e-Notes of Lecture Sessions of

e-Classes conducted by
Dr. M.S. Raviprakasha
Professor of E&E Engg.,
Malnad College of Engineering, Hassan
Subject: COMPUTER TECHNIQUES IN POWER SYSTEMS
Code: EE72 No. of Hrs.: 33
SUBJECT EXPERTS:
Dr. MS Raviprakasha, Professor of E&EE, MCE, Hassan
Dr. K Umarao, Prof. & HOD of E&EE, RNSIT, Bangalore
LECTURE SCHEDULE:
Dr. M.S. Raviprakasha: Chapters # 1, 3 (Introduction, Linear
graph theory, Z
BUS
Building, Load flow analysis (15 Hrs.): (Plus
2 concluding Sessions)
Dr. K Umarao: Chapter # 2,4,5 (Load frequency control,
economic operation of power systems, transient stability
studies) (15 Hrs.) : (Plus 1 concluding Session)

PROGRAMME SCHEDULE:
17 Hrs. of Classes by Dr. MS Raviprakasha, MCE, Hassan
16 Hrs. of Classes by Dr. K Umarao, RNSIT, Bangalore
3 Classes per week: Wednesday, Thursday and Friday (All
classes at 11am -12 noon)
3 concluding sessions to discuss on solutions to the recent
question papers of VTU, VIIEEE class on the selected subject
of CTPS (EE72).
2
CONTENTS (CHAPTER-WISE)
Chapter 1 LINEAR GRAPH THEORY
PART A: INCIDENCE AND NETWORK MATRICES
Definitions of important terms
Incidence matrices: Element node incidence matrix and Bus
incidence matrix
Primitive networks and primitive network matrices
Performance of primitive networks
Frames of reference
Singular transformation analysis
Formation of bus admittance matrix by rule of inspection and
singular transformation
Examples
PART B: Z
BUS
BUILDING ALGORITHMS
Partial networks
Addition of Branch and addition of Link
Algorithms for formation of Z
BUS
for single phase systems
Special cases without and with mutually coupled elements
Deletion of an element, changing the impedance value of an
element, etc.
Examples
Chapter 2
LOAD FREQUENCY CONTROL
Turbine speed governing system
Modeling and block diagram representation of single area
Steady state and dynamic response
Area control error
Two area load frequency control
Examples
3
Chapter 3
LOAD FLOW STUDIES
PART A: REVIEW OF SOLUTION OF EQUATIONS
Solution of linear, nonlinear and differential equations
Importance of iterative methods
Flow diagrams for the iterative methods
Comparison of methods
Examples
PART B: BASIC LOAD FLOW ANALYSIS
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
Importance of acceleration factor
Newton Raphson method in polar coordinates
Concept of Jacobian matrix
Comparison of load flow methods
Examples
PART C: ADVANCED LOAD FLOW ANALYSIS
Decoupled load flow
Fast decoupled load flow analysis
Representation of TCUL transformers
Examples
Chapter 4
ECONOMIC OPERATION OF POWER SYSTEMS
Optimal distribution of loads within a plant
Langrangian multipliers
Transmission loss as a function of plant generation
4
Determination of loss coefficients
Automatic economic load dispatch
Examples
Chapter 5
TRANSIENT STABILITY STUDIES
Basic Stability terms
Swing Equation & Swing Curve
Numerical solution of differential equations
Modified Eulers method
Runge-Kutta IV order method
Milnes predictor corrector method
Representation of synchronous machine, loads, etc. for TS studies
Network performance equations
Solution techniques and flow charts
Examples
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.
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.
PREREQUISITE SUBJECTS:
1. Power System Analysis
2. Synchronous Machines
3. Numerical analysis
5
ELECTRICAL POWER SYSTEMS-
THE STATE-OF-THE-ART:
To Begin With
No.
TOPICS
SUB-TOPICS
1.
Representation of
Power systems
- p.u. React./ Imp. Diagram
2.
Electric Power
System
- Generation - Machines
- Transmission - Distribution - Utilization
Tariffs
3. Fault studies
- Sym. Faults
- Sym. Components
- Seq. Imps. / Networks
- Unsymmetrical Faults
4. System Stability
- SSS, TS, DS
- Angle Stability
- Solution of Equations
- EAC, Clarkes Diagram
5.
Linear Graph Theory
(Linear Equations)
- Incidence Matrices
- Singular/ NS Transformations
- Network Matrices
- Z
BUS
Building
6.
Power Flow Studies
(NL Equations)
- Buses, Y
BUS
Advs., Loads flow equations
- Iterative Methods
- GS, NR, FDLF & DCLF
6
Present Scenario
No.
TOPICS
SUB-TOPICS
7.
Reactive Power
Management
- Importance of VArs
- Compensation Devices,
Sizing, Placement,
Design, Optimality,
- VAr Dispatch
- VAr Co-ordination
8.
Gen. Expansion
Planning
- Optimality
- Load Prediction: Short,
Medium and Long Term
Forecasting
9.
Operation and
Control
-EMS: EMC, SLDC,RLDC
-ALFC, Voltage Control
-Tie-line Power Control
10.
System
Reliability
- Requirements
- Methods
11. Economic Operation
- Unit Commitment
- Parallel Operation
- Optimal Load Dispatch
- Constraints
12. Instrumentation
- CTs, PTs
13. State Estimation
- SCADA
- Bad Data Elimination
- Security/ Cont. Studies
.. and so on.
7
Future Trends
No.
TOPICS
SUB-TOPICS
14. Voltage Stability
- Importance
- Angle/ Voltage stability
15.
Power System
Simulators
- Requirements
- Control Blocks
- Data-Base Definition
16. Energy Auditing
- Deregulation
17.
Demand Side
Management
- Time of Use Pricing
18. Renewable Energy
- The Paradigm
19.
Sparsity Oriented
Programming
- Sparsity: Y
BUS
- Ordering Schemes
- LU- Factorization: Fills
- Pivoting
- UD Table Storage
20.
Recent Computer
Applications
- AI
- Expert Systems
- ANN,
- Genetic Algorithms
- Fuzzy Logic, Etc.
__________________-
8
CHAPTER- 1-A
INCIDENCE AND NETWORK MATRICES
[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]
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
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.
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
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
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
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
9
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.
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
need not mark the voltage polarities in the oriented graph.
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:
e = number of elements = 6
n = number of nodes = 4
b = number of branches = n-1 = 3
l = number of links = e-b = 3
Tree = T(1,2,3) and
Co-tree = T(4,5,6)
Sub-graph : sG is a sub-graph of G if the following conditions are satisfied:
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
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
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 (Kirchhoffs Voltage Law) for the loop is
stated as follows: In any lumped network, the algebraic sum of the branch voltages
around any of the loops is zero.
10
Fig 1a. Single line diagram of a power system
Fig 1b. Reactance diagram
Fig 1c. Oriented Graph
11
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
two separate unconnected sub-graphs.
The removal of all but one of the branches of the set, leaves the remaining
graph connected.
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 (Kirchhoffs Current
Law) for the cutset is stated as follows: In any lumped network, the algebraic sum of
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 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.
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,
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
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)
12
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 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 particular
tree. Since each branch is associated with a basic cut-set, the number of basic cut-sets
is equal to the number of branches.
Examples on Basics of LG Theory:
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.
Fig. E1a. Single line diagramof Example System
Fig. E1b. Oriented Graph of Fig. E1a.
13
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.
Fig. E1c. Basic Cutsets of Fig. E1a.
Fig. E1d. Basic Loops of Fig. E1a.
14
INCIDENCE MATRICES
Elementnode incidence matrix: A

