Habtamu Eshete PDF

ADDIS ABABA UNIVERSITY
INSTITUTE OF TECHNOLOGY
SCHOOL OF ELECTRICAL AND COMPUTER ENGINEERING
STUDY ON POWER LOSS MINIMIZATION FOR DISTRIBUTION NETWORK

RECONFIGURATION USING GENETIC ALGORITHM
CASE STUDY: ADDIS NORTH 132/15 KV SUBSTATION
By
HABTAMU ESHETE
ADVISOR
Mr. KIROS TESFAY
A Thesis Submitted to the School of Graduate Studies of Addis Ababa University in Partial
Fulfillment of the Requirements for the Degree of Masters of Science in Electrical Engineering
June, 2018
Addis Ababa, Ethiopia

ADDIS ABABA UNIVERSITY
INSTITUTE OF TECHNOLOGY
SCHOOL OF ELECTRICAL AND COMPUTER ENGINEERING
STUDY ON POWER LOSS MINIMIZATION FOR DISTRIBUTION NETWORK

RECONFIGURATION USING GENETIC ALGORITHM
CASE STUDY: ADDIS NORTH 132/15 KV SUBSTATION
Habtamu Eshete
Approval by Board of Examiners
___________________________ ____________________
Chairman, Dept. Graduate Committee Signature
_______________________ ______________________
Advisor Signature
________________________ _______________________
Internal Examiner Signature
_______________________ ______________________
External Examiner Signature
Habtamu Eshete, AAiT, MSc. Thesis Page I

Declaration
I, the undersigned, declare that this thesis is my original work, has not been Presented for a
degree in this or any other university, and all sources of materials used for the thesis have been
fully acknowledged.
Habtamu Eshete ____________________
Name Signature
Place: Addis Ababa Institute of Technology
Date of Submission: ________________
This thesis has been submitted for examination with my approval as a university advisor.
___________________________ _______________________
Advisor‟s Name Signature
Habtamu Eshete, AAiT, MSc. Thesis Page II

ACKNOWLEDGEMENT
I would like to express my sincere gratitude to my advisor Mr. Kiros Tesfay for his help
throughout the thesis including suggesting this interesting idea and guiding me to the end. I thank
also, Addis Ababa University for giving me this opportunity to continue the MSc program. I
would like to acknowledge Mr. Kassay and Abiy, who are Addis North 132/15 KV Substation
worker for their cooperation and support.
Habtamu Eshete, AAiT, MSc. Thesis Page III

ABSTRACT
Distribution system is a largest portion of network of electrical power system. Different actions
have been taken to improve the efficiency of the distribution network. One of the most basic and
the commonest way to improve the performance of the distribution network is the network
reconfiguration. Electric distribution systems reconfiguration comprises tie and sectionalizing
switches. Tie switches are normally open, and sectionalizing switches are normally closed. By
opening and closing these switches, the distribution network can be reconfigured. This
reconfiguration can be done for the objective of loss minimization. In order to get feasible results
(loss minimization), the reconfiguration must meet some constraints, like Kirchhoff‟s voltage
and current laws, other equality and inequality constraints.
Distribution network reconfiguration is an optimization problem and needs a suitable algorithm

(method).The method used for this optimization problem is a genetic algorithm optimization
method. The genetic algorithm has been described in detail and then applied specifically to the
network reconfiguration problem. In the optimization process, load flow of the distribution
system was computed. Then, computer simulation was performed by using DIgSILENT
PowerFactory software for analyzing the distribution network reconfiguration. In addition,
optimized genetic algorithm was used as a tool for network reconfiguration.
Addis North 132/15 kV substation feeders are used as test system for this particular study.
Before the reconfiguration, the power loss is 3.983783 MW, after reconfiguration the power loss
has decreased from 3.983783 MW to 1.594640 MW, which is 2.389143 MW (59.91716%)
reduction. Besides, the maximum voltage drop before reconfiguration is 0.352639, and the
reconfiguration increased it to 5.050102. The minimum voltage before reconfiguration is
0.992618 p.u., and the minimum voltage is found to be 0.947288 p.u. after reconfiguration. In
addition, the maximum voltage before reconfiguration is 0.992069 p.u., and it is found to be
0.9953500 p.u. after reconfiguration. Based on the findings of this research, it is concluded that
reconfiguration of distribution networks can reduce power loss and operating cost as well as
improves the voltage profile of distribution systems. Hence, it is recommended that all the
distribution network feeders of Addis Ababa city to be reconfigured for the betterment of the
Ethiopian Electric Utility (EEU) services.
Keywords: Network Reconfiguration, Genetic Algorithm Optimization, Power Loss Minimization
Habtamu Eshete, AAiT, MSc. Thesis Page IV

TABLE OF CONTENTS
ACKNOWLEDGEMENT ............................................................................................................ III
ABSTRACT .................................................................................................................................. IV
LIST OF FIGURES ................................................................................................................... VIII
LIST OF TABLES ........................................................................................................................ IX
LIST OF ACRONYMS ................................................................................................................. X
Chapter 1 Introduction .................................................................................................................... 1
1.1 Background ................................................................................................................... 1

1.2 Motivation ..................................................................................................................... 3
1.3 Statement of the problem .............................................................................................. 3
1.4 Scope of the research .................................................................................................... 3
1.5 Objectives of the Study ................................................................................................. 3
1.5.1 General Objective .......................................................................................... 3
1.5.2 Specific Objectives ........................................................................................ 4
1.6 Methodology ................................................................................................................. 4

1.7 Organization of the thesis ............................................................................................. 5
Chapter 2 ......................................................................................................................................... 6
Theoretical Background and Literature Review ............................................................................. 6
2.1 Theoretical Background ................................................................................................ 6

2.2 Literature Review.......................................................................................................... 6
2.2.1 Methods Based on Evolutionary Techniques for Network Reconfiguration . 6
2.2.1.1 Genetic Algorithm (GA) .................................................................. 7
2.2.1.2 Simulated Annealing ........................................................................ 8
2.2.1.3 Artificial Neural Network ................................................................ 8
2.2.1.4 Ant Colony Optimization ................................................................. 9
2.2.1.5 Particle Swarm Optimization ........................................................... 9
2.2.2 Knowledge Based Methods ......................................................................... 10
Habtamu Eshete, AAiT, MSc. Thesis Page V

2.2.2.1 Heuristic optimization methods For Network Reconfiguration ... 10
2.2.2.2 Linear programming Techniques ................................................... 12
2.2.2.3 Fuzzy Logic Method ...................................................................... 12
2.2.2.4 Tabu Search Method ...................................................................... 13
2.2.3 Mixed Methods ............................................................................................ 13
Chapter 3 ....................................................................................................................................... 15
Research Methodology ................................................................................................................. 15
3.1 Problem formulation ................................................................................................... 15

3.1.1 Mathematical Model of the Network Reconfiguration Problem ................. 15
3.2 Genetic Algorithm ...................................................................................................... 18

3.2.1 Genetic Algorithm Optimization ................................................................. 18
3.2.2 How Genetic Algorithms Work ................................................................... 18
3.2.3 Genetic Algorithm Operators ....................................................................... 21
3.2. 4 Encoding ..................................................................................................... 30
3.3 DIgSILENT PowerFactory Overview ........................................................................ 31

Chapter 4 ....................................................................................................................................... 33
Simulation Studies and Analysis of Results ................................................................................. 33
4.1 Addis North 132/15 KV Substation Test Systems......................................................... 33

4.1.1 Addis North 132/15 KV Substation System Description............................... 33
4.1.2 Addis North 132/15 KV Substation Simulation Result ................................. 33
4.2 Genetic Algorithm optimization result ....................................................................... 40

4.3 Feasibility Analysis ..................................................................................................... 43
Chapter 5 ....................................................................................................................................... 44
Conclusions, Recommendations and Future Work ....................................................................... 44
5.1 Conclusions ................................................................................................................. 44

5.2 Recommendations ....................................................................................................... 44
5.3 Future Work ................................................................................................................ 45
Habtamu Eshete, AAiT, MSc. Thesis Page VI

References ..................................................................................................................................... 46
Appendix A: Addis North 132/15 kV Substation feeders before reconfiguration ....................... 51
Appendix B: Addis North 132/15 kV Substation feeders network after reconfiguration............. 61
Appendix C: Addis North 132/15 kV Substation feeders data ..................................................... 71
Appendix D: Matlab program for reconfiguration of distribution network using Genetic

algorithm ...................................................................................................................................... 80
Habtamu Eshete, AAiT, MSc. Thesis Page VII

LIST OF FIGURES
Figure1.1 Addis North 132/15 kV Substation one line diagram .................................................... 1
Figure 1.2 Distribution system structure ......................................................................................... 2
Figure 3.1 Optimization Process ................................................................................................... 18
Figure 3.2 Flow chart of Genetic Algorithm Optimization technique ......................................... 20
Figure 3.3 Population, Chromosomes and Genes ......................................................................... 21
Figure 3.4 Roulette wheel selections. [29], [30] ............................................................................... 24
Figure 3.5 Rank selection effects. (a) Before ranking. (b) After ranking [31], [32], [33], [34] ............. 25
Figure 3.6 Crossover point ........................................................................................................... 26
Figure 3.7 Exchanging genes among parents .............................................................................. 26
Figure 3.8 New offspring ............................................................................................................. 26
Figure 3.9 Single Point Crossover ............................................................................................... 27
Figure 3.10 Multi point crossover ................................................................................................ 28
Figure 3.11 Uniform crossovers .................................................................................................. 28
Figure 3.12 Mutation: Before and After ....................................................................................... 29
Figure 3.13 Single point mutation ............................................................................................... 29
Figure 3.14 Multi point mutation ................................................................................................. 30
Figure 4.1 Feeders output calculation analysis chart before optimization.................................... 35
Figure 4.2 Feeders output calculation analyses chart after optimization ...................................... 40
Figure A.1 Addis North 132/15 kV substation outgoing feeders online diagram ........................ 51
Figure A.2 Addis North 132/15 kV substation feeder 1 one line diagram ................................... 52
Figure A.8 Addis North 132/15 kV Substation feeders network after reconfiguration ................ 70
Habtamu Eshete, AAiT, MSc. Thesis Page VIII

LIST OF TABLES
Table 4-1 Feeder output calculation analysis before optimization .............................................. 34
Table 4-2 Tie Open Point Optimization........................................................................................ 35
Table 4-3 Optimal Tie Open Positions ......................................................................................... 36
Table 4-4 Necessary Switching Actions (From Initial to Optimized Configuration) ................... 36
Table 4-5 Feeders Results ............................................................................................................. 38
Table 4-6 Feeder output calculation analysis after optimization shown table below ................... 39
Table 4-7 Genetic algorithm test results for feeder 1 and feeder 2 ............................................... 41
Table A-1 Network Data of Feeder 1 of Addis North 132/15 kV Substation .............................. 71
Table A-2 Network Data of Feeder 2 of Addis North 132/15 kV substation ............................... 74
Table A- 4 Network Data of Feeder 4 of Addis North 132/15 kV substation .............................. 77
Table A- 6 Network Data of Feeder 6 of Addis North 132/15 kV Substation ............................. 79
Habtamu Eshete, AAiT, MSc. Thesis Page IX

LIST OF ACRONYMS
F1 Feeder One
F2 Feeder Two
F3 Feeder Three
F4 Feeder Four
F5 Feeder Five
F6 Feeder Six
R Resistance
X Reactance
Z Impedance
I Current
P Active Power
Q Reactive Power
S Apparent Power
Sloss Loss Power
Va Voltage at node a
Vb Voltage at node b
Vmin Minimum Voltage
Vmax Maximum Voltage
Habtamu Eshete, AAiT, MSc. Thesis Page X

Chapter 1 Introduction
1.1 Background
Distribution system is a largest portion of network of electrical power system. Different actions
have been taken to improve the efficiency of the distribution network. One of the most basic and
the commonest way to improve the performance of the distribution network is the network
reconfiguration. Feeder reconfiguration is performed by opening sectionalizing (normally closed)
and closing tie (normally open) switches of the network.
Addis North 132/15 KV substation has six outgoing feeders. The one line diagram is shown in
figure 1.1, which is obtained from Addis North 132/15 KV substation.
Figure1.1 Addis North 132/15 KV Substation one line diagram
Habtamu Eshete, AAiT, MSc. Thesis Page 1

Figure 1.2 Distribution system structure
a. Radial: For radial, only one path between substation or service transformer and customer.
The power flow is from substation to customer along single path. Furthermore, radial is
cheap and predictable to use and simple to analyze. Besides, it is easy to analyze minimize
power losses.
b. Mesh: For mesh, multiple paths between all point in the network. Power flow between any
two points is split along several paths. It is most reliable method of distributing electricity.
If failure occurs, power instantly reroutes itself. Other than that, it is usually used in high
density urban area where maintenance and repairs are difficult and costly.
c. Loop: For loop, two paths between substation transformer and customer. Power flow is
usually from both sides to the middle. The equipment is rated so that service can be
maintained if an open point occurs in the system.

