now assume that the bus impedance matrix Zbus is known for a partial network of m
A

buses and a known reference bus. Thus, Zbus of the partial network is of dimension
m  m. If now a new element is added between buses p and q we have the following
two possibilities:

37
 (i) p is an existing bus in the partial network and q is a new bus; in this case
p-q is a branch added to the p-network as shown in Fig 1a, and
(ii) both p and q are buses existing in the partial network; in this case p-q is a
link added to the p-network as shown in Fig 1b.

d
1
2

l
Partial

or
Network
p
q

W
BUS i
m
0 Ref.
TU
Fig 1a. Addition of branch p-q

1
2
Partial
JN

Network
p

Z
BUS q
ll

m
0 Ref.
A

Fig 1b. Addition of link p-q

38
 If the added element ia a branch, p-q, then the new bus impedance matrix would be
of order m+1, and the analysis is confined to finding only the elements of the new
row and column (corresponding to bus-q) introduced into the original matrix.

If the added element ia a link, p-q, then the new bus impedance matrix will remain
unaltered with regard to its order. However, all the elements of the original matrix

d
are updated to take account of the effect of the link added.

ADDITION OF A BRANCH

l
Consider now the performance equation of the network in impedance form with the

or
added branch p-q, given by
Z Z Z Z  I 
 E1   Z11 12 1p 1m
E  Z 1q
 I 1 
 2   21 Z Z Z Z  2 
   22 2p 2m 2q
 

W
  Z  I 
E p   p1 Z Z Z Z  p  (11)
  pq
 
 p2 pp pm

   Z Z Z  
Z Z  I m 
 Em   m1 m2 mp mm
Z  
mq
E  Z
 q   q1 Z I
q2 Z Z qm qq
  qp
 q 
TU
It is assumed that the added branch p-q is mutually coupled with some elements of the
partial network and since the network has bilateral passive elements only, we have
Vector ypq-rs is not equal to zero and Zij= Zji i,j=1,2,…m,q (12)

To find Zqi:
JN

The elements of last row-q and last column-q are determined by injecting a current of
1.0 pu at the bus-i and measuring the voltage of the bus-q with respect to the reference
bus-0, as shown in Fig.2. Since all other bus currents are zero, we have from (11) that
Ek = Zki Ii = Zki  k = 1, 2,…i.…...p,….m, q (13)
Hence, Eq = Zqi ; Ep = Zpi ………
Also, Eq=Ep -vpq ; so that Zqi = Zpi - vpq  i =1, 2,…i.…...p,….m, ≠q (14)
ll

To find vpq:
In terms of the primitive admittances and voltages across the elements, the current
A

through the elements is given by

39
 y v
ipq  pq,pq y pq,rs  pq 

     (15)
i y y v
 rs   rs,pq rs,rs  rs 

d
1
2

l
Partial

or
Network v
pq
p
q

Z i

W
BUS Ii =1pu
m
0 Ref.
TU
Fig.2 Calculation for Zqi

where i pq is current through element p-q

irs is vector of currents through elements of the partial network

v pq is voltage across element p-q

JN

y pq , pq is self – admittance of the added element

y pq,rs is the vector of mutual admittances between the added elements p-q and
elements r-s of the partial network.
vrs is vector of voltage across elements of partial network.

yrs, pq is transpose of y pq,rs .

ll

yrs ,rs is the primitive admittance of partial network.

A

Since the current in the added branch p-q, is zero, i pq  0 . We thus have from (15),

i y v 
pq pq , pq pq y pq ,rs v rs  0 (16)

40
 y pq,rs vrs
Solving, v pq  y or
pq, pq
  
v y pq,rs Er Es
pq  y (17)

d
pq, pq

Using (13) and (17) in (14), we get


Z Z y pq,rs Z ri  Z si 

l
qi pi  y i  1,2......m; i  q (18)

or
pq, pq

To find zqq:
The element Zqq can be computed by injecting a current of 1pu at bus-q, Iq = 1.0 pu.
As before, we have the relations as under:
Ek = Zkq Iq = Zkq  k = 1, 2,…i.…...p,….m, q (19)

W
Hence, Eq = Zqq ; Ep = Zpq ; Also, Eq =Ep - vpq ; so that Zqq = Zpq - vpq (20)
Since now the current in the added element is ipq Iq 1.0 , we have from (15)

i y v
pq pq , pq pq  y pq ,rsv rs 1
y pq,rs v rs
TU
Solving, v pq 1  y
pq, pq
 
v y pq,rs E r  E s
pq 1  y (21)
pq, pq

Using (19) and (21) in (20), we get

 

JN

Z Z 1y pq,rs
Z rq Z sq
qq pq  y (22)
pq, pq

Special Cases
The following special cases of analysis concerning ZBUS building can be considered
with respect to the addition of branch to a p-network.