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

incidence matrix, A

. An element a
ij
of A

is defined as under:
a
ij
= 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.
= 0 if the branch-i is not at all incident on the node-j.
Thus the dimension of A

is en, where e is the number of elements and n is the

number of nodes in the network. For example, consider again the sample system with
its oriented graph as in fig. 1c. the corresponding element-node incidence matrix, is
obtained as under:
Nodes
0 1 2 3
Elements
1 1 -1
2 1 -1
A

=
3 1 -1
4 1 -1
5 1 -1
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
matrix is less than n. Thus in general, the matrix A

satisfies the identity:

n
a
ij
= 0 i = 1,2,..e. (3)
j=1
15
Bus incidence matrix: A
By selecting any one of the nodes of the connected graph as the reference node, the
corresponding column is deleted from A

to obtain the bus incidence matrix, A. The

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:
Buses
1 2 3
Elements
1 -1
A
b
Branches
2 -1
A = 3 -1
=
4 1 -1
A
l
Links 5 1 -1
6 1 -1
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 A
b
and A
l
as shown, where,
(i) A
b
is of dimension (bxb) corresponding to the branches and
(ii) A
l
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
T
A i = 0 (4)
where A
T
is the transpose of matrix A and i is the vector of branch currents. Similarly
for the branch voltages we can write,
v = A
bus
E (5)
16
Examples on Bus Incidence Matrix:
Example-2: For the sample network-oriented graph shown in Fig. E2, by selecting a
tree, T(1,2,3,4), obtain the incidence matrices A and A

. Also show the partitioned

form of the matrix-A.
Fig. E2. Sample Network-Oriented Graph
nodes
A

= Elements

1 0 1 0 0 7
0 0 1 1 0 6
0 1 1 0 0 5
1 1 0 0 0 4
1 0 0 0 1 3
0 0 1 0 1 2
0 0 0 1 1 1
4 3 2 1 0 \ n e
buses
A = Elements

1 0 1 0 7
0 0 1 1 6
0 1 1 0 5
1 1 0 0 4
1 0 0 0 3
0 0 1 0 2
0 0 0 1 1
4 3 2 1 \ b e
17
Corresponding to the Tree, T(1,2,3,4), matrix-A can be partitioned into two sub-
matrices as under:
buses
A
b
= branches