1.2 Motivation
In Addis Ababa, the growing industrialization and increasing standard of living considerably
increase the usage of energy. The increasing demand of the electrical energy is quietly related to
the power demand. The increasing electricity demand has made distribution system more complex
and often caused power loss. To reduce the power loss, the network distribution system needs to be
reconfigured.
1.3 Statement of the problem

In order to remain competitive, it is becoming more and more imperative for power distribution
companies to be able to meet the demands of their customers efficiently. This means that one of
their goals is finding an operating state for a large, balanced, three-phase, distribution network
which minimizes the cost for the power company supplying the power, while satisfying the
requirements of the customer. The demand for the electricity is rising due to the population growth.
The distribution system has become more complex. The increasing of current draw during the
distribution of electricity lead to the instability that often is the cause for power losses. Coordinated
reconfiguration and upgrading of the distribution system can be a solution to this problem.
1.4 Scope of the research
This thesis is delimited to the study of power loss minimization in Addis North 132/15 KV
Substation. The power loss and distribution system of the selected area has been first studied. This
has been followed with reconfiguration of the networking system aiming to minimize the power
loss.
1.5 Objectives of the Study

1.5.1 General Objective
The main objective of this thesis is to minimize the electrical power losses by considering
reconfiguration method and applying a Genetic Algorithm for optimization in Addis North 132/15
KV distribution substation.

1.5.2 Specific Objectives
The specific objectives of the thesis are:
 To study the Addis North 132/15 kV substation and the distribution system.
 To formulate the objective function and solve the optimal reconfiguration problem using
genetic algorithm.
 To study reconfiguration method of power loss reduction.
 To reconfigure the distribution system by the use of tie and sectionalizing switches.
 To minimize distribution losses by optimal reconfiguration of the existing network.
 To evaluate the distribution loss with and without reconfiguration of the distribution
network.
 To investigate the voltage profile with and without reconfiguration of the distribution
system.
 To draw conclusions based on the finding of this research and make recommendation to
Ethiopia Electric Utility (EEU) for possible improvement in performance of Addis Ababa
distribution system.
1.6 Methodology
The following methodology are adopted in carrying out this research and are itemized as follows;
i. Collection of data for radial distribution network of Addis North 132/15 KV distribution
substation network.
ii. Establishment of the most suitable method for performing power flow analysis for the radial
distribution network.
iii. Based on ii, a power flow analysis is carried out to estimate the initial configuration for
distribution network.
iv. Development of genetic algorithm model used to determine the optimal location of tie and
sectionalizing switches considering active power loss and total voltage profile.
v. Testing and validation of the developed model.
vi. A comparison of the reconfigure system with the original system is carried out, to
determine the extent of loss reduction.

1.7 Organization of the thesis
The thesis is organized as follows. In Chapter 2, a theoretical background and literature about
network reconfiguration and different optimization methods used earlier is reviewed. In chapter 3,
a problem formulation and the new genetic algorithm methodology which is used in this thesis is
discussed in detail. In Chapter 4, the Addis North 132/15 KV substation distribution network is
reconfigured using DIgSILENT PowerFactory and the genetic algorithm optimization using Matlab
softwares simulation results and discussions are discussed in detail. In chapter 5, Conclusion,
recommendation and future work are made in the last chapter.

Chapter 2
Theoretical Background and Literature Review
2.1 Theoretical Background
More than 40 years, the French engineers Merlin and Back [8] introduced the opportunity to reduce
technical losses by exploring a change in the status of normally closed and normally open switches.
They proposed the “network reconfiguration problem”, for which the solution should provide the
best status for all the switches in a primary distribution network, best in the sense that they provide
a radial configuration supplying loads with the minimum of power loss. Between 1988 and 1990,
heuristic methods were used to solve the problem. The developments during this period focused on
increasing the number of operating constraints. From 1990, new solution strategies appeared: linear
programming simulated annealing, and genetic algorithms, whose objective function is power
losses minimization and the operating constraints previously used. In addition, load models are
improved with more precise models.
In 1993, solutions to the problem were presented through neural networks, which initially model
few operating constraints and simple load models. In 1997, models with more constraints were
used. Between 1995 and 1996 the heuristic method was proposed again in order to optimize energy
losses using more precise load models. From 1997 until now, the techniques used are combinations
of the previous techniques, aiming to complement each method's strengths. Also new methods like
Ant Colony (ACO) and Particle Swarm Optimization (PSO) are used. According to the historic
development, the computational searching methods are classified into three large groups:
 Methods based on evolutionary techniques for network reconfiguration.

 Knowledge-based methods for network reconfiguration.
 Mixed methods for network reconfiguration.
2.2 Literature Review

2.2.1 Methods Based on Evolutionary Techniques for Network Reconfiguration
Evolutionary algorithms (EAs) are developed to arrive at optimum or near-optimum solutions to a
large scale optimization problem. The problem having very large number of decision variables and
non-linear objective functions are often solved by EAs. EAs mimic the metaphor of natural
biological evolution or social behavior like how ants find the shortest route to a source of food and

how birds find their destination during migration. The behavior of such species is guided by
learning and adaptation. The evolutionary algorithms are based on population based search
procedures that incorporate random variation and selection.
EAs have an important characteristic: the lack of a rigorous mathematical formulation that allows
establishing their operation in each situation with certainty. These techniques start from a solution
and improve it. Genetic algorithm, simulated annealing, neural networks, ant colony (ACO),
particle swarm optimization (PSO), etc. belong to this category. Genetic algorithm is a search
based on the mechanism of natural selection and natural genetics. It can be used to solve the multi
objective optimization problem. It is used by [1], [2], [3]. Simulated annealing method can avoid
local optima but requires an excessive computation time. It is used by [4]. Artificial neural network
methods were also used by [5].
2.2.1.1 Genetic Algorithm (GA)

The first evolutionary-based optimization technique was the genetic algorithm (GA).Genetic
algorithms have become very popular as a method of finding global optimums. As applied to
reconfiguration [1], [2], the switch states are encoded in strings of 0/1 "chromosomes", and a
population of, for example, 30 topologies is built at random. At each iteration, two parent
topologies are selected at random for crossbreeding which is a process of combining the
chromosomes according to some defined algorithm. Then mutation, a random alteration of some
chromosomes, may occur with a certain probability. If the resulting child is better, it replaces an
existing topology in the population of 30. This process of crossbreeding continues for a number of
iterations. The population also has to be re-seeded periodically with random strings to avoid in
breeding. As the population evolves, there will always be a best solution that should steadily
improve.
Genetic algorithms are most attractive for parallel processing environments, and when each child
can be evaluated quickly. The cross breeding and mutation algorithms must be custom designed
and tested for each application. Parameters such as the number of crossbreeding per generation,
mutation probability, number of generations, population size, and percentage of population
reseeded must all be determined by testing.
Applications to reconfiguration have used simplified network analysis because many thousands of
topologies are considered, so the resulting solution may not be optimal with a more detailed model.

Nara [2] used the genetic algorithm (GA) which is a search algorithm based on the mechanism of
natural selection and natural genetics. It combines the adaptive nature of the natural genetics or the
evolution procedure of organs with functional optimization. The simple feature of GA makes it
suitable for different multi objectives optimization problem. The principle problem in using GA
rests on an efficient coding and decoding mechanism of the chromosome representing the
distribution network and the structure of fitness function.
Fudou, Fukuyama and Nakanishi [3] present a GA using three phases unbalanced load flow. A
proper string representation for loads and power supplies is devised and a method to yield a good
problem dependent initial string population is presented. A repair operator which modifies the
string so as to improve the objective function of the problem and to satisfy the radial network
constraints. A modification to the fitness function is made to reinforce the satisfaction of the power
source limits and voltage as well as current constraints.
2.2.1.2 Simulated Annealing

The name and inspiration came from annealing process in metallurgy, a technique involving
heating and controlled cooling of a material to increase the size of its crystals and reduce their
defects. The heat causes the atoms to become unstuck from their initial positions (a local
minimum of the internal energy) and wander randomly through states of higher energy; the slow
cooling gives them more chances of finding configurations with lower internal energy than the
initial one. In the simulated annealing method each point of the search space is compared to a
state of some physical system, and the function to be minimized is interpreted as the internal
energy of the system in that state. Therefore the goal is to bring the system, from an arbitrary
initial state, to a state with the minimum possible energy. Ray Daniel Zimmerman [4] used
simulated annealing for loss reduction of three phase power distribution system. It is used on
larger, more complex, unbalanced three-phase system. Computer program was developed and was
able to find the optimal configuration of the 147 bus, 12 switch example systems, demonstrating
the feasibility of such an approach for the solution of the network reconfiguration problem.
2.2.1.3 Artificial Neural Network

Ali Reza Fereidunian, Hamid Lesani and Caro Lucas [5] used an intelligent neural optimizer with
two objective functions which is designed for electrical distribution systems. The presented method
is faster than alternative optimization methods and is comparable with the most powerful and

precise ones. The optimizer is much smaller than similar neural systems. The proposed method has
established a relation between two applications of neural networks: Optimization and Pattern
Recognition.
H. Kim [6] presented the strategy of feeder reconfiguration to reduce the power loss by artificial
neural (ANN) network. This approach developed is basically different in the aspects that the load
transfer and the corresponding load flow solution during the search process are not required. The
set of ANN is the optimal system topology corresponding to various load patterns which minimizes
the load under given conditions.
2.2.1.4 Ant Colony Optimization

Ant colony (ACO) optimization is based on ant social behavior. In the real world, ants (initially)
wander randomly, and upon finding food return to their colony while laying down pheromone
trails. If other ants find such a path, they are likely not to keep traveling at random, but instead
follow the trail laid by earlier ants, returning and reinforcing it, if they eventually find any food.
Over time, however, the pheromone trail starts to evaporate, thus reducing its attractive strength.
The more time it takes for an ant to travel down the path and back again, the more time the
pheromones have to evaporate. A short path, by comparison, gets marched over faster, and thus the
pheromone density remains high as it is laid on the path as fast as it can evaporate. Thus, when one
ant finds a good (short) path from the colony to a food source, other ants are more likely to follow
that path, and such positive feedback eventually leaves all the ants following a single path.
It is used by [9], [10], [11] used ACO for optimal reconfiguration of distribution network for loss
minimization and load balancing. Matlab program which uses ACO was developed and tested on
Baran and Wu 33 bus system and two large practical distribution systems. The results were also
validated using Power World simulation software.
2.2.1.5 Particle Swarm Optimization

The Particle Swarm Optimization (PSO) is inspired by the social behavior of a flock of migrating
birds trying to reach an unknown destination. In PSO, each solution is a bird in the flock and is
referred to as a particle. A particle is analogous to a chromosome in GAs. As opposed to GAs, the
evolutionary process in the PSO does not create new birds from parent ones. Rather, the birds in
the population only evolve their social behavior and accordingly their movement towards a
destination. Each bird looks in a specific direction, and then when communicates together, they
identify the bird that is in the best location. Accordingly, each bird speeds towards the best bird
using a velocity that depends on its current position. Each bird, then investigates the search space
from its new local position, and the process repeats until the flock reaches a desired destination
Tamer M. Khalil, and Alexander V. Gorpinich [7] used selective particle swarm optimization
(SPSO) for loss reduction. This algorithm is a simple modification to the binary particles warm
optimization (BPSO). The search space of the algorithm is a set of branches (switches) which are
normally closed or normally opened, this search space may be dissimilar for different dimensions.
The process of solving reconfiguration problem is divided in to two steps. First, finding search
spaces after closing all switches and second, using SPSO to find switches that would be opened.
The presented technique is applied to a 33-node system and a 69-node system. The results obtained
via SPSO are compared with some previous methods to demonstrate the effectiveness of the
proposed algorithm.
2.2.2 Knowledge Based Methods

They are based on the operators' experience in the system operations. Based on this Knowledge,
algorithms have been designed to facilitate searching for the new distribution network
configuration, trying to find an option close to the optimal. Heuristic methods, linear programming,
expert systems, fuzzy logic etc, belong to this category. Heuristic methods used many times by
different researchers. The solution process leads to the optimum or near optimum in less
computation time.
2.2.2.1 Heuristic optimization methods For Network Reconfiguration