Case (a): If there is no mutual coupling then elements of y pq ,rs are zero. Further, if
ll

p is the reference node, then Ep=0. thus,

Zpi = 0 i  1,2......m : i  q
A

And Zpq = 0.
Hence, from (18) (22) Zqi = 0 i  1,2.......m; i  q
Z z
And qq pq, pq \ (23)

41
 Case (b): If there is no mutual coupling and if p is not the ref. bus, then, from (18)

and (22), we again have,

Z
qi  Z pi , i  1,2....m; i  q

d
Z z
qq  Z pq pq , pq (24)

l
ADDITION OF A LINK

or
Consider now the performance equation of the network in impedance form with the
added link p-l, (p-l being a fictitious branch and l being a fictitious node) given by

Z Z  I 

W
E1   Z11 12 Z1 p Z1m
E  Z 1q
 1 
 2   21 Z Z  I2 
   22 Z2 p Z2m 2q
 
E  Z  I 
 p   p1 Z Z Z Z  p  (25)
 
pq
 p2 pp pm
 
 Z Z Z

Em 


Z Z  I
m

m1 m2 mp mm
TU
mq
  
Z Z  
E l2
ZZli Zlm ll  Il 
 l   l1
It is assumed that the added branch p-q is mutually coupled with some elements of the
partial network and since the network has bilateral passive elements only, we have

Vector ypq-rs is not equal to zero and Zij= Zji i,j=1,2,…m,l. (26)
JN

To find Zli:
The elements of last row-l and last column-l are determined by injecting a current of
1.0 pu at the bus-i and measuring the voltage of the bus-q with respect to the
reference bus-0, as shown in Fig.3. Further, the current in the added element is made
zero by connecting a voltage source, el in series with element p-q, as shown. Since
ll

all other bus currents are zero, we have from (25) that
Ek = Zki Ii = Zki  k = 1, 2,…i.…...p,….m, l (27)
A

Hence, el = El = Zli ; Ep = Zpi ; Ep = Zpi ………

Also, el = Ep - Eq - vpq ;
So that Zli = Zpi - Zqi - vpq  i=1,2,…i.…p,...q,….m, ≠l (28)

42
 To find vpq:
In terms of the primitive admittances and voltages across the elements, the current
through the elements is given by

ipl ypl, pl ypl,rsvpl
 

d
(29)
i y y v
 rs  rs,pl rs,rs  rs 

l
or
1
2
Partial
Network p
v

W
pl

q el
i l
Z
BUS
Ii =1pu
m
TU
0 Ref.

Fig.3 Calculation for Zli

JN

where ipl is current through element p-q

i rs is vector of currents through elements of the partial network

v
is voltage across element p-q
pl
y
pl , pl is self – admittance of the added element

y pl , rs is the vector of mutual admittances between the added elements p-q and
ll

elements r-s of the partial network.

vrs is vector of voltage across elements of partial network.
A

yrs, pl is transpose of y pl , rs .

yrs ,rs is the primitive admittance of partial network.

43
 Since the current in the added branch p-l, is zero, ipl  0 . We thus have from (29),

i y v  0
pl pl , pl pl y pl , rs v rs (30)
y pl , rs v rs
Solving, vpl  y or

d
pl , pl

v y pl , rs E r  Es 
pl  y (31)

l
pl , pl

However,

or
y pl , rs  y pq, rs
y y
And pl , pl pq, pq (32)
Using (27), (31) and (32) in (28), we get


W

y pq, rs
Z ri Z si 
Z Z
li pi  Zqi  y i  1,2......m;i  l (33)
pq, pq

To find Zll:
The element Zll can be computed by injecting a current of 1pu at bus-l, Il = 1.0 pu.
TU
As before, we have the relations as under:
Ek = Zkl Il = Zkl  k = 1, 2,…i.…...p,…q,….m, l (34)
Hence, el = El = Zll ; Ep = Zpl ;
Also, el = Ep - Eq - vpl ;
So that Zll = Zpl - Zql - vpl  i=1,2,…i.…p,...q,….m, ≠l (35)
JN