1 1 0 0 4
1 0 0 0 3
0 0 1 0 2
0 0 0 1 1
4 3 2 1 \ b b
buses
A
l
= links

1 0 1 0 7
0 0 1 1 6
0 1 1 0 5
4 3 2 1 \ b l
Example-3: For the sample-system shown in Fig. E3, obtain an oriented graph. By
selecting a tree, T(1,2,3,4), obtain the incidence matrices A and A

. Also show the

partitioned form of the matrix-A.
Fig. E3a. Sample Example network
Consider the oriented graph of the given system as shown in figure E3b, below.
18
Fig. E3b. Oriented Graph of system of Fig-E3a.
Corresponding to the oriented graph above and a Tree, T(1,2,3,4), the incidence
matrices and A can be obtained as follows:
e\n 0 1 2 3 4
1 1 -1
2 1 -1
A

=
3 1 -1
4 1 -1
5 1 -1
6 -1 1
7 1 -1
8 -1 1
9 -1 1
e\b 1 2 3 4
1 -1
2 -1
A =
3 -1
4 -1
5 1 -1
6 -1 1
7 1 -1
8 -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:
e\b 1 2 3 4
1 -1
A
b
=
2 -1
3 -1
4 -1
e\b 1 2 3 4
5 1 -1
A
l
=
6 -1 1
7 1 -1
8 -1 1
9 -1 1
19
PRIMITIVE NETWORKS
So far, the matrices of the interconnected network have been defined. These matrices
contain complete information about the network connectivity, the 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 behaviour of the
individual elements which make the network, along with the 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
contain active or passive components. Figure 2 represents the alternative impedance
and admittance forms of representation of a general network component.
E
p
p E
p
p
i
pq
(i
pq
+j
pq
)
j
pq
v
pq
= E
p
- E
q
i
pq
i
pq
E
q
q E
q
q
Fig.2 Representation of a primitive network element
(a) Impedance form (b) Admittance form
e
pq
z
pq
y
pq
20
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,
v
pq
= voltage across the element p-q,
e
pq
= source voltage in series with the element p-q,
i
pq
= current through the element p-q,
j
pq
= source current in shunt with the element p-q,
z
pq
= self impedance of the element p-q and
y
pq
= self admittance of the element p-q.
Performance equation: Each element p-q has two variables, v
pq
and i
pq
. The
performance of the given element p-q can be expressed by the performance equations
as under:
v
pq
+ e
pq
= z
pq
i
pq
(in its impedance form)
i
pq
+ j
pq
= y
pq
v
pq
(in its admittance form) (6)
Thus the parallel source current jpq in admittance form can be related to the series
source voltage, epq in impedance form as per the identity:
j
pq
= - y
pq
e
pq
(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. 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)
Primitive network matrices:
A diagonal element in the matrices, [z] or [y] is the self impedance z
pq-pq
or self
admittance, y
pq-pq
. An off-diagonal element is the mutual impedance, z
pq-rs
or mutual
admittance, y
pq-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
21
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.
Examples on Primitive Networks:
Example-4: Given that the self impedances of the elements of a network referred by
the bus incidence matrix given below are equal to: Z
1
=Z
2
=0.2, Z
3
=0.25, Z
4
=Z
5
=0.1
and Z
6
=0.4 units, draw the corresponding oriented graph, and find the primitive
network matrices. Neglect mutual values between the elements.
-1 0 0
0 -1 0
A = 0 0 -1
1 -1 0
0 1 -1
1 0 -1
Solution:
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
A

=
1 0 0 -1
0 1 -1 0
0 0 1 -1
0 1 0 -1
22
Based on the conventional definitions of the elements of A