Several methods have been proposed to solve the reconfiguration problem. In 1975, Merlin and
Back [8] proposed a branch and bound type heuristic method to determine the network
configuration for minimum line losses. Its solution scheme starts with a meshed network by
initially closing all switches in the network. The switches are then opened one at a time until a new
radial configuration is reached. In this process the switch to be opened at each stage is selected in
order to minimize line losses of the resulting network.
Branch and bound method will work better if the initial solution is close to optimal, because more
pruning will occur. Branch and bound also benefits from breaking the problem into sub problems,
each of which can be optimized separately. Even so, branch and bound is a combinatorial method
and hence too slow for practical use. Therefore, most of the recent work on reconfiguration has

used either a branch exchange method or sequential switch opening method. Heuristics are applied
in most cases to reduce the number of switching options considered.
D. Shirmomohammadi and Hong [12] improved the method of Merlin and Back. As a result, it
shares the two principle benefits of that methodology, convergence to the optimum or near
optimum solution and the independence of the final solution from the initial status of the network
switches. At the same time, this method avoids all the major drawbacks of Merlin and Back.
Civanlar [13] developed a branch exchange method. In this method, loss reduction is achieved by
exchange operation corresponds to the selection of a pair of switches, one for opening and the other
for closing so that the resulting network has lower line losses while remaining connected and
radial.
The major drawbacks of this method are:
 The final network reconfiguration is dependent on the initial state of the network switches.
 Optimum solution is not guaranteed.
 Selection of each switches exchange operation becomes very time consuming.
Baran and Wu [14] presented a heuristic reconfiguration methodology based on the branch
exchange to reduce losses and balance the loads in the feeders. To assist in the search, two
approximated load flows for radial networks with different degrees of accuracy are used. They are
simple power flow method and back and forward update of power flow method. The method is
very time consuming due to the complicated combinations in large scale system and converges to a
local optimum solution, that is, convergence to the global optimum is not guaranteed.
It is used by [15], [16] used a combined method, the switch exchange (SEM) and sequential switch
opening method (SSOM) for reconfiguration of the network for loss reduction.
Its main advantage is optimum or near optimum configuration is obtained using one of the two
basic methods, SEM where closing switch and opening another in the loop formed. SSOM where
all tie switches are initially closed and an optimal load flow is obtained. The system is returned to a
radial configuration by successive opening of the switches having the least current flow until
network radiality is obtained. SEM requires less computation time and SSOM is independent from
the initial configuration and thus more likely leads to the actual optimum.

Broadwater and Khan [17] suggested a reconfiguration algorithm which calculates switch patterns
as a function of time. Either seasonal or daily time studies may be performed. Both manual and
automatic switches are used to reconfigure the system for seasonal studies whereas only automatic
switches are considered for daily studies. Such a continuous reconfiguration is allowed by today's
distribution automation, information technology and equipment. It is shown that switching at the
system peak can reduce losses but cause a marginal increase in system peak. The practical aspects
of such an optimization remain to be carefully analyzed through costs, transient effect and
influence to system reliability.
R. Safri, M.M.A Salama and Y. Chickani [18] proposed an algorithm that is based on network
portioning into groups of load buses such that the line section losses between the groups of nodes
are minimized, the proposed method overcomes the size restriction imposed by reconfiguration
techniques. By dividing the distribution network into groups of buses, the combinatorial nature of
reconfiguration problem is reduced while simultaneously minimizing losses.
2.2.2.2 Linear programming Techniques

A. Borghetti, M. Paolone and C.A. Nucci [19] used mixed integer linear programming (MILP)
model for the solution of the minimum loss reconfiguration problem of distribution networks,
including embedded generation. The proposed MILP model takes into account the main operating
constraints other than radiality, such as the lower and upper bounds of the bus voltages and the
upper limits of the line currents. The solution of the MILP model does not require the knowledge
of an initial feasible configuration of the network.
G. Celli, M. Loddo, F. Pilo and A. Abur [20] formulated and solved the network reconfiguration
problem with DGs using a simple linear programming approach. Optimal configurations are
determined by considering the effects of DG outputs, load variations, and various other
contingences such as faults and maintenance outages. Demand Side Management actions have also
been taken into account.
2.2.2.3 Fuzzy Logic Method

Ramadoni Syahputra [21] presented a fuzzy multi-objective approach for achieving the minimum
active power loss and the maximum voltage magnitude in order to improve the efficiency of radial
distribution networks with distributed generations. Multi-objective function are considered for load
balancing among the feeders, minimization of the real power loss, deviation of nodes voltage, and
branch current constraint violation, while subject to a radial network structure in which all loads
must be energized. The effectiveness of the method has been demonstrated by a 70-node
distribution network test system.
Abhisek Ukil and Willy Siti [22] used Fuzy Logic for load balancing. They presented a fuzzy
logic-based load balancing system along with a combinatorial optimization-based implementation
system for implementing the load changes. The input to the fuzzy step is the total load (kW) per
phase of the feeders. Output of the fuzzy step is the load change values, negative value for load
releasing and positive value for load receiving. Sum of the positive and negative values is zero, i.e.,
the total load remains unchanged for the entire phase balancing.
2.2.2.4 Tabu Search Method

Tabu Search is a meta-heuristic that guides a local heuristic search procedure to explore the
solution space beyond local optimality. One of the main components of Tabu search is its use of
adaptive memory, which creates a more flexible search behavior. Memory-based strategies are
therefore the hallmark of Tabu search approaches, founded on a quest for “integrating principles,”
by which alternative forms of memory are appropriately combined with effective strategies for
exploiting them.
N. Rugthaicharoencheep and S. Sirisumrannukul [23] used Tabu Search (TS) for loss reduction of
distribution networks with distributed generation. The developed methodology is tested with a 69-
bus distribution system having 48 load points. The study results indicate that for a given set of
distributed generators and their locations, the proposed method can identify optimal on/off patterns
of the switches that yield the minimum loss while satisfying the constraints.
2.2.3 Mixed Methods

These methods are in use since 1996 and are combinations of previous methods to gather their
combined strengths; hence better results are obtained. Mixed methods like linear programming
with heuristics [24] and Fuzzy Logic and heuristics [25] were used.
King and Radha [26] used a fuzzy logic controller to adapt the cross over and mutation
probabilities based on the fitness function. The main advantages of fuzzy control system over the
conventional method are: ability of modeling the quantities aspects of human knowledge and
reasoning process, model free estimator, robustness, and easy implementation. The fuzzy logic

controlled GA always finds the global optimum and has proved to have faster convergence than a
GA using fixed cross over and adaptive mutation.
Mehdi Assadian [27] investigated the ability of particle swarm optimization (PSO) in cooperation
with graph theory for network reconfiguration to reduce the power loss and voltage profile
enhancement of distribution system.

Chapter 3
Research Methodology
3.1 Problem formulation
It is becoming more and more important to power distribution companies to be able to meet
efficiently and reliably the demands of their customers. This means that one of their goals is to be
able to find an operating state for a large, three-phase, distribution network which minimizes the
loss for the power company supplying the power, while satisfying the requirements of the
customer. In developing countries like Ethiopia, the utility companies should also reconfigure and
redesign their distribution system in order to improve the reliability of their distribution system,
which is unacceptably low by many standards. This chapter states introduces some mathematical
model of network reconfiguration problem and presents a formulation of the network
reconfiguration problem of loss reduction and besides, Genetic Algorithm Optimization.
3.1.1 Mathematical Model of the Network Reconfiguration Problem
Radial distribution system reconfiguration is done by opening/closing two types of switches, tie
switches and sectionalizing switches. A feeder may be served for another feeder by closing a tie
switch linking the two while a particular sectionalizing switch must be opened to maintain radial
structures. In case of power loss reduction, the problem here to be addressed is to identify tie and
sectionalizing switches that should be closed and opened, respectively, to achieve a maximum
reduction in power losses. Theoretically, it is a straightforward matter to determine whether or
not, the new system obtained through a feeder reconfiguration would incur lower power losses.
The reduction in power losses can easily be computed from the results of two load flow studies of
the system configurations before and after the feeder reconfiguration.
Let us consider a distribution network of n nodes. The optimization problem is then finding an
optimal radial network u among all possible radial networks of search space S generated with the
switch condition changes that minimize the objective function without violation of the constraints.
a. Objective Function
Assuming that S = P + jQ, and it turns out that |S| is the total power that is transported through the
circuit component in order to get a usable active power P. We call S the apparent power. It is clear
from the definitions that a smaller phase difference results in a smaller reactive (useless) power,
hence a smaller difference between the active power and the apparent power. In electrical

engineering it is important to take into account the reactive power as much as the active power. So
we should work with the complex value S, and not just with the real value P. Now we can calculate
the power loss through a line. If I is the complex current through
a line from a to b with voltages Va, Vb, and Z = R + jX the impedance of this line, then the
apparent power loss Sloss can be calculated simply as the difference between the powers at a and b.
3.1
=( ) 3.2
= 3.3
= Z |I| 2 3.4
= R |I| 2 + j X |I| 2 3.5
Using Ohm‟s law, now the reactance X only gives information about the phase difference between
the current and the voltages due to the cable, where the resistance R induces the permeability of the
cable. Only the latter plays a role for power loss, so the actual power loss is P loss = |I| 2R. So that
the objective of the optimal feeder reconfiguration problem to minimize the total power loss can be
expressed as:
Minimize ∑ 3.6
Where, PL = total power loss

Nb = number of branches
Ii = current flow in ithbranch
Ri = Resistance of ithbranch.
b. Constraints:
The objective function in equation 1.6 is subjected to the following constraints.
i. Bus voltage limits

Bus voltage limits is well known that a small change in nodal voltage affects the flow of reactive
power whereas active power practically does not change. Further, the operating voltage at each
node must be in safety range as given below.
( )
Where,

Minimum voltage limits of ith node respectively.
Maximum voltage limits of ith node respectively.
Voltage at ith node
Number of buses.
ii. Feeder capacity limits

Power flow in each branch must be less than or equal to its maximum capacity as given below.
|Ii| ≤ i ɛ {1, 2, 3 . . ., }
Where, maximum current capacity of i branch
Ii = current in ith branch

iii. Bus isolation
All nodes or buses have to be served after reconfiguration. The node must not be isolated without
output supply from any feeder. It means only one switch should be opened in a loop.

3.2 Genetic Algorithm
Genetic Algorithm (GA) is a search-based optimization technique based on the principles
of Genetics and Natural Selection. It is frequently used to find optimal or near-optimal solutions to
difficult problems which otherwise would take a lifetime to solve. It is frequently used to solve
optimization problems, in research, and in machine learning.
One of the more challenging aspects of using genetic algorithms is to choose the configuration
parameter settings. Discussion of Genetic Algorithm (GA) theory provides little guidance of proper
selection of the settings. The population size, the mutation rate, and the type of recombination have
the largest effect on search performance. They are used to control the run of a Genetic Algorithm
(GA). They can influence the Population and the Reproduction part of the Genetic Algorithm GAs.
In traditional GAs the parameters have fixed values [28].
3.2.1 Genetic Algorithm Optimization
Genetic algorithm optimization is the process of making something better. In any process, we have
a set of inputs and a set of outputs as shown in figure 3.1.
Set of inputs Process Set of outputs
Figure 3.1 Optimization Process

Besides, optimization refers to finding the values of inputs in such a way that we get the best
output values. The definition of best varies from problem with problem, but in mathematical terms,
it refers to maximizing or minimizing one or more objective functions, by varying the input
parameters. The set of all possible solutions or values which the inputs can take make up the
search space. In this search space, lies a point or a set of points which give the optimal solution.
The aim at optimization is to find that points or set of points in the search space.
3.2.2 How Genetic Algorithms Work

Genetic algorithm maintains a population of individuals, say P(t), for generation t. Each individual
represents a potential solution to the problem at hand. Each individual is evaluated to give some

measure of its fitness. Some individuals undergo stochastic transformations by means of genetic
operations to form new individuals. There are two type of transformation:
a. Mutation, which creates new individuals by making changes in a single individual.

b. Crossover, which creates new individuals by combining parts from two individuals.
The new individuals, called offspring C (t), are then evaluated. A new population is formed by
selecting the more fit individuals from the parent population and offspring population. After
several generations, genetic algorithm converges to the best individual, which hopefully represents
an optimal or suboptimal solution to the problem.
The general structure of the Genetic algorithms is as follow:
Begin
t=0;
Initialize P(t);
Evaluate P(t);
While (not termination condition) do Begin
Apply crossover and mutation to P(t) to yield C(t);
Evaluate C(t);
Select P(t+1) from P(t) and C(t);
t=t+1;
End
End