Since now the current in the added element is ipl Il 1.0 , we have from (29)

i y v 
pl pl , pl pl y pl , rs v rs 1
v y pl , rs vrs
y
Solving, pl 1  pl , pl

v y pl , rs Er  Es 
ll

pl 1  y (36)
pl , pl

However,
A

y pl , rs  y pq, rs
y y
And pl , pl pq , pq (37)
Using (34), (36) and (37) in (35), we get

44
 
1  y pq, rs Z rl  Z sl 
Z
ll  Z pl  Zql  y (38)
pq, pq

Special Cases Contd….

The following special cases of analysis concerning ZBUS building can be considered

d
with respect to the addition of link to a p-network.

Case (c): If there is no mutual coupling, then elements of y pq,rs are zero. Further, if p

l
is the reference node, then Ep=0. thus,

or
Z
li Zqi , i  1,2....m;i  l
Z Z  z
ll ql pq , pq (39)
From (39), it is thus observed that, when a link is added to a ref. bus, then the
situation is similar to adding a branch to a fictitious bus and hence the following

W
steps are followed:
1. The element is added similar to addition of a branch (case-b) to obtain the
new matrix of order m+1.
2. The extra fictitious node, l is eliminated using the node elimination algorithm.

Case (d): If there is no mutual coupling, then elements of y pq,rs are zero. Further, if p
is not the reference node, then
TU
Zli = Zpi - Zqi

Zll = Zpl – Zql – zpq,pq

= Zpp + Zqq – 2 Zpq+ zpq,pq (40)

MODIFICATION OF ZBUS FOR NETWORK CHANGES

JN

An element which is not coupled to any other element can be removed easily. The
Zbus is modified as explained in sections above, by adding in parallel with the
element (to be removed), a link whose impedance is equal to the negative of the
impedance of the element to be removed. Similarly, the impedance value of an
element which is not coupled to any other element can be changed easily. The Zbus
ll

is modified again as explained in sections above, by adding in parallel with the

element (whose impedance is to be changed), a link element of impedance value
chosen such that the parallel equivalent impedance is equal to the desired value of
A

impedance. When mutually coupled elements are removed, the Zbus is modified by
introducing appropriate changes in the bus currents of the original network to reflect
the changes introduced due to the removal of the elements.

45
 Examples on ZBUS building

Example 1: For the positive sequence network data shown in table below, obtain
ZBUS by building procedure.
Pos. seq.
p-q
Sl. No. reactance
(nodes)

d
in pu
1 0-1 0.25
2 0-3 0.20

l
3 1-2 0.08
4 2-3 0.06

or
Solution:
The given network is as shown below with the data marked on it. Assume the
elements to be added as per the given sequence: 0-1, 0-3, 1-2, and 2-3.

W
TU

Fig. E1: Example System

Consider building ZBUS as per the various stages of building through the
JN

consideration of the corresponding partial networks as under:

Step-1: Add element–1 of impedance 0.25 pu from the external node-1 (q=1) to
internal ref. node-0 (p=0). (Case-a), as shown in the partial network;
ll
A

1
(1)
ZBUS = 1 0.25

46
 Step-2: Add element–2 of impedance 0.2 pu from the external node-3 (q=3) to
internal ref. node-0 (p=0). (Case-a), as shown in the partial network;

ld
or
1 3
(2) 1 0.25 0

W
ZBUS = 3 0 0.2

Step-3: Add element–3 of impedance 0.08 pu from the external node-2 (q=2) to
internal node-1 (p=1). (Case-b), as shown in the partial network;
TU
JN

1 3 2

(3)
1 0.25 0 0.25
ZBUS = 3 0 0.2 0
2 0.25 0 0.33
Step-4: Add element–4 of impedance 0.06 pu between the two internal nodes,
node-2 (p=2) to node-3 (q=3). (Case-d), as shown in the partial network;
ll
A

47
 l d
or
1 3 2 l
1 0.25 0 0.25 0.25
(4) 3 0 0.2 0 -0.2

W
ZBUS = 2 0.25 0 0.33 0.33
l 0.25 -0.2 0.33 0.59

The fictitious node l is eliminated further to arrive at the final impedance matrix as
under:

1 3 2
TU
1 0.1441 0.0847 0.1100
(final)
ZBUS = 3 0.0847 0.1322 0.1120
2 0.1100 0.1120 0.1454

Example 2: The ZBUS for a 6-node network with bus-6 as ref. is as given below.
Assuming the values as pu reactances, find the topology of the network and the
parameter values of the elements involved. Assume that there is no mutual coupling
of any pair of elements.
JN

1 2 3 4 5
1 2 0 0 0 2
2 0 2 0 2 0
3
ZBUS = 0 0 2 0 0
4 0 2 0 3 0
5 2 0 0 0 3
ll

Solution:
The specified matrix is so structured that by its inspection, we can obtain the
network by backward analysis through the various stages of ZBUS building and p-
networks as under:
A

48
 A
ll
JN
TU
W

49
or
ld
 l d
or
W
Thus the final network is with 6 nodes and 5 elements connected as follows with
the impedance values of elements as indicated.
TU
JN

Fig. E2: Resultant network of example-2

ll
A

50
 Example 3: Construct the bus impedance matrix for the system shown in the figure
below by building procedure. Show the partial networks at each stage of building
the matrix. Hence arrive at the bus admittance matrix of the system. How can this
result be verified in practice?

l d
or
Solution: The specified system is considered with the reference node denoted by

W
node-0. By its inspection, we can obtain the bus impedance matrix by building
procedure by following the steps through the p-networks as under:

Step1: Add branch 1 between node 1 and reference node. (q =1, p = 0)

TU
JN

Step2: Add branch 2, between node 2 and reference node. (q = 2, p = 0).

ll
A

51
 d
Step3: Add branch 3, between node 1 and node 3 (p = 1, q = 3)

l
or
W
TU

Step 4: Add element 4, which is a link between node 1 and node 2. (p = 1, q = 2)

JN
ll
A

52
 l d
Now the extra node-l has to be eliminated to obtain the new matrix of step-4,
using the algorithmic relation:

or
new old
Yij = Yij – Yin Ynj / Ynn i,j = 1,2, 3.

W
TU
Step 5: Add link between node 2 and node 3 (p = 2, q=3)
JN
ll
A

53
 l d
or
Thus, the new matrix is as under:

W
TU
Node l is eliminated as shown in the previous step:
JN

Further, the bus admittance matrix can be obtained by inverting the bus
impedance matrix as under:
ll
A

As a check, it can be observed that the bus admittance matrix, YBUS can also be
obtained by the rule of inspection to arrive at the same answer.

54
 Example 4: Form the bus impedance matrix for the network shown below.

l d
or
Solution:
Add the elements in the sequence, 0-1, 1-2, 2-3, 0-3, 3-4, 2-4, as per the various

W
steps of building the matrix as under:

Step1: Add element 1, which is a branch between node-1 and reference node.
TU
Step2: Add element 2, which is a branch between nodes 1 and 2.
JN

Step3: Add element 3, which is a branch between nodes 2 and 3

ll
A

Step4: Add element 4, which is a link from node 3 to reference node.

55
 d
Eliminating node l,

l
or
W
Step5: Add element 5, a branch between nodes 3 and 4.
TU
JN

Step 6: Add element 6, a link between nodes 2 & 4.

ll

Eliminating node l we get the required bus impedance , matrix

A

56
 l d
Example 5: Form the bus impedance matrix for the network data given below.

or
Self Impedance Mutual Impedance
Element Bus zpq, pq Bus zpq, rs
p-q (pu) r-s (pu)
1 1 – 2(1) j0.6
2 1 – 2(2) j0.4 1 – 2(1) j0.2

W
Solution:
Let bus-1 be the reference. Add the elements in the sequence 1-2(1), 1-2(2). Here,
in the step-2, there is mutual coupling between the pair of elements involved.
TU

Step1: Add element 1 from bus 1 to 2, element 1-2(1). ( p=1, q=2, p is the
reference node)
JN

Step2: Add element 2, element 1-2(2), which is a link from bus1 to 2,

mutually coupled with element 1, 1-2(1).
ll
A

57
 Consider the primitive impedance matrix for the two elements given by

d
Thus the primitive admittance matrix is obtained by taking the inverse of [z] as

l
or
Thus,

So that we have,

W
TU
JN

Thus, the network matrix corresponding to the 2-node, 1-bus network given, is
obtained after eliminating the extra node-l as a single element matrix, as under:
ll
A

-------------------

58