, the oriented graph can be

formed as under:
Fig. E4 Oriented Graph
Thus the primitive network matrices are square, symmetric and diagonal matrices of
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
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
0 5.0 0 0 0 0
[y] =
0 0 4.0
0 0 0
0 0 0 10 0 0
0 0 0 0 10 0
0 0 0 0 0 2.5
23
Example-5: Consider three passive elements whose data is given in Table E5 below.
Form the primitive network impedance matrix.
Table E5
Element
number
Self impedance (z
pq-pq
) Mutual impedance, (z
pq-rs
)
Bus-code,
(p-q)
Impedance in
p.u.
Bus-code,
(r-s)
Impedance in
p.u.
1 1-2 j 0.452
2 2-3 j 0.387 1-2 j 0.165
3 1-3 j 0.619 1-2 j 0.234
Solution:
1-2 2-3 1-3
1-2
j 0.452 j 0.165 j 0.234
[z] =
2-3
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
The off-diagonal elements are mutual impedances between the corresponding
elements.
Matrices [z] and [y] are inter-invertible.
24
FORMATION OF Y
BUS
AND Z
BUS
The bus admittance matrix, Y
BUS
plays a very important role in computer aided power
system analysis. It can be formed in practice by either of the methods as under:
1. Rule of Inspection
2. Singular Transformation
3. Non-Singular Transformation
4. Z
BUS
Building Algorithms, etc.
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) relating
the bus vectors of currents and voltages through the bus impedance matrix and bus
admittance matrix:
E
BUS
= Z
BUS
I
BUS
I
BUS
= Y
BUS
E
BUS
(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
currents and voltages through the branch impedance matrix and branch admittance
matrix:
E
BR
= Z
BR
I
BR
I
BR
= Y
BR
E
BR
(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 voltages through the branch impedance matrix and branch admittance matrix:
E
LOOP
= Z
LOOP
I
LOOP
I
LOOP
= Y
LOOP
E
LOOP
(11)
Of the various network matrices refered above, the bus admittance matrix (Y
BUS
) and
the bus impedance matrix (Z
BUS
) are determined for a given power system by the rule
of inspection as explained next.
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 Kirchhoffs Current
Law principle at the nodal points, we get the relations as under:
At node 1: I
1
=Y
1
V
1
+ Y
3
(V
1
-V
3
) + Y
6
(V
1
V
2
)
At node 2: I
2
=Y
2
V
2
+ Y
5
(V
2
-V
3
) + Y
6
(V
2
V
1
)
At node 3: 0 = Y
3
(V
3
-V
1
) + Y
4
V
3
+ Y
5
(V
3
V
2
) (12)
25
Fig. 3 Example System for finding Y
BUS
These are the performance equations of the given network in admittance form and
they can be represented in matrix form as:
I
1
= (Y
1
+Y
3
+Y
6
) -Y
6
-Y
3
V
1
I
2
= -Y
6
(Y
2
+Y
5
+Y
6
) -Y
5
V
2
0 = -Y
3
-Y
5
(Y
3
+Y
4
+Y
5
) V
3
(13)
In other words, the relation of equation (9) can be represented in the form
I
BUS
= Y
BUS
E
BUS
(14)
Where, Y
BUS
is the bus admittance matrix, I
BUS
& E
BUS
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
the given system diagram:
Diagonal elements: A diagonal element (Y
ii
) of the bus admittance matrix,
Y
BUS
, 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 (Y
ij
) of the bus admittance
matrix, Y
BUS
, is equal to the negative of the admittance value of the
connecting element present between the buses I and j, if any.
This is the principle of the rule of inspection. Thus the algorithmic equations for the
rule of inspection are obtained as:
Y
ii
= y
ij
(j = 1,2,.n)
Y
ij
= - y
ij
(j = 1,2,.n) (15)
26
For i = 1,2,.n, n = no. of buses of the given system, y
ij
is the admittance of element
connected between buses i and j and y
ii
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
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.
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:
Example 6: Obtain the bus admittance matrix for the admittance network shown
aside by the rule of inspection
Example 7: Obtain Y
BUS
for the impedance network shown aside by the rule of
inspection. Also, determine Y
BUS
for the reduced network after eliminating the eligible
unwanted node. Draw the resulting reduced system diagram.
16 -8 -4
Y
BUS
= j -8 24 -8
-4 -8 16
-9.8 5 4
Y
BUS
= j 5 -16 10
4 10 -14
Z
BUS
= Y
BUS
-1
27
SINGULAR TRANSFORMATIONS
The primitive network matrices are the most basic matrices and depend purely on 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, Y
BUS
and Bus impedance matrix, Z
BUS
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 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
Y
BUS
New
= Y
A
-Y
B
Y
D
-1
Y
C
Y
BUS =
-8.66 7.86
7.86 -8.86
Y
BUS
New
= Y
A
-Y
B
Y
D
-1
Y
C
Y
BUS
= j -8.66 7.86
7.86 -8.66
28
performance equation relating the bus voltages to bus current injections in bus frame
of reference in admittance form is given by
I
BUS
= Y
BUS
E
BUS
(17)
Where E
BUS
= vector of bus voltages measured with respect to reference bus
I
BUS
= Vector of currents injected into the bus
Y
BUS
= bus admittance matrix
The performance equation of the primitive network in admittance form is given by
i + j = [y] v
Pre-multiplying by A
t
(transpose of A), we obtain
A
t
i +A
t
j = A
t
[y] v (18)
However, as per equation (4),
A
t
i =0,
since it indicates a vector whose elements are the algebraic sum of element currents
incident at a bus, which by Kirchhoffs law is zero. Similarly, A
t
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,
A
t
j = I
BUS
(19)
Thus from (18) we have, I
BUS
= A
t
[y] v (20)
However, from (5), we have
v =A E
BUS
And hence substituting in (20) we get,
I
BUS
= A
t
[y] A E
BUS
(21)
Comparing (21) with (17) we obtain,
Y
BUS
= A
t
[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 ,
Z
BUS
= Y
BUS
-1
(23)
Note: This transformation can be derived using the concept of power invariance,
however, since the transformations are based purely on KCL and KVL, the
transformation will obviously be power invariant.
29
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 T(1,2,3).
The data is given in Table E8.
Fig E8 System for Example-8
Table E8: Data for Example-8
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 -
Solution:
The bus incidence matrix is formed taking node 1 as the reference bus.
30
A =