The flowchart explains how genetic algorithms work is shown in figure 3.2.
Figure 3.2 Flow chart of Genetic Algorithm Optimization technique

3.2.3 Genetic Algorithm Operators
There are five basic genetic algorithms operators which are Population, Fitness function, Selection,
Crossover and mutation.
A. Population
The process begins with a set of individuals which is called a Population. Each individual is a
solution to the problem you want to solve. An individual is characterized by a set of parameters
(variables) known as Genes. Genes are joined into a string to form a Chromosome (solution).
In a genetic algorithm, the set of genes of an individual is represented using a string, in terms of an
alphabet. Usually, binary values are used (string of 1s and 0s). We say that we encode the genes in a
chromosome, as shown in fig 3.3.
Figure 3.3 Population, Chromosomes and Genes

The population size is one of the most important parameters that play a significant role in the
performance of the genetic algorithms. The population size dictates the number of individuals in
the population. Larger population sizes increase the amount of variation present in the initial
population at the expense of requiring more fitness evaluations. It is found that the best population
size is both applications dependent and related to the individual size (number of chromosomes
within). A good population of individuals contains a diverse selection of potential building blocks
resulting in better exploration. If the population loses diversity the population is said to have
“premature convergence” and little exploration is being done. For larger individuals and
challenging optimization problems, larger population sizes are needed to maintain diversity (higher

diversity can also be achieved through higher mutation rates and uniform crossover) and hence
better exploration. Many researchers suggest population sizes between 25 and 100 individual, while
others suggest that it must be very much larger (1000 individual or more).
B. Fitness Function
The fitness function determines how fit an individual is (the ability of an individual to compete
with other individuals). It gives a fitness value to each individual. The probability that an individual
will be selected for reproduction is based on its fitness value.
C. Selection
The idea of selection phase is to select the fittest individuals and let them pass their genes to the
next generation. Two pairs of individuals (parents) are selected based on their fitness values [36].
Individuals with high fitness have more chance to be selected for reproduction.
Selection is the process of determining the number of times a particular individual is chosen for
reproduction and, thus, the number of offspring that an individual will produce. The principle of
genetic algorithms is essentially Darwinian natural selection. Selection provides the driving force in
genetic algorithms. With too much force, genetic search will terminate prematurely. While with too
little force, evolutionary progress will be slower than necessary.
Typically, a lower selection pressure is indicated at the start of genetic search in favor of a wide
exploration of the search space, while a higher selection pressure is recommended to the end to
narrow the search space. In this way, the selection directs the genetic search for promising regions
in the search space and that will improve the performance of genetic algorithms. Many selection
methods have been proposed, examined and compared. The most common types are:
i. Roulette wheels selection iv. Steady state selection

ii. Rank selection v. Elitism
iii. Tournament selection
i. Roulette Wheel Selection
Roulette wheel selection is the commonest selection method used in genetic algorithms for
selecting potentially useful individuals (solutions) for crossover and mutation.
In roulette wheel selection, as in all selection methods, possible solutions are assigned fitness by
the fitness function. This fitness level is used to associate a probability of selection of each

individual. While candidate solutions to a higher fitness will be less likely to be eliminated, there is
still a chance that they may be. With roulette wheel selection there is a chance some weaker
solutions may survive the selection process; this is an advantage, as though a solution may be weak,
it may include some component which could prove useful following the recombination process.
The analogy between a roulette wheel can be envisaged by imagining a roulette wheel in which
each candidate solution represents a pocket on the wheel; the size of the pockets is proportionate to
the probability of selection of the solution. Selecting N individual from the population is equivalent
to playing N games on the roulette wheel, as each candidate is drawn independently, as shown in
figure 3.4.
Figure 3.4 Roulette wheel selections. [29], [30]
ii. Rank Selection
In ranking selection, the individuals in the population are sorted through best to worst according to
their fitness values. Each individual in the population is assigned a numerical rank based on fitness,
and selection is based on this ranking rather than differences in fitness. The advantage of this
method is that it can prevent very fit individuals from gaining dominance early at the expense of
less fit ones, which would reduce the population's genetic diversity and might hinder attempts to
find an acceptable solution. The disadvantage of this method is that it required sorting the entire
population by rank which is a potentially time consuming procedure. Rank selection effect is
shown in figure 3.5 (a and b).

Figure 3.5 Rank selection effects. (a) Before ranking. (b) After ranking [31], [32], [33], [34]
iii. Tournament Selection
This method randomly chooses a set of individual and picks out the best individual for
reproduction. The number of individual in the set is called the tournament size. A common
tournament size is two, this is called binary tournament. By adjusting tournament size, the selection
pressure can be made arbitrarily large or small. For example, using large Tournament size has the
effect of increasing the selection pressure, since below average individuals are less likely to win a
tournament while above average individuals are more likely to win it [35].
iv. Steady State Selection

The steady state selection will eliminate the worst of individuals in each generation [41]. It works;
the offspring of the individuals selected from each generation go back into the pre-existing
population, replacing some of the less fit members of the previous generation.

v. Elitism
Elitism is an addition to many selection methods that force genetic algorithms to retain some
number of the best individual at each generation. It improves the selection process and save the best
individuals. With elitist selection, the quality of the best solution in each generation monotonically
increases from time. Without elitist selection, it is possible to lose the best individuals due to
stochastic errors (due to crossover, mutation or selection pressure) [42].
D. Crossover
Crossover is the most significant phase in a genetic algorithm. For each pair of parents to be mated,
a crossover point is chosen at random from within the genes. For example, consider the crossover
point to be three, as shown below in figure 3.6.
A1 0 0 0 0 0 0
A2 1 1 1 1 1 1
Crossover point
Figure 3.6 Crossover point
Offspring are created by exchanging the genes of parents among themselves until the crossover
point is reached, as shown in figure 3.7. And the new offspring are added to the population, as
shown in figure 3.8.
A3 0 0 0 0 0 0
A4 1 1 1 1 1 1
Crossover point
Figure 3.7 Exchanging genes among parents

A5 1 1 1 0 0 0
A6 0 0 0 1 1 1
Figure 3.8 New offspring

One of the unique aspects of the work involving genetic algorithms (GAs) is the important role that
Crossover (recombination) plays in the design and implementation of robust evolutionary systems.
In most GAs, individuals are represented by fixed-length strings and crossover operates on pairs of
individuals (parents) to produce new strings (offspring) by exchanging segments from the parents‟
strings. Crossover rate determines the probability that crossover will occur. The crossover will
generate new individuals in the population by combining parts of existing individuals. The
crossover rate is usually high and „application dependent‟. Many researchers suggest crossover rate
to be between 0.6 and 0.95. [37], [38], [39], [40]
vi. Single Point Crossover
A commonly used method of crossover is called single point crossover. In this method, a single
point crossover position (called cut point) is chosen at random (e.g., between the 10th and 5th
point) and the parts of two parents after the crossover position are exchanged to form two offspring,
as shown in figure 3.9. Cross point
1 0 0 0 1 1 1 1 0 0 1 1 1 0 0 Parent 1
0 0 1 1 1 0 0 1 1 0 1 0 1 1 0 Parent2
Cross point
1 0 0 0 1 1 1 1 0 0 1 0 1 1 0 offspring1
0 0 1 1 1 0 0 1 1 0 1 1 1 0 0 Offspring 2
Figure 3.9 Single Point Crossover

vii. Multi Point Crossover
Multi-point crossover is a generalization of single point crossover, introducing a higher number of
cut-points. In this case multi positions are chosen at random and the segments of them are
exchanged, as shown in figure 3.10.
Cross point Cross point
1 0 0 0 1 1 1 1 0 0 1 1 1 0 0 Parent 1
0 0 1 1 1 0 0 1 1 0 1 0 1 1 0 Parent2
Cross point Cross point
1 0 1 1 1 0 0 1 1 0 1 0 1 1 0 offspring1
0 0 0 0 1 1 1 1 0 0 1 1 1 0 0 Offspring 2
Figure 3.10 Multi point crossover

viii. Uniform Crossover
Uniform crossover does not use cut-points, but simply uses a global parameter to indicate the
likelihood that each variable should be exchanged with two parents, as shown in figure 3.11.
1 0 0 0 1 1 1 1 0 0 1 1 1 0 0 Parent 1
0 0 1 1 1 0 0 1 1 0 1 0 1 1 0 Parent2
1 0 1 1 1 1 1 1 1 0 1 0 1 1 0 offspring1
0 0 0 0 1 1 1 1 0 0 1 1 1 0 0 Offspring 2
Figure 3.11 Uniform crossovers

E. Mutation
Mutation rate determines the probability that a mutation will occur. Mutation is employed to give
new information about the population (uncover new chromosomes) and also prevents the
population of becoming saturated with similar chromosomes, simply said to avoid premature
convergence. Large mutation rates increase the probability that good schemata will be destroyed,
but increase population diversity. The best mutation rate is „application dependent‟. For most
applications, mutation rate is between 0.001 and 0.1[43].
In certain new offspring formed, some of their genes can be subjected to a mutation with a low
random probability. This implies that some of the bits in the bit string can be flipped, as shown in
fig 3.12.
A5 1 1 1 0 0 0
Before mutation
A5 1 1 0 1 1 0
After Mutation
Figure 3.12 Mutation: Before and After
i. Single Point Mutation

Single gene (chromosome or even individual) is randomly selected to be mutated and its value is
changed depending on the encoding type used, as shown in figure 3.13.
1 0 0 0 1 1 1 1 0 0 1 1 1 0 0 Parent
1 0 0 0 1 1 1 1 0 0 1 1 1 0 0 Offspring
Figure 3.13 Single point mutation

ii. Multi Point Mutation
Multi genes (chromosomes or even individuals) are randomly selected to be mutated and there
values are changed depending on the encoding type used, as shown in figure 3.14.
1 0 0 0 1 1 1 1 0 0 1 1 1 0 0 Parent
1 0 0 0 1 1 1 1 0 0 1 1 1 0 0 Offspring
Figure 3.14 Multi point mutation
3.2. 4 Encoding
How to encode the solutions to the problem of chromosomes is a key issue when using genetic
algorithms. One outstanding problem associated with encoding is that some individuals correspond
to infeasible or illegal solutions to a given problem. This may become very severe for constrained
optimization problems and combinatorial optimization problems.
Various encoding methods have been created for particular problems to provide effective
implementation of genetic algorithms. According to what kind of symbol is used as the alleles of a
gene, the encoding methods can be classified as follows:
i. Binary encoding
ii. Real-number encoding
iii. Integer or literal permutation encoding
i. Binary Encoding
Binary encoding (i.e., the bit strings) are the most common encoding used for several of reasons.
One is historical: in their earlier work, Holland and his students concentrated on such encodings
and genetic algorithms practices have tended to follow this lead. Another reason for that was
because much of existing GAs theories are based on the assumption of using binary encoding. In
spite of all that, binary encoding for function optimization problems is known to severe drawbacks
due to the existence of Hamming cliffs, pairs of encoding having a large Hamming distance (The
Hamming distance between two bit strings is defined as the number of corresponding positions in

these bit strings where the bits have a different value) while belonging to points of minimal
distance in phenotype space.
The probability that crossover and mutation will occur can be very small. In this sense, the binary
code doses not preserve the locality of points in the phenotype space. For many problems in the
industrial engineering world, it is nearly impossible to represent their solution with binary
encoding.
ii. Real Number Encoding

Real number encoding is best used for function optimization problems. It has been widely
confirmed that real number encoding perform better than binary encoding for function optimization
and constrained optimizations problems. In real number encoding, the structure of genotype space
is identical to that of the phenotype. Therefore, it is easy to form effective genetic operators by
borrowing useful techniques from conventional methods.
iii. Integer or Literal Permutation Encoding

Integer or literal permutation encoding is best used for combinational optimization Problems
because the essence of this kind of problems is to search for the best permutation or combination of
items subject to constrains.
3.3 DIgSILENT PowerFactory Overview
DIgSILENT power factory software is a powerful commercial program which converges on

solving distribution system load flows in microseconds even though it uses full Newton-Raphson
method.
The calculation program PowerFactory, as written by DIgSILENT, is a computer aided engineering
tool for the analysis of transmission, distribution, and industrial electrical power systems. It has
been designed as an advanced integrated and interactive software package dedicated to electrical
power system and control analysis in order to achieve the main objectives of planning and
operation optimization. “DIgSILENT" is an acronym for “DIgital SImuLation of Electrical
NeTworks". DIgSILENT Version 7 was the world‟s first power system analysis software with an
integrated graphical single-line interface. That interactive single-line diagram included drawing
functions, editing capabilities and all relevant static and dynamic calculation features.
PowerFactorywas designed and developed by qualified engineers and programmers with many
years of experience in both electrical power system analysis and programming fields. The accuracy

and validity of results obtained with PowerFactoryhas been confirmed in a large number of
implementations, by organizations involved in planning and operation of power systems throughout
the world. To address users power system analysis requirements, PowerFactorywas designed as an
integrated engineering tool to provide a comprehensive suite of power system analysis functions
within a single executable program. Key features include:
1. PowerFactory core functions: definition, modification and organization of cases; core numerical
routines; output and documentation functions.
2. Integrated interactive single line graphic and data case handling.
3. Power system element and base case database.
4. Integrated calculation functions (e.g. line and machine parameter calculation based on
geometrical or nameplate information).
5. Power system network configuration with interactive or on-line SCADA access.
6. Generic interface for computer-based mapping systems.
Use of a single database, with the required data for all equipment within a power system (e.g. line
data, generator data, protection data, harmonic data, controller data), means that PowerFactory can
easily execute all power simulation functions within a single program environment - functions such
as load-flow, short-circuit calculation, harmonic analysis, protection coordination, stability
calculation, and modal analysis. Although PowerFactory includes some sophisticated power system
analysis functions, the intuitive user interface makes it possible for new users to very quickly
perform common activities such as load-flow and short-circuit calculations.
In this thesis DIgSILENT PowerFactory, is used for simulating the distribution network
reconfiguration system.