1 0 1
0 0 1
1 1 0
0 1 0
0 0 1
The primitive incidence matrix is given by,
[z]=

2 . 0 0 . 0 0 . 0 0 . 0 0 . 0
0 . 0 4 . 0 0 . 0 0 . 0 2 . 0
0 . 0 0 . 0 5 . 0 0 . 0 0 . 0
0 . 0 0 . 0 0 . 0 5 . 0 1 . 0
0 . 0 2 . 0 0 . 0 1 . 0 6 . 0
j
j j
j
j j
j j j
The primitive admittance matrix [y] = [z]
-1
and given by,
[y]=

0 . 5 0 . 0 0 . 0 0 . 0 0 . 0
0 . 0 0208 . 3 0 . 0 2083 . 0 0417 . 1
0 . 0 0 . 0 0 . 2 0 . 0 0 . 0
0 . 0 2083 . 0 0 . 0 0833 . 2 4167 . 0
0 . 0 0417 . 1 0 . 0 4167 . 0 0833 . 2
j
j j j
j
j j j
j j j
The bus admittance matrix by singular transformation is obtained as
Y
BUS
= A
t
[y] A =

0 . 7 0 . 2 0 . 5
0 . 2 0833 . 4 2083 . 0
0 . 5 2083 . 0 0208 . 8
j j j
j j j
j j j
Z
BUS
= Y
BUS
-1
=

3609 . 0 1885 . 0 2299 . 0

1885 . 0 3437 . 0 1264 . 0
2299 . 0 1264 . 0 2713 . 0
j j j
j j j
j j j
31
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
or admittances.
Network equations can be formulated for solution of the network using graph theory,
independent of the nature of elements. In the graph of a network, the tree-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.
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,
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.
32
Chapter-1-B
FORMATION OF BUS IMPEDANCE MATRIX
[CONTENTS: Node elimination by matrix algebra, generalized algorithms for Z
BUS
building, addition of BRANCH, addition of LINK, special cases of analysis,
removal of elements, changing the impedance value of an element, examples]
NODE ELIMINATION BY MATRIX ALGEBRA
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 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:
I
BUS
= Y
BUS
E
BUS
(1)
Where I
BUS
and E
BUS
are n-vectors of injected bus current and bus voltages and Y
BUS
is the square, symmetric, coefficient bus admittance matrix of order n.
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.,
I
BUS
= Y
BUS
new
E
BUS
(2)
Where Y
BUS
new
is the bus admittance matrix of the reduced network and the vectors
I
BUS
and E
BUS
are of order m. It is assumed in (1) that I
BUS
and E
BUS
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,
33
m p
I
BUS-m
m
Y
A
Y
B
E
BUS-m
I
BUS-p
=
p
Y
C
Y
D
E
BUS-p
(3)
Where the self and mutual values of Y
A
and Y
D
are those identified only with the
nodes to be retained and removed respectively and Y
C
=Y
B
t
is composed of only the
corresponding mutual admittance values, that are common to the nodes m and p.
Now, for the p nodes to be eliminated, it is necessary that, each element of the vector
I
BUS-p
should be zero. Thus we have from (3):
I
BUS-m
= Y
A
E
BUS-m
+ Y
B
E
BUS-p
I
BUS-p
= Y
C
E
BUS-m
+ Y
D
E
BUS-p
= 0 (4)
Solving, E
BUS-p
= - Y
D
-1
Y
C
E
BUS-m
(5)
Thus, by simplification, we obtain an expression similar to (2) as,
I
BUS-m
= {Y
A
- Y
B
Y
D
-1
Y
C
} E
BUS-m
(6)
Thus by comparing (2) and (6), we get an expression for the new bus admittance
matrix in terms of the sub-matrices of the original bus admittance matrix as:
Y
BUS
new
= {Y
A
- Y
B
Y
D
-1
Y
C
} (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
Y
D
(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 more
advantageous to eliminate the unwanted nodes from the given network by considering
one node only at a time for elimination, as discussed next.
34
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 Y
D
then
would be a single element matrix and hence it inverse would be just equal to its own
reciprocal value. Thus the generalized algorithmic equation for finding the elements
of the new bus admittance matrix can be obtained from (6) as,
Y
ij
new
= Y
ij
old
Y
in
Y
nj
/ Y
nn
i,j = 1,2, n. (8)
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:
Example-1: Obtain Y
BUS
for the impedance network shown below by the rule of
inspection. Also, determine Y
BUS
for the reduced network after eliminating the eligible
unwanted node. Draw the resulting reduced system diagram.
The admittance equivalent network is as follows:
35
The bus admittance matrix is obtained by RoI as:
The reduced matrix after elimination of node 3 from the given system is determined
as per the equation:
Y
BUS
New
= Y
A
-Y
B
Y
D
-1
Y
C
n/n 1 2
Y
BUS
new
=
1 -j8.66 j7.86
2
j7.86 -j8.66
Alternatively,
Y
ij
new
= Y
ij
old
Y
i3
Y
3j
/ Y
33
i,j = 1,2.
Y
11
= Y
11
-Y
13
Y
31
/ Y
33
= -j8.66
Y
22
= Y
22
Y
23
Y
32
/ Y
33
= -j8.66
Y
12
= Y
21
= Y
12
Y
13
Y
32
/Y
33
= j7.86
Thus the reduced network can be obtained again by the rule of inspection as shown
be low.
-9.8 5 4
Y
BUS
= j 5 -16 10
4 10 -14
36
Example-2: Obtain Y
BUS
for the admittance network shown below by the rule of
inspection. Also, determine Y
BUS
for the reduced network after eliminating the eligible
unwanted node. Draw the resulting reduced system diagram.
n/n 1 2 3 4
1 -j50 0 j20 j10
Y
BUS
= 2
0 -j60 0 j72
=
Y
A
Y
B
3 j20 0 -j72 j50 Y
C
Y
D
4 j10 j72 j50 -j81
Y
BUS
New
= Y
A
-Y
B
Y
D
-1
Y
C
n/n 1 2
Y
BUS
new
=
1 -j32.12 j10.32
2
j10.32 -j51.36
Thus the reduced system of two nodes can be drawn by the rule of inspection as
under:
37
Z
BUS
building
FORMATION OF BUS IMPEDANCE MATRIX
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
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
[ ]
bus bus bus
I Z E = (9)
When expanded so as to refer to a n bus system, (9) will be of the form
n n k k
I Z I Z I Z I Z E
1 1 2 12 1 11 1
... . ....... + + + + =