Chapter 4
Simulation Studies and Analysis of Results
This chapter discusses the overall simulation results obtained from DIgSILENT PowerFactory on
the distribution system of Addis North 132/15 KV substation and Matlab for Genetic Algorithm
optimization. The overall Addis North 132/15 KV substation distribution system simulation
modeling is represented in the DIgSILENT PowerFactory as shown in figures A-1 in the
appendix.
4.1 Addis North 132/15 KV Substation Test Systems
4.1.1 Addis North 132/15 KV Substation System Description
The proposed network reconfiguration has been computed in Addis North 132/15 KV substation,
Addis Ababa, which comprises six feeders as shown in Appendix-A before reconfiguration and
Appendix-B after reconfiguration and For Genetic Algorithm optimization result the Matlab
algorithm is shown in Appendix-D. The DIgSILENT PowerFactory simulation result is shown from
Table 4-1 to Table 4- 6 and Genetic Algorithm optimization result is shown from Table 4-7 to Table 4-
9. The system is a three phase system, 15 KV and the input data is shown in table Appendix-C, which
is obtained from Ethiopia Electric Utility (EEU). The following conductors are used in the feeder.
1. Underground cables used for each feeder are listed below, which is obtained from Addis North
132/15 KV Substation.
 Feeder 1: 240 X 2 sqmm AC  Feeder 4: 120/ 3φ/ sqmm Cu

 Feeder 2: 240 X 2 sqmm AC  Feeder 5: 300 X 1 sqmm Cu
 Feeder 3: 240 X 2 sqmm AC  Feeder 6: 240 X 2 sqmm AC
2. Overhead lines used for all feeders, which is obtained from Ethiopia Electric Utility (EEU).
 AAC 50 sqmm, AAC 25 sqmm, 50 ACSR sqmm
3. The proposed overhead lines for all feeders
 AAAC 240 sqmm , AAC 95 sqmm, AAC 50 sqmm
4.1.2 Addis North 132/15 KV Substation Simulation Result
a. Load flow Calculation

 “External Grid” is local reference in separated area of bus “B”.

 Start Newton – Raphson algorithm . . .
o Load flow iteration: 1
o Load flow iteration: 2
 Newton – Raphson converged with 2 iterations.
 Load flow calculation successful.
b. Feeder output calculation analysis before optimization
In table 4 -1, the six feeders input current and total load have been obtained before optimization of
the distribution network. As shown in the table 4-1, the total power loss is 3.983783 MW.
Name Feeder Input Total Generation Losses Max.Loadin Mini.Voltage

Current Load g
[MW] [MW] [p.u.]
[kA] [MW] [%]
BF_F1 F1 0.193 4.960 0.000 0.035 92.79 0.986
BF_F2 F2 0.589 13.660 0.000 1.550 107.07 0.845
BF_F3 F3 0.486 11.310 0.000 1.232 106.24 0.850
BF_F4 F4 0.115 2.980 0.000 0.007 95.40 0.993
BF_F5 F5 0.127 3.280 0.000 0.013 96.39 0.991
BF_F6 F6 0.503 11.850 0.000 1.146 105.87 00.858
Table 4-1 Feeder output calculation analysis before optimization

Figure 4.1 Feeders output calculation analysis chart before optimization
c. Tie Open Point Optimization

Objective Function Minimization of Losses
Consider Thermal Constraints YES
Global constraint for all components YES
Maximum thermal loading of components 100.0 %
Consider Voltage Constraints YES
Constraints Global constraint for all terminals YES
Lower voltage limit 0.95 p.u.
Upper voltage limit 1.05 p.u.
Consider Voltage Drop/Rise Constraints NO
Ignore all constraints for nominal voltage <= 1.00 kV
Table 4-2 Tie Open Point Optimization

d. Optimal Tie Open Positions
Name Feeder 1 Feeder 2
Switch/Line033 BF_F3 BF_F4
Switch/Line(32) BF_F4
Switch/Line(27) BF_F5
Switch/Line033 BF_F4
Table 4-3 Optimal Tie Open Positions
e. Necessary Switching Actions (From Initial to Optimized Configuration)

Name Action
Switch/Line033 OPEN
Switch/Line028 OPEN
Switch/Line37 OPEN
Switch/Line CLOSE
Switch/Line CLOSE
Switch/Line CLOSE
Table 4-4 Necessary Switching Actions (From Initial to Optimized Configuration)

f. Feeders Results Before and After Reconfiguration
The detail description of the feeders‟ result of loss, total load, number of customer, maximum
voltage drop, maximum voltage rise, minimum and maximum voltages before optimization, after
optimization and their difference is given in the table below.
Feeder Losses Total Load Number of Maximum Maximu Minimum Maximum

m
[MW] [MW] Customers Voltage Voltage Voltage
Voltage
Drop [%] [p.u.] [p.u.]
Rise [%]
Before 0.035081 4.960000 10.000000 0.977528 0.000000 0.986369 0.994853
Optimization
After 0.333706 9.673195 28.000000 5.674761 0.000000 0.941041 0.995182

BF_F1
Optimization
Difference 0.298625 4.713195 18.000000 4.697234 0.000000 -0.045328 0.000328
Before 1.550367 13.660318 46.000000 15.117099 0.000000 0.844974 0.992069

Optimization
After 0.385894 8.946123 28.000000 5.872831 0.000000 0.939061 0.995350

BF_F2
Optimization
Difference -1.164473 -4.714195 -18.000000 -9.244267 0.000000 0.094087 0.003281
Before 1.232391 11.309954 59.000000 14.645229 0.000000 0.849692 0.992789

Optimization
After 0.218436 6.733634 33.000000 4.821054 0.000000 0.949578 0.995979

BF_F3
Optimization
Difference -1.013956 -4.576319 -26.000000 -9.824175 0.000000 0.099886 0.003190
Before 0.007463 2.980000 6.000000 0.352639 0.000000 0.992618 0.995374

Optimization
After 0.224793 7.555761 32.000000 5.050102 0.000000 0.947288 0.995765

BF_F4
Optimization
Difference 0.217330 4.575761 26.000000 4.697463 0.000000 -0.045330 0.000391

Before 0.012796 3.280000 8.000000 0.516859 0.000000 0.990976 0.995294
Optimization
After 0.220597 7.752629 29.000000 4.470545 0.000000 0.953083 0.995716

BF_F5
Optimization
Difference 0.207801 4.472629 21.000000 3.953686 0.000000 -0.037893 0.000421
Before 1.145686 11.849606 58.000000 13.767013 0.000000 0.858475 0.992681

Optimization
After 0.211215 7.376473 37.000000 4.480650 0.000000 0.952982 0.995816

BF_F6
Optimization
Difference -0.934471 -4.473133 -21.000000 -9.286363 0.000000 0.094508 0.003135
Before 3.983783 48.039878 187.000000 0.352639 0.000000 0.992618 0.992069

Optimization
After 1.594640 48.037816 187.000000 5.050102 0.000000 0.947288 0.995350

TOTAL
Optimization
Difference 2.389143 -0.002062 0.000000 4.697463 0.000000 -0.045330 0.003281
Difference[%] 59.97172 -0.004292 0.000000 1332.08644 100.0000 4.566754 0.330711
Table 4-5 Feeders Results

The distribution network is reconfigured for the real power loss minimization. As can be seen from
Table 4-5, the loss after reconfiguration is 1.594640 MW and the difference is 2.389143 MW. The
power loss before reconfiguration is 3.983783 MW. Therefore the reduction in power loss is Power
loss reduction=3.983783 MW - 2.389143 MW = 1.594640 MW, which is approximately 59.97172
% of the loss after reconfiguration is reduced. High percentage of reduction in power loss obtained
because the proposed tie open point switches are strategic ones and their overhead lines (cables)
cross sections deliberately selected to be higher ones.
The total energy loss before and after optimization in a year

Before reconfiguration total power loss = 3.983783 MW

 Before reconfiguration total energy loss for one year = 365 X 3.983783 MW X 24h
= 34, 897, 939.08 KWh
After optimization, can be saved = 2.389143 MW
 After reconfiguration total energy, can be saved for one year = 365 X 2.389143
MW X 24h
= 20,928,892.68 KWh
g. Feeder output calculation analysis after optimization shown table below
In the table below, the 6 feeders input current and total load have been obtained after optimization
of the distribution network. As shown in the table below, the total power loss is 1.594640.
Name Feeder Input Total Generation Losses Max.Loading Mini.Voltage

Current Load
[MW] [MW] [%] [p.u.]
[kA] [MW]
BF_F1 F1 0.386 9.673 0.000 0.334 185.59 0.941
BF_F2 F2 0.360 8.946 0.000 0.386 65.45 0.939
BF_F3 F3 0.268 6.734 0.000 0.218 58.68 0.950
BF_F4 F4 0.300 7.556 0.000 0.225 248.05 0.947
BF_F5 F5 0.308 7.753 0.000 0.221 233.01 0.953
BF_F6 F6 0.293 7.376 0.000 0.211 61.62 0.953
Table 4-6 Feeder output calculation analysis after optimization shown table below

12
10
8
Input Current [kA]
6
Total Load [MW]
4 Losses [MW]
0
F1 F2 F3 F4 F5 F6
Figure 4.2 Feeders output calculation analyses chart after optimization
4.2 Genetic Algorithm optimization result

 Program to optimize the power for feeder 1 and feeder 2
o Enter the size of initial population::30
o Enter the bit size of each string::7
o Enter the number of iterations to perform the process of genetic algorithm::1000
o Enter the number of times you want to perform tournament selection::50
With the parameters chosen as above, the best results were achieved. A higher number of iterations
always results in better output. The test results achieved when using the above parameter values can
be found in Table 4-7, for 10 iterations and for 1000 iterations. For this test result the Matlab
algorithm is shown in Appendix D.
Power before Power optimization Power optimization after

optimization after 10 iterations 1000 iterations
P_1 for feeder 1 (MW) 4.9600 8.8204 9.1174
P_2 for feeder 2 (MW) 13.6600 9.7996 9.5026
Total Power 18.6200 18.6200 18.6200
No of switches in feeder 1 (u) 10 23 24

No of switches in feeder 2 (V) 46 33 32
Total No of Switches (u + v) 56 56 56
Table 4-7 Genetic algorithm test results for feeder 1 and feeder 2

Power before Power optimization Power optimization

optimization after 10 iterations after 100 iterations
P_3 for feeder 3 (MW) 11.3100 6.9010 7.0927
P_4 for feeder 4 (MW) 2.9800 7.3890 7.1973
Total Power 14.29 14.29 14.29

Power before Power optimization Optimization after

optimization after 10 iterations 1000 iterations
Power Loss
P_5 for feeder 5 (MW) 3.2800 6.9576 7.5705 0.214
P_6 for feeder 6 (MW) 11.8500 8.1724 7.5595 0.216
Total Power 15.1300 15.1300 15.1300 0.43

4.3 Feasibility Analysis
After reconfiguration total energy, can be saved for one year is 20,928.89268 MWh
By using current average electric sale bill of EEU, it can be converting in ETB.
Average electric sale bill = 0.655 cent per KWh
Hence, 20,928,892.68 X 0.655 = 13,708,424.7054 ETB = $498,125.889435 per year
Tie and Sectionalizing switch specification
 Price: $5.2/piece
 Minimum order: 10 piece
 Operating capacity: Medium voltage
 Operation: automatic type
 Total number of distribution network switches : 187
 $5.2 X 187 = $972.4
Assuming that the installation costs are 10% and annual maintenance costs are 2%.
 10% of $498,125.889435 = $49,812.6
 2% of $498,125.889435 = $9,962.52
 Total cost = $972.4 + $49,812.6 + $9,962.52 = $60,747.52
 difference $498,125.889435-$60,747.52 = $437,378.369435