n kn k kk k k k
I Z I Z I Z I Z E + + + + + = .... ......
2 2 1 1

n nn k nk n n n
I Z I Z I Z I Z E + + + + + = ..... .........
2 2 1 1
(10)
Now assume that the bus impedance matrix Z
bus
is known for a partial network of m
buses and a known reference bus. Thus, Z
bus
of the partial network is of dimension
mm. If now a new element is added between buses p and q we have the following
two possibilities:
38
(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.
q
Ref.
Fig 1a. Addition of branch p-q
Ref.
Fig 1b. Addition of link p-q
1
2
Partial
Network
p
Z
BUS i
m
0
1
2
Partial
Network
p
Z
BUS q
m
0
39
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 are
updated to take account of the effect of the link added.
ADDITION OF A BRANCH
Consider now the performance equation of the network in impedance form with the
added branch p-q, given by

q
m
p
qq qm qp q q
mq mm mp m m
pq pm pp p p
q m p
q m p
q
m
p
I
I
I
I
I
Z Z Z Z Z
Z Z Z Z Z
Z Z Z Z Z
Z Z Z Z Z
Z Z Z Z Z
E
E
E
E
E

2
1
2 1
2 1
2 1
2 2 2 22 21
1 1 1 12 11
2
1
(11)
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 y
pq-rs
is not equal to zero and Z
ij
= Z
ji
i,j=1,2,m,q (12)
To find Z
qi
:
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
E
k
= Z
ki
I
i
= Z
ki
k = 1, 2,i....p,.m, q (13)
Hence, E
q
= Z
qi
; E
p
= Z
pi

Also, E
q
=E
p
-v
pq
; so that Z
qi
= Z
pi
- v
pq
i =1, 2,i....p,.m, q (14)
To find v
pq
:
In terms of the primitive admittances and voltages across the elements, the current
through the elements is given by
40

rs
pq
rs rs pq rs
rs pq pq pq
rs
pq
v
v
y y
y y
i
i
, ,
, ,
(15)
v
pq
q
I
i
=1pu
Ref.
Fig.2 Calculation for Z
qi
where
pq
i is current through element p-q
rs
i is vector of currents through elements of the partial network
pq
v is voltage across element p-q
pq pq
y
,
is self admittance of the added element
rs pq
y
,
is the vector of mutual admittances between the added elements p-q and
elements r-s of the partial network.
rs
v is vector of voltage across elements of partial network.
pq rs
y
,
is transpose of
rs pq
y
,
.
rs rs
y
,
is the primitive admittance of partial network.
Since the current in the added branch p-q, is zero, 0 =
pq
i . We thus have from (15),
0
, ,
= + =
rs rs pq pq pq pq pq
v y v y i (16)
1
2
Partial
Network
p
i
Z
BUS
m
0
41
Solving,
pq pq
rs rs pq
pq
y
v y
v
,
,
= or
( )
pq pq
s r rs pq
pq
y
E E y
v
,
,

= (17)
Using (13) and (17) in (14), we get
( )
pq pq
si ri rs pq
pi qi
y
Z Z y
Z Z
,
,

+ = q i m i = ; ...... 2 , 1 (18)
To find zqq:
The element Z
qq
can be computed by injecting a current of 1pu at bus-q, I
q
= 1.0 pu.
As before, we have the relations as under:
E
k
= Z
kq
I
q
= Z
kq
k = 1, 2,i....p,.m, q (19)
Hence, E
q
= Z
qq
; E
p
= Z
pq
; Also, E
q
=E
p
- v
pq
; so that Z
qq
= Z
pq
- v
pq
(20)
Since now the current in the added element is 0 . 1 = =
q pq
I i , we have from (15)
1
, ,
= + =
rs rs pq pq pq pq pq
v y v y i
Solving,
pq pq
rs rs pq
pq
y
v y
v
,
,
1+ =
( )
pq pq
s r rs pq
pq
y
E E y
v
,
,
1