Chapter 5
Conclusions, Recommendations and Future Work
5.1 Conclusions
Network reconfiguration means restructuring the power lines which connect various buses in a
power system. Restructuring of specific lines leads to alternate system configurations. System
reconfiguration can be accomplished by placing line interconnection switches into network.
Opening and closing a switch connects or disconnect a line to the existing network.
The main objective of this thesis is to minimize the electrical power loss by considering network
reconfiguration and applying a Genetic Algorithm in an electric distribution power system in the case of Addis
North 132/15 KV substation, Addis Ababa. Obviously, Addis Ababa is experiencing frequent power
interruptions in recent times as the distribution system is not designed and built to handle the current load
demands. The power interruption can be minimized by using network reconfiguration.
As can be referred from the simulation and result discussion of chapter 4, the test system has been
computed in Addis North 132/15 KV substation distribution network. The network reconfiguration
has been simulated using DIgSILENT PowerFactory and Genetic Algorithm has been optimized
using Matlab algorithm. An optimization problem is a maximization or minimization problem
which involves finding the best solution out of a set of possible alternatives. It can be completely
characterized by the search space and objective function. The search space is a finite or countable
infinite set of possible solutions, and the objective function maps each point in the search space into
the real line, to give a measure of how good a solution is relative to the others.
After network reconfiguration for loss minimization, the power loss has been reduced, which is
59.91716% reduction. Hence, reconfiguration of distribution networks can reduce power loss and
operating cost as well as improves the voltage profile of distribution systems.
5.2 Recommendations
The following are the recommendations based on the study on power loss minimization for network
reconfiguration.
1. From this thesis reconfiguration of distribution network has shown to decrease the power loss
significantly. Therefore, it is highly recommended if all the distribution network feeders of Addis
Ababa city to be reconfigured for the objectives wanted by the Ethiopian Eclectic Utility
(EEU).

2. Most of the conductors used on the distribution network are 25 sq mm AAC, 50 sq mm AAC and
50 sq mm ACSR conductors‟ cross-section areas. In this case, some lines are loaded. Therefore,
it is better to use AAAC 240 sqmm, AAC 95 sqmm, AAC 50 sqmm or above conductors
cross-sectional areas, which is recommended for main lines of cities like Addis North 132/15
KV Substation, Addis Ababa, based on the loads connected. And the existing Addis Ababa
distribution network is not designed and built according to international standards. Large cross-
sections area conductors are used after small cross-section conductors on main lines and also on
branch lines of the distribution network. In order to useful capacity of the large cross-section of
conductors, it should be built before the small cross-sectional area conductors.
3. Most of the feeders network of the Addis Ababa city have no tie switches to a feeder, even
though there are tie switches to feeders. Since these tie switches are very useful for network
reconfiguration as shown by this thesis and has also other advantages, it is highly
recommended if the distribution feeders have these tie switches.
5.3 Future Work

In the future researches can be done on the distribution networks of Addis Ababa city including the
following suggestions
1. The simulation of the distribution network can also be used for optimization for capacitor
placement. To reduce real power losses, improve the voltage profile and improve the load
balancing of the network.
2. The simulation of the distribution network can be made to include strategically placed
distribution generators (DG) and the network can be reconfigured. In doing so the voltage
profiles, the real power losses and the system load balancing index can be improved.
3. Other optimization algorithms, some of which are described in literature review, can also be
used to improve the reconfiguration of the distribution network and find global optimum
solutions.

References
[1] K. Nara, A. Shiose, M. Kitagawa, T. Ishihara,“Implementation of genetic algorithm for
distribution systems loss minimum re-configuration”, IEEE Transactions on Power systems,
Vol.7, No 3, August 1992, pp. 1044-1051.
[2] K. Nara, A. Shioss, M. Kitagwa and T. Ishihwara, “Implementation of GA for distribution

system loss minimum reconfiguration”, IEEE Transaction on Power Systems, Vol. 7, No. 3,
1992, pp. 1044-1051.
[3] H. Fudou, T. Genji, Y. Fukuyamam and Y. Nakanishi, “A genetic algorithm for network
reconfiguration using three unbalanced load flow”, Intelligent System Applications to Power
System (ISAP‟97), Seoul, Korea, 1997, pp. 1-5.
[4] Ray Daniel Zimmerman, Network Reconfiguration for Loss Minimization in Three Phase
Power Distribution System, Cornel University, May 1992
[5] Ali Reza Fereidunian, Hamid Lesani and Caro Lucas “Distribution systems reconfiguration
using pattern recognizer neural networks” IJE International: Applications Vol. 15, No. 2, July
2002 – 135.
[6] H. Kim, “ANN based feeder reconfiguration for loss reduction in distribution system”, IEEE
Transaction on Power Delivery, Vol. 8,No. 3, 1993, pp. 1356-1366.
[7] Tamer M. Khalil, and Alexander V. Gorpinich, “Reconfiguration for Loss Reduction of
Distribution Systems Using Selective Particle Swarm Optimization” International Journal Of
Multidisciplinary Sciences And Engineering, Vol. 3, No. 6, June 2012.
[8] A. Merlin, H. Back, “Search for a minimal-loss operating spanning tree configuration in an
urban power distribution system” Proceedings of 5th Power System Computation Conference
(PSCC), Cambridge, UK, (1975), pp. 1-18.
[9] E.Dolatdar, S.Soleymani, B.Mozafari “A New Distribution Network Reconfiguration Approach

Using a Tree Model”, World Academy of Science, Engineering and Technology Vol: 3 2009-
10-26.

[10] Hugh Rudnick , Ildefonso Harnisch , Raúl Sanhueza , “Reconfiguration of Electric Distribution
Systems, Revista Facultad De Ingenieria”, U.T.A. (Chile), Vol. 4, 1997.
[11] Abdullah M. Alshehri,“Optimal Reconfiguration of Distribution Networks Using Ant

Colony Method”, King Saud University, June, 2007.
[12] D. Shirmohammadi and H.W Hong,“Reconfiguration of electric distribution networks for

resistive line losses reduction”, IEEE Transactions on Power Delivery, Vol. 4, No. 2, 1989,
pp. 1492-1498
[13] S. Civanlar, J.J. Grainger, H. Yin, S.S. Lee, “Distribution Feeder Reconfiguration for Loss
Reduction”, IEEE Transactions on Power Delivery, Vol. 3, No. 3, July 1988, pp. 1217-1223.
[14] M.E. Baran, F.F. Wu, “Network Reconfiguration in Distribution Systems for Loss Reduction
and Load Balancing”, IEEE Transactions on Power Delivery, Vol. 4, No. 2, April 1989, pp.
1401-1407.
[15] T. Taylor, D. Lubkeman, “Implementation of Heuristic Search Strategies for Distribution

Feeder Reconfiguration”, IEEE Transactions on Power Delivery, Vol. 5, No. 1, January 1990,
pp. 239-246.
[16] G.J Peponis, M.P. Popadopoulos and N.D Hatziargyrion, “Distribution network
reconfiguration for resistive line losses”, IEEE Transactions on Power Delivery, Vol. 10,
No. 3, 1995, pp. 1338-1342.
[17] R.P Broadwater, A.H. Khan, H.E Shalaan and R.E. Lee, “Time varying load analysis to
reduce distribution losses through reconfiguration”, IEEE Transactions on Power Delivery,
Vol. 8, No. 1, 1993, pp. 294-300.
[18] R. Safri, M.M.A Salama and Y. Chickani, “Distribution system reconfiguration for loss
reduction: an algorithm based on network portioning theory”,a IEEE Transactions on Power
System, Vol. 11, No. 1, 1996, pp. 504-510.

[19] A. Borghetti, M. Paolone and C.A. Nucci “ A mixed integer linear programming approach to
the Optimal configuration of electrical distribution Networks with embedded generators” 17th
Power Systems Computation Conference, Stockholm Sweden - August 22-26, 2011.
[20] G. Celli, M. Loddo, F. Pilo and A. Abur ” On-line network reconfiguration for loss reduction
in distribution networks with distributed generation” 18th International Conference on
Electricity Distribution, Turin, 6-9 June 2005.
[21] Ramadoni Syahputra ” Fuzzy multi-objective approach for the Improvement of distribution
network efficiency by considering DG” International Journal of Computer Science &
Information Technology (IJCSIT) Vol 4, No 2, April 2012.
[22] Abhisek Ukil and Willy Siti “Feeder Load Balancing using Fuzzy Logic and Combinatorial
Optimization-based Implementation”
[23] N. Rugthaicharoencheep and S. Sirisumrannukul “Feeder Reconfiguration for Loss Reduction

in Distribution System with Distributed Generators by Tabu Search”, GMSARN International
Journal 3 (2009), pp 47 – 54.
[24] T.P. Wagner, A.Y. Chikhani, R. Hackam, “Feeder Reconfiguration for Loss Reduction: An
Application of Distribution Automation”, IEEE Transactions on Power Delivery, Vol. 6, No.
4, July 1991, pp. 1922-1931.
[25] Q. Zhou, D. Shirmohammadi, W. H. Liu, ”Distribution Feeder Reconfiguration for Service

Restoration and Load Balancing”, IEEE Transactions on Power Systems, Vol. 12, No. 2, May
1997, pp. 724-729.
[26] R.T.F Ah King, B. Radha and H.C.S Rughooputh, “A fuzzy logic controlled GA for optimal
electric distribution network reconfiguration”, Proc.IEEE, International Confernce on
Networking Sensing and Control, Taipei, Taiwan, 2004, pp. 577-582.
[27] Mehdi Assadian, Malihe M. Farsangi , Hossein Nezamabadi-pour ” Optimal

Reconfiguration of Distribution System by PSO and GA using graph theory” Proceedings of
the 6th WSEAS International Conference on Applications of Electrical Engineering, Istanbul,
Turkey, May 27-29, 2007.

[28] D. E. Goldberg, Genetic Algorithms in Search, Optimization and Machine
Learning, Addison Wesley Publishing Company, January 1989.
[29] Bäck, Thomas, Evolutionary Algorithms in Theory and Practice (1996), p. 120, Oxford
Univ. Press
[30] A. Lipowski, Roulette-wheel selection via stochastic acceptance (arXiv:1109.3627)
[31] Cedeno, W. and V. Vemuri (1994). Multi-niche crowding in genetic algorithms and its
application to the assembly of DNA restriction fragments, Evolutionary Computation, Vol.
2, No. 4, pp 321-345. MIT Press.
[32] Cooper, M.G. (1994). Genetic Design of Rule-Based Fuzzy Controllers. PhD dissertation,
Department of Computer Science, University of California, Los Angeles.
[33] DeJong, K.A. (1975). Analysis of the Behavior of a Class of Genetic Adaptive Systems.
PhD dissertation, Department of Computer and Communication Sciences, University of
Michigan.
[34] Goldberg, D.E. (1989). Genetic Algorithms in Search, Optimization, and Machine Learning.
Addison-Wesley.
[35] D. Whitley, “The GENITOR algorithm and selection pressure: why rank based allocations
of reproductive trials is best”, Proc. ICGA 3, pp. 116-121, 1989.
[36] J. E. Baker, “Adaptive Selection Methods for Genetic Algorithms”, Proc.

ICGA 1, pp. 101-111, 1985.
[37] G. Syswerda, “Uniform crossover in genetic algorithms”, Proc. ICGA 3, pp. 2-9, 1989.
[38] W. M. Spears and K. A. De Jong, “On the Virtues of Parameterised Uniform Crossover”,
Proc. ICGA 4, pp.230-236, 1991.
[39] G. Syswerda, “Uniform crossover in genetic algorithms”, Proc. ICGA 3, pp. 2-9, 1989.
[40] W. M. Spears and K. A. De Jong, “An Analysis of Multi-Point Crossover”, In Foundations

of Genetic Algorithms, J. E. Rawlins (Ed.), pp. 301-315, 1991.