+ = (21)
Using (19) and (21) in (20), we get
( )
pq pq
sq rq rs pq
pq qq
y
Z Z y
Z Z
,
,
1 +
+ = (22)
Special Cases
The following special cases of analysis concerning Z
BUS
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
rs pq
y
,
are zero. Further, if p
is the reference node, then E
p
=0. thus,
Z
pi
= 0 q i m i = : ...... 2 , 1
And Z
pq
= 0.
Hence, from (18) (22) Z
qi
= 0 q i m i = ; ....... 2 , 1
And
pq pq qq
z Z
,
= \ (23)
42
Case (b): If there is no mutual coupling and if p is not the ref. bus, then, from (18)
and (22), we again have,
q i m i Z Z
pi qi
= = ; .... 2 , 1 ,
pq pq pq qq
z Z Z
,
+ = (24)
ADDITION OF A LINK
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

l
m
p
ll lm li l l
mq mm mp m m
pq pm pp p p
q m p
q m p
l
m
p
I
I
I
I
I
Z Z Z Z Z
Z Z Z Z Z
Z Z Z Z Z
Z Z Z Z Z
Z Z Z Z Z
E
E
E
E
E

2
1
2 1
2 1
2 1
2 2 2 22 21
1 1 1 12 11
2
1
(25)
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 y
pq-rs
is not equal to zero and Z
ij
= Z
ji
i,j=1,2,m,l. (26)
To find Z
li
:
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, e
l
in series with element p-q, as shown. Since all other
bus currents are zero, we have from (25) that
E
k
= Z
ki
I
i
= Z
ki
k = 1, 2,i....p,.m, l (27)
Hence, e
l
= E
l
= Z
li
; E
p
= Z
pi
; E
p
= Z
pi

Also, e
l
= E
p
- E
q
- v
pq
;
So that Z
li
= Z
pi
- Z
qi
- v
pq
i=1,2,i.p,...q,.m, l (28)
43
To find v
pq
:
In terms of the primitive admittances and voltages across the elements, the current
through the elements is given by

rs
pl
rs rs pl rs
rs pl pl pl
rs
pl
v
v
y y
y y
i
i
, ,
, ,
(29)
p
v
pl
q e
l
l
I
i
=1pu
Ref.
Fig.3 Calculation for Z
li
where
pl
i is current through element p-q
rs
i is vector of currents through elements of the partial network
pl
v is voltage across element p-q
pl pl
y
,
is self admittance of the added element
rs pl
y
,
is the vector of mutual admittances between the added elements p-q and
elements r-s of the partial network.
rs
v is vector of voltage across elements of partial network.
pl rs
y
,
is transpose of
rs pl
y
,
.
rs rs
y
,
is the primitive admittance of partial network.
1
2
Partial
Network
i
Z
BUS
m
0
44
Since the current in the added branch p-l, is zero, 0 =
pl
i . We thus have from (29),
0
, ,
= + =
rs rs pl pl pl pl pl
v y v y i (30)
Solving,
pl pl
rs rs pl
pl
y
v y
v
,
,
= or
( )
pl pl
s r rs pl
pl
y
E E y
v
,
,

= (31)
However,
rs pq rs pl
y y
, ,
=
And
pq pq pl pl
y y
, ,
= (32)
Using (27), (31) and (32) in (28), we get
( )
pq pq
si ri rs pq
qi pi li
y
Z Z y
Z Z Z
,
,

+ = l i m i = ; ...... 2 , 1 (33)
To find Z
ll
:
The element Z
ll
can be computed by injecting a current of 1pu at bus-l, I
l
= 1.0 pu. As
before, we have the relations as under:
E
k
= Z
kl
I
l
= Z
kl
k = 1, 2,i....p,q,.m, l (34)
Hence, e
l
= E
l
= Z
ll
; E
p
= Z
pl
;
Also, e
l
= E
p
- E
q
- v
pl
;
So that Z
ll
= Z
pl
- Z
ql
- v
pl
i=1,2,i.p,...q,.m, l (35)
Since now the current in the added element is 0 . 1 = =
l pl
I i , we have from (29)
1
, ,
= + =
rs rs pl pl pl pl pl
v y v y i
Solving,
pl pl
rs rs pl
pl
y
v y
v
,
,
1+ =
( )
pl pl
s r rs pl
pl
y
E E y
v
,
,
1