[41] J. E. Baker, “Reducing bias and inefficiency in the selection algorithm”, Proc. ICGA 2, pp.
14-21, Lawrence Erlbaum Associates, Publishers, 1987.
[42] R. A. Caruana, L. A. Eshelman, J. D. Schaffer, “Representation and hidden bias II:

Eliminating defining length bias in genetic search via shuffle crossover”, In
Eleventh International Joint Conference on Artificial Intelligence, N. S. Sridharan (Ed.),
Vol. 1, pp. 750 755, Morgan Kaufmann Publishers, 1989.
[43] M. F. Bramlette, “Initialization, Mutation and Selection Methods in Genetic Algorithms for
Function Optimization”, Proc ICGA 4, pp. 100-107, 1991.

Appendix A: Addis North 132/15 kV Substation feeders before reconfiguration
Figure A.1 Addis North 132/15 kV substation outgoing feeders online diagram

Figure A.2 Addis North 132/15 kV substation feeder 1 one line diagram





Figure A.7 Addis north 132/15 kV substation feeder 6 one line diagram

Appendix B: Addis North 132/15 kV Substation feeders network after
reconfiguration

Figure A.1 Addis North 132/15 kV Substation feeders network after reconfiguration

Appendix C: Addis North 132/15 kV Substation feeders data
This appendix consists of tables for Addis North 132/15 KV Substation feeders data, which is
obtained from Ethiopia Electric Utility (EEU).
Load Load Bus Bus R X
No. (MW) (From) (To) (ohm) (ohm)
01 0.25 BF BF1 0.0578 0.036
02 0.25 BF1 2 0.085544 0.05328
03 1.00 2 3 0.085544 0.05328
04 0.49 3 4 0.083232 0.05184
05 0.49 4 5 0.087856 0.05472
06 1.00 5 6 0.087856 0.05472
07 0.25 6 7 0.087856 0.05472
08 0.49 7 8 0.083232 0.05184
09 0.25 8 9 0.083232 0.05184
10 0.49 9 10 0.085544 0.05328
Table A-1 Network Data of Feeder 1 of Addis North 132/15 KV Substation

01 0.2981436 BF BF2 0.0578 0.036
02 0.2981436 BF2 01 0.10404 0.0648
03 0.2771436 01 02 0.07225 0.045
04 0.2771436 02 03 0.08381 0.0522
05 0.2771436 03 04 0.10115 0.063
06 0.2771436 04 05 0.10115 0.063
07 0.2771436 05 06 0.08381 0.0522
08 0.1495091 06 07 0.10693 0.0666
09 0.2771436 07 08 0.10693 0.10693
10 0.1495091 08 09 0.07225 0.045
11 0.1495091 09 10 0.10115 0.063
12 0.2771436 10 11 0.08381 0.0522
13 0.1495091 11 12 0.10404 0.0648
14 0.2771436 12 13 0.10693 0.0666
15 0.1495091 13 14 0.07225 0.045
16 0.934432 14 15 0.07225 0.045
17 0.934432 15 16 0.10693 0.0666
18 0.5747545 16 17 0.10115 0.063
19 0.2771436 17 18 0.08381 0.0522
20 0.1495091 18 19 0.10404 0.0648

21 0.2771436 19 20 0.08381 0.0522
22 0.1495091 20 21 0.10115 0.063
23 0.2771436 21 22 0.07225 0.045
24 0.2771436 22 23 0.07225 0.045
25 0.5747545 23 24 0.10693 0.0666
26 0.2354769 24 25 0.08381 0.0522
27 0.5747545 25 26 0.10115 0.063
28 0.1495091 26 27 0.08381 0.0522
29 0.1495091 27 28 0.10693 0.0666
30 0.1495091 28 29 0.07225 0.045
31 0.1495091 29 30 0.10115 0.063
32 0.5747545 30 31 0.10115 0.10115
33 0.2771436 31 32 0.08381 0.0522
34 0.1495091 32 33 0.10693 0.0666
35 0.1495091 33 34 0.10693 0.0666
36 0.2771436 34 35 0.07225 0.045
37 0.1495091 35 36 0.08381 0.0522
38 0.2771436 36 37 0.10404 0.0648
39 0.1495091 37 38 0.10693 0.0666
40 0.2771436 38 39 0.08381 0.0522
41 0.2771436 39 40 0.10115 0.063
42 0.2771436 40 41 0.07225 0.045

43 0.2771436 41 42 0.10693 0.0666
44 0.2771436 42 43 0.07225 0.045
45 0.2771436 43 44 0.10115 0.063
46 0.5980365 44 45 0.10404 0.0648
Table A-2 Network Data of Feeder 2 of Addis North 132/15 KV substation
1 0.231548 BF BF3 0.0578 0.036

2 0.7867537 BF3 01 0.13294 0.0828
3 0.1157399 01 02 0.07514 0.0468
4 0.07350729 02 03 0.07514 0.0468
5 0.23154800 03 04 0.0578 0.036
6 0.07350729 04 05 0.0867 0.054
7 0.115774 05 06 0.0867 0.054
8 0.07350729 06 07 0.13294 0.0828
9 0.115774 07 08 0.07514 0.0468
10 0.1102609 08 09 0.0867 0.054
11 0.08269571 09 10 0.0578 0.036
12 0.231548 10 11 0.07514 0.0468
13 0.231548 11 12 0.0867 0.054
14 0.231548 12 13 0.13294 0.0828
15 0.07350729 13 14 0.0578 0.036
16 0.231548 14 15 0.0578 0.036

17 0.07350729 15 16 0.0867 0.054
18 0.07350729 16 17 0.0867 0.054
19 0.231548 17 18 0.07514 0.0468
20 0.115774 18 19 0.0867 0.054
21 0.07350729 19 20 0.07514 0.0468
22 0.1102609 20 21 0.0867 0.054
23 0.7867537 21 22 0.13294 0.0828
24 0.115774 22 23 0.07514 0.0468
25 0.2940292 23 24 0.0867 0.054
26 0.07350729 24 25 0.07514 0.0468
27 0.07350729 25 26 0.0867 0.054
28 0.231548 26 27 0.0578 0.036
29 0.1102609 27 28 0.07514 0.0468
30 0.1102609 28 29 0.0867 0.054
31 0.231548 29 30 0.07514 0.0468
32 0.231548 30 31 0.07514 0.0468
33 0.78675370 31 32 0.13294 0.0828
34 0.231548 32 33 0.0867 0.054
35 0.231548 33 34 0.0867 0.054
36 0.115774 34 35 0.0578 0.036
37 0.07350729 35 36 0.07514 0.0468
38 0.23154800 36 37 0.0867 0.054
39 0.7867537 37 38 0.13294 0.0828
40 0.231548 38 39 0.13294 0.0828
41 0.115774 39 40 0.0867 0.054
42 0.07350729 40 41 0.07514 0.0468
43 0.23154800 41 42 0.0578 0.036

44 0.07350729 42 43 0.0867 0.054
45 0.07350729 43 44 0.07514 0.0468
46 0.07350729 44 45 0.0867 0.054
47 0.23154800 45 46 0.0578 0.036
48 0.07350729 46 47 0.0867 0.054
49 0.1102609 47 48 0.07514 0.0468
50 0.115774 48 49 0.0867 0.054
51 0.231548 49 50 0.07514 0.0468
52 0.115774 50 51 0.13294 0.0828
53 0.1102609 51 52 0.13294 0.0828
54 0.07350729 52 53 0.07514 0.0468
55 0.231548 53 54 0.07514 0.0468
56 0.07350729 54 55 0.0867 0.054
57 0.231548 55 56 0.07514 0.0468
58 0.231548 56 57 0.13294 0.0828
59 0.2021451 57 58 0.07514 0.0468
Table A-3 Network Data of Feeder 3 of Addis North 132/15KV substation
1 0.25 BF BF4 0.0578 0.036

2 0.99 BF4 01 0.0867 0.054
3 0.25 01 02 0.06936 0.0432
4 0.25 02 03 0.06936 0.0432
5 0.25 03 04 0.06358 0.0396

6 0.99 04 05 0.0867 0.054
Table A- 4 Network Data of Feeder 4 of Addis North 132/15 KV substation
1 0.25 BF BF5 0.0578 0.036

2 0.25 BF5 01 0.0867 0.054
3 0.32 01 02 0.09826 0.0612
4 0.75 02 03 0.0867 0.054
5 0.75 03 04 0.06936 0.0432
6 0.32 04 05 0.0578 0.036
7 0.32 05 06 0.06936 0.0432
8 0.32 06 07 0.09826 0.0612
Table A-5 Network Data of Feeder 5 of Addis North 132/15 KV substation
1 0.2951094 BF BF6 0.0578 0.036

2 0.09585703 BF6 01 0.09248 0.0576
3 0.09585703 01 02 0.07514 0.0468
4 0.4173342 02 03 0.0578 0.036
5 0.2951094 03 04 0.0578 0.036
6 0.4173342 04 05 0.07514 0.0468
7 0.2951094 05 06 0.07514 0.0468

8 0.09585703 06 07 0.0578 0.036
9 0.04792851 07 08 0.09248 0.0576
10 0.04792851 08 09 0.0578 0.036
11 0.09585703 09 10 0.0578 0.036
12 0.4173342 10 11 0.07514 0.0468
13 0.3019496 11 12 0.0578 0.036
14 0.09585703 12 13 0.09248 0.0576
15 0.2951094 13 14 0.09248 0.0576
16 0.04792851 14 15 0.0578 0.036
17 0.1437855 15 16 0.07514 0.0468
18 0.4173342 16 17 0.0578 0.036
19 0.4173342 17 18 0.07514 0.0468
20 0.2663594 18 19 0.0578 0.036
21 0.09585703 19 20 0.0578 0.036
22 0.4173342 20 21 0.09248 0.0576
23 0.09585303 21 22 0.07514 0.0468
24 0.4173342 22 23 0.0578 0.036
25 0.09585703 23 24 0.07514 0.0468
26 0.09585703 24 25 0.0578 0.036
27 0.02396426 25 26 0.07514 0.0468
28 0.04792851 26 27 0.0578 0.036
29 0.09585703 27 28 0.09248 0.0576
30 0.09585703 28 29 0.09248 0.0576
31 0.4173342 29 30 0.0578 0.036
32 0.02396426 30 31 0.0578 0.036
33 0.04792851 31 32 0.07514 0.0468
34 0.2663594 32 33 0.0578 0.036

35 0.09585703 33 34 0.07514 0.0468
36 0.4173342 34 35 0.0578 0.036
37 0.02396426 35 36 0.07514 0.468
38 0.09585703 36 37 0.0578 0.036
39 0.2663594 37 38 0.0578 0.036
40 0.09585703 38 39 0.09248 0.0576
41 0.2663594 39 40 0.0578 0.036
42 0.4173342 40 41 0.0578 0.036
43 0.4173342 41 42 0.07514 0.0468
44 0.2663594 42 43 0.0578 0.036
45 0.09585703 43 44 0.0578 0.036
46 0.09585703 44 45 0.09248 0.0576
47 0.09585703 45 46 0.0578 0.036
48 0.4173342 46 47 0.09248 0.0576
49 0.04792851 47 48 0.0578 0.036
50 0.09585703 48 49 0.07514 0.0468
51 0.2663594 49 50 0.0578 0.036
52 0.04792851 50 51 0.09248 0.0576
53 0.1437855 51 52 0.0578 0.036
54 0.4173342 52 53 0.0578 0.036
55 0.2663594 53 54 0.07514 0.0468
56 0.09585703 54 55 0.0578 0.036
57 0.1437855 55 56 0.0578 0.036
58 0.4173342 56 57 0.09248 0.0576
Table A- 6 Network Data of Feeder 6 of Addis North 132/15 Substation