+ = (36)
However,
rs pq rs pl
y y
, ,
=
And
pq pq pl pl
y y
, ,
= (37)
Using (34), (36) and (37) in (35), we get
45
( )
pq pq
sl rl rs pq
ql pl ll
y
Z Z y
Z Z Z
,
,
1 +
+ = (38)
Special Cases Contd.
The following special cases of analysis concerning Z
BUS
building can be considered
with respect to the addition of link to a p-network.
Case (c): If there is no mutual coupling, then elements of
rs pq
y
,
are zero. Further, if p
is the reference node, then E
p
=0. thus,
l i m i Z Z
qi li
= = ; .... 2 , 1 ,
pq pq ql ll
z Z Z
,
+ = (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 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
rs pq
y
,
are zero. Further, if p
is not the reference node, then
Z
li
= Z
pi
- Z
qi
Z
ll
= Z
pl
Z
ql
z
pq,pq
= Z
pp
+ Z
qq
2 Z
pq
+ z
pq,pq
(40)
MODIFICATION OF Z
BUS
FOR NETWORK CHANGES
An element which is not coupled to any other element can be removed easily. The
Z
bus
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 Z
bus
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 impedance. When mutually
coupled elements are removed, the Z
bus
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.
46
Examples on Z
BUS
building
Example 1: For the positive sequence network data shown in table below, obtain
Z
BUS
by building procedure.
Sl. No.
p-q
(nodes)
Pos. seq.
reactance
in pu
1 0-1 0.25
2 0-3 0.20
3 1-2 0.08
4 2-3 0.06
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.
Fig. E1: Example System
Consider building Z
BUS
as per the various stages of building through the consideration
of the corresponding partial networks as under:
Step-1: Add element1 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;
Z
BUS
(1)
=
1
1 0.25
47
Step-2: Add element2 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;
Z
BUS
(2)
=
1 3
1 0.25 0
3 0 0.2
Step-3: Add element3 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;
Z
BUS
(3)
=
1 3 2
1 0.25 0 0.25
3 0 0.2 0
2 0.25 0 0.33
Step-4: Add element4 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;
48
Z
BUS
(4)
=
1 3 2 l
1 0.25 0 0.25 0.25
3 0 0.2 0 -0.2
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:
Z
BUS
(final)
=
1 3 2
1 0.1441 0.0847 0.1100
3 0.0847 0.1322 0.1120
2 0.1100 0.1120 0.1454
Example 2: The Z
BUS
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.
Z
BUS
=
1 2 3 4 5
1
2 0 0 0 2
2
0 2 0 2 0
3
0 0 2 0 0
4
0 2 0 3 0
5
2 0 0 0 3
Solution:
The specified matrix is so structured that by its inspection, we can obtain the network
by backward analysis through the various stages of Z
BUS
building and p-networks as
under:
49
50
Thus the final network is with 6 nodes and 5 elements connected as follows with the
impedance values of elements as indicated.
Fig. E2: Resultant network of example-2
51
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?
Solution: The specified system is considered with the reference node denoted by
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)
Step2: Add branch 2, between node 2 and reference node. (q = 2, p = 0).
52
Step3: Add branch 3, between node 1 and node 3 (p = 1, q = 3)
Step 4: Add element 4, which is a link between node 1 and node 2. (p = 1, q = 2)
53
Now the extra node-l has to be eliminated to obtain the new matrix of step-4, using
the algorithmic relation:
Y
ij
new
= Y
ij
old
Y
in
Y
nj
/ Y
nn
i,j = 1,2, 3.
Step 5: Add link between node 2 and node 3 (p = 2, q=3)
54
Thus, the new matrix is as under:
Node l is eliminated as shown in the previous step:
Further, the bus admittance matrix can be obtained by inverting the bus impedance
matrix as under:
As a check, it can be observed that the bus admittance matrix, Y
BUS
can also be
obtained by the rule of inspection to arrive at the same answer.
55
Example 4: Form the bus impedance matrix for the network shown below.
Solution:
Add the elements in the sequence, 0-1, 1-2, 2-3, 0-3, 3-4, 2-4, as per the various steps
of building the matrix as under:
Step1: Add element 1, which is a branch between node-1 and reference node.
Step2: Add element 2, which is a branch between nodes 1 and 2.
Step3: Add element 3, which is a branch between nodes 2 and 3
Step4: Add element 4, which is a link from node 3 to reference node.
56
Eliminating node l,
Step5: Add element 5, a branch between nodes 3 and 4.
Step 6: Add element 6, a link between nodes 2 & 4.
Eliminating node l we get the required bus impedance , matrix
57
Example 5: Form the bus impedance matrix for the network data given below.
Element
Self Impedance Mutual Impedance
Bus
p-q
z
pq, pq
(pu)
Bus
r-s
z
pq, rs
(pu)
1 1 2(1) j0.6
2 1 2(2) j0.4 1 2(1) j0.2
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.
Step1: Add element 1 from bus 1 to 2, element 1-2(1). ( p=1, q=2, p is the reference
node)
Step2: Add element 2, element 1-2(2), which is a link from bus1 to 2, mutually
coupled with element 1, 1-2(1).
58
Consider the primitive impedance matrix for the two elements given by
Thus the primitive admittance matrix is obtained by taking the inverse of [z] as
Thus,
So that we have,
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:
-------------------