Appendix D: Matlab program for reconfiguration of distribution network using
Genetic algorithm
%%%%%GENETIC ALGORITHM FOR F1 AND F2 POWER OPTIMIZATION %%%%%%
clc
clear all
close all
display('Program to optimize the power');
p_no=input ('Enter the size of initial population ::');
bits=input('Enter the bit size of each strings::') ;
final=[];
iterations=input('Enter the number of iterations to perform the process of genetic algorithm::');
tournament_no=input ('Enter the number of times you want to perform tournament selection ::');
%% power before optimization
p_1=0;
p_2=0;
I_1=0.926354; %in ampere
I_2=0.716774444747; %in ampere
R=0.578; %in ohm
u=10; %number of switches in feeder 1
v=46; %number of switches in feeder 2
n=u+v;
for i=1:u
p_1=p_1+(I_1^2)*R;
end;
for i=1:v
p_2=p_2+(I_2^2)*R;
end;
display ('power before optimization')
p_1
p_2
%% generate the bits randomly

p=rand(p_no,bits);
for i=1:p_no
for j=1:bits
if (p(i,j)>=0.5)
popu(i,j)=1;
else
popu(i,j)=0;
end;
end;
end;
for z=1:iterations
p_t=p_1+p_2;
val=zeros(p_no,1);
%%calculate the value of strings
for i=1:p_no
for j=bits:-1:1
k=bits-j;
val(i,1)=val(i,1)+pow2(popu(i,j),k);
end;
end;
val;
%% generate the random numbers to do the tournament selection
for i=1:tournament_no
select(i,:)=[floor(1+rand(1)*(p_no-1)) ceil(1+rand(1)*(p_no-1))];%% have to make changes
here
end;
select;
%%calculate the fitness value and perform tournament selection
tour_val_1=zeros(tournament_no,2);

tour_val_11=zeros(select(i,1),1);
for j=2:val(select(i))+1
tour_val_11(j)=tour_val_11(j-1)+(I_2^2)*R;
tour_val_1(i)=tour_val_11(j);
end;
for j=2:val(select(i,2))+1
end;
tor_min1=min(abs(tour_val_1(i)-p_t/2),abs(tour_val_2(i)-p_t/2));
if tor_min1==abs(tour_val_1(i)-p_t/2)
fitest(i,:)=popu(select(i,1),:);
else
end;
end;
tour_val_1;
tour_val_2;
fitest;
prob_crossover=0.8;
prob_mutation=0.01;
crossover_point=2;
for x=1:ceil(p_no/2)
%%selecting of parents
parents=[];
crossover=[];
crossover_val=[];
crossover_val_mapped=[];
crossover_fitness=[];

parents=[fitest(floor(1+rand(1)*(tournament_no-1)),:); fitest(ceil(1+rand(1)*(tournament_no-
1)),:)];
%selecting of childrens
val_gen=rand(1);
child=[];
if (val_gen<prob_crossover)
child=[[parents(1,1:crossover_point)
parents(2,(crossover_point+1):bits)];[parents(2,1:crossover_point)
parents(1,(crossover_point+1):bits)]];
end;
%%calculate the value of parents and children
crossover=[parents;child];
crossover_val=zeros(size(crossover,1),1);
for i=1:size(crossover,1)
for j=bits:-1:1
k=bits-j;
crossover_val(i,1)=crossover_val(i,1)+pow2(crossover(i,j),k);
end;
end;
crossover_val;
%%calculate the fitness value to select the the most suitable solution
for j=1:size(crossover_val,1)
p_21=zeros(size(crossover_val,1),1);
% p_2=zeros(size(crossover_val,1),1);
for i=2:crossover_val(j,1)+1
p_21(i,1)=p_21(i-1,1)+(I_2^2)*R;
p_2(j,1)=p_21(i,1);
end;
crossover_fitness=p_2;
end

select_min1=min(abs(crossover_fitness-p_t/2));
for i=1:size(crossover_val,1)
if (select_min1==abs(crossover_fitness(i,1)-p_t/2))
pos1=i;
break;
end;
end;
%%the position of the max fitness value after crossover
pos1;
% to select the 2nd bst fitness value.
select_min2=abs(crossover_fitness(1,1)-p_t/2);
pos2=1;
if select_min2==select_min1
pos2=2;
end;
for i=1:size(crossover_fitness,1)
if(select_min1==abs(crossover_fitness(i,1)-p_t/2))
continue;
elseif (abs(crossover_fitness(i,1)-p_t/2)~=select_min2)
select_min2=abs(crossover_fitness(i,1)-p_t/2);
pos2=i;
end;
end;
% the values of 2nd best fitness values and its position
select_min2;
pos2;
% taking the best 2 fitest solution and creating a different mating pool
final=[final;crossover(pos1,:);crossover(pos2,:)];
end;%% end of crossover
%performing mutation over the solutions generated after crossover.

for i=1:size(final,1)
for j=1:size(final,2)
val_gen=rand(1);
if (val_gen<prob_mutation)
final1(i,j)=xor(final(i,j),1);
else
final1(i,j)=final(i,j);
end;
end;
end;
final=[];
popu=final1;
p_no=size(popu,1);
val=zeros(p_no,1);
for i=1:p_no
for j=bits:-1:1
k=bits-j;
end;
end;
u2=val(1,1);
v2=n-val(1,1);
p_1=0;
p_2=0;
for i=1:u2
p_2=p_2+(I_2^2)*R;
end;
p_1=p_t-p_2;
end;%%end to the number of iterations for the entire process;
display('The final best solution is given as....')
v2

u2
display('power after optimization')
p_1
p_2
%%%%%%GENETIC A LGORITHM FOR F3 AND F4 POWER OPTIMIZATION %%%%

clc
clear all
close all
display('Program to optimize the power ');
p_no=input ('Enter the size of initial population ::');
final=[];
tournament_no=input('Enter the number of times you want to perform tournament selection::');
p_3=0;
p_4=0;
I_3=0.57589244; %in ampere
I_4=0.9269762; %in ampere
R=0.578; %in ohm
n=u+v;
for i=1:u
p_3=p_3+(I_3^2)*R;
end;
for i=1:v
p_4=p_4+(I_4^2)*R;
end;
display('power before optimization')

p_3
p_4

p=rand(p_no,bits);
for i=1:p_no
for j=1:bits
if (p(i,j)>=0.5)
popu(i,j)=1;
else
popu(i,j)=0;
end;
end;
end;
for z=1:iterations
p_t=p_3+p_4;
val=zeros(p_no,1);
for i=1:p_no
for j=bits:-1:1
k=bits-j;
end;
end;
val;
here
end;
select;

end;
end;
else
end;
end;
tour_val_1;
tour_val_2;
fitest;
prob_crossover=0.8;
prob_mutation=0.01;
crossover_point=2;
parents=[];
crossover=[];

crossover_val=[];
1)),:)];
val_gen=rand(1);
child=[];
end;
for j=bits:-1:1
k=bits-j;
end;
end;
crossover_val;
% p_4=zeros(size(crossover_val,1),1);
p_31(i,1)=p_31(i-1,1)+(I_3^2)*R;
p_3(j,1)=p_31(i,1);

end;
end
pos1=i;
break;
end;
end;
pos1;
pos2=1;
pos2=2;
end;
continue;
pos2=i;
end;
end;
select_min2;
pos2;

% crossover1=dec2bin(crossover_fitness,7);
% taking the best 2 fitest solution and creating a different mating pool
val_gen=rand(1);
else
end;
end;
end;
final=[];
popu=final1;
p_no=size(popu,1);
val=zeros(p_no,1);
for i=1:p_no
for j=bits:-1:1
k=bits-j;
end;
end;
u2=val(1,1);
v2=n-val(1,1);
p_3=0;
p_4=0;
for i=1:u2
p_3=p_3+(I_3^2)*R;

end;
p_4=p_t-p_3;
u2
v2
p_3
p_4
%%%%%% GENETIC A LGORITHM FOR F5 AND F6 POWER OPTIMIZATION %%%%%

clc
clear all
close all
display('Program to optimize the power ');
p_no=input('Enter the size of initial population::');
final=[];
tournament_no=input('Enter the number of times you want to perform tournament selection ::');
p_5=0;
p_6=0;
I_5=0.84222477; %in ampere
I_6=0.59454; %in ampere
R=0.578; %in ohm
n=u+v;
for i=1:u
p_5=p_5+(I_5^2)*R;

end;
for i=1:v
p_6=p_6+(I_6^2)*R;
end;
display('power before optimization')
p_5
p_6
p=rand(p_no,bits);
for i=1:p_no
for j=1:bits
if (p(i,j)>=0.5)
popu(i,j)=1;
else
popu(i,j)=0;
end;
end;
end;
for z=1:iterations
p_t=p_5+p_6;
val=zeros(p_no,1);
for i=1:p_no
for j=bits:-1:1
k=bits-j;
end;
end;
val;

here
end;
select;
end;
end;
else
end;
end;
tour_val_1;
tour_val_2;
fitest;
prob_crossover=0.8;
prob_mutation=0.01;
crossover_point=2;

parents=[];
crossover=[];
crossover_val=[];
1)),:)];
val_gen=rand(1);
child=[];
end;
for j=bits:-1:1
k=bits-j;
end;
end;
crossover_val;

p_61(i,1)=p_61(i-1,1)+(I_6^2)*R;
p_6(j,1)=p_61(i,1);
end;
end
pos1=i;
break;
end;
end;
pos1;
pos2=1;
pos2=2;
end;
continue;
pos2=i;
end;
end;

select_min2;
pos2;
% crossover1=dec2bin(crossover_fitness,7);
% taking the best 2 fittest solution and creating a different mating pool
val_gen=rand(1);
else
end;
end;
end;
final=[];
popu=final1;
p_no=size(popu,1);
val=zeros(p_no,1);
for i=1:p_no
for j=bits:-1:1
k=bits-j;
end;
end;
u2=val(1,1);
v2=n-val(1,1);
p_5=0;
p_6=0;

for i=1:u2
p_6=p_6+(I_6^2)*R;
end;
p_5=p_t-p_6;
v2
u2
p_5
p_6

Habtamu Eshete PDF

Uploaded by

Copyright:

Available Formats

Habtamu Eshete PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Habtamu Eshete PDF

Uploaded by

Copyright:

Available Formats

ADDIS ABABA UNIVERSITY

STUDY ON POWER LOSS MINIMIZATION FOR DISTRIBUTION NETWORK

Mr. KIROS TESFAY

Addis Ababa, Ethiopia

STUDY ON POWER LOSS MINIMIZATION FOR DISTRIBUTION NETWORK

CASE STUDY: ADDIS NORTH 132/15 KV SUBSTATION

Approval by Board of Examiners

Chairman, Dept. Graduate Committee Signature

Internal Examiner Signature

External Examiner Signature

Habtamu Eshete, AAiT, MSc. Thesis Page I

Habtamu Eshete ____________________

Place: Addis Ababa Institute of Technology

Date of Submission: ________________

Habtamu Eshete, AAiT, MSc. Thesis Page II

Habtamu Eshete, AAiT, MSc. Thesis Page III

Distribution network reconfiguration is an optimization problem and needs a suitable algorithm

Keywords: Network Reconfiguration, Genetic Algorithm Optimization, Power Loss Minimization

Habtamu Eshete, AAiT, MSc. Thesis Page IV

ACKNOWLEDGEMENT ............................................................................................................ III

LIST OF FIGURES ................................................................................................................... VIII

LIST OF TABLES ........................................................................................................................ IX

LIST OF ACRONYMS ................................................................................................................. X

Chapter 1 Introduction .................................................................................................................... 1

1.1 Background ................................................................................................................... 1

1.5.2 Specific Objectives ........................................................................................ 4

1.6 Methodology ................................................................................................................. 4

Theoretical Background and Literature Review ............................................................................. 6

2.1 Theoretical Background ................................................................................................ 6

2.2.1.1 Genetic Algorithm (GA) .................................................................. 7

2.2.1.2 Simulated Annealing ........................................................................ 8

2.2.1.3 Artificial Neural Network ................................................................ 8

2.2.1.4 Ant Colony Optimization ................................................................. 9

2.2.1.5 Particle Swarm Optimization ........................................................... 9

2.2.2 Knowledge Based Methods ......................................................................... 10

Habtamu Eshete, AAiT, MSc. Thesis Page V

2.2.2.2 Linear programming Techniques ................................................... 12

2.2.2.3 Fuzzy Logic Method ...................................................................... 12

2.2.2.4 Tabu Search Method ...................................................................... 13

2.2.3 Mixed Methods ............................................................................................ 13

Research Methodology ................................................................................................................. 15

3.1 Problem formulation ................................................................................................... 15

3.2 Genetic Algorithm ...................................................................................................... 18

3.2.2 How Genetic Algorithms Work ................................................................... 18

3.2.3 Genetic Algorithm Operators ....................................................................... 21

3.2. 4 Encoding ..................................................................................................... 30

3.3 DIgSILENT PowerFactory Overview ........................................................................ 31

Simulation Studies and Analysis of Results ................................................................................. 33

4.1 Addis North 132/15 KV Substation Test Systems......................................................... 33

4.1.2 Addis North 132/15 KV Substation Simulation Result ................................. 33

4.2 Genetic Algorithm optimization result ....................................................................... 40

Conclusions, Recommendations and Future Work ....................................................................... 44

5.1 Conclusions ................................................................................................................. 44

Habtamu Eshete, AAiT, MSc. Thesis Page VI

Appendix A: Addis North 132/15 kV Substation feeders before reconfiguration ....................... 51

Appendix B: Addis North 132/15 kV Substation feeders network after reconfiguration............. 61

Appendix C: Addis North 132/15 kV Substation feeders data ..................................................... 71

Appendix D: Matlab program for reconfiguration of distribution network using Genetic

Habtamu Eshete, AAiT, MSc. Thesis Page VII