DP 2 Tarun PDF

MINIMIZE POWER LOSSES USING
PARTICLE SWARM OPTIMIZATION

TECHNIQUES
By
Patel Tarunkumar Dasharthbhai
(170910707018)
Under the guidance of

Assistant. Prof. A. G. ACHARYA
Electrical Engineering Department

SSSRGI College vadasma
A Thesis Submitted to Gujarat Technological University in Partial Fulfilment of the Requirements for
the Master of Engineering Degree in Electrical Engineering
MARCH 2019
Shri satsangi saketdham ram ashram group of institutions, vadasma

-1-
CERTIFICATE
This is to certify that work embodied in this dissertation titled “Minimize power loss using
particle swarm optimization techniques ” was carried out by Patel Tarunkumar Dashrathbhai
(Enrollment no. 170910707018) at SSSRGI VADASMA (091) for Master of Engineering
degree in Electrical Engineering to be awarded by Gujarat Technological University. This work
has been carried out under my guidance and supervision and it is up to my satisfaction.
Date:
Place:
Signature and Name of Signature and Name of
Guide Head of department
Assistant prof. A.G. Acharya Assistant prof. A.G. Acharya
SSSRGI Vadasma SSSRGI Vadasma
2
COMPLIANCE CERTIFICATE
This is to certify that research work embodied in this dissertation titled “Minimize power loss using
particle swarm optimization techniques” was carried out by a Patel Tarunkumar Dashrathbhai
(Enrollment no. 170910707018) at SSSRGI Vadasma (091) for partial fulfillment of Master of
Engineering degree to be awarded by Gujarat Technological University. He has complied to the
comments given by the Dissertation phase – I as well as Mid Semester Thesis Reviewer to my
satisfaction.
Date:
Place:
Signature and Name of Signature and Name of
Guide Head of department
Assistant prof. A.G. Acharya Assistant prof. A.G. Acharya
SSSRGI Vadasma SSSRGI Vadasma
3
PAPER PUBLICATION CERTIFICATE
(Enrollment no. 170910707018) at SSSRGI Vadasma (091) for partial fulfillment of Master of
Engineering degree to be awarded by Gujarat Technological University has published article Minimize
power loss using particle swarm optimization techniques for publication by the International Journal
of Electrical Engineering and Technology (IJEET) at Chennai, Tamilnadu during March- April
2019 in volume 10 issue 2.
Date:
Place:
Candidate Name Signature and Name of Guide
Patel Tarunkumar D. Assistant prof. A.G. Acharya
Enrollment No.( 170910707018 ) SSSRGI Vadasma
4
THESIS APPROVAL CERTIFICATE
(Enrollment no. 170910707018) at SSSRGI VADASMA (091) is approved for the degree of
Master of Engineering with specialization in Electrical Engineering by Gujarat Technological
University.
Date:
Place:
Examiner’s Sign and Name:
---------------------------------- ----------------------------------
( ) ( )
5
DECLARATION OF ORIGINALITY
We hereby certify that we are the sole authors of this thesis and that neither any part of this thesis
nor the whole of the thesis has been submitted for a degree to any other University or Institution.
We certify that, to the best of our knowledge, the current thesis does not infringe upon anyone’s
copyright nor violate any proprietary rights and that any ideas, techniques, quotations or any other
material from the work of other people included in our thesis, published or otherwise, are fully
acknowledged in accordance with the standard referencing practices. Furthermore, to the extent that
we have included copyrighted material that surpasses the boundary of fair dealing within the
meaning of the Indian Copyright (Amendment) Act 2012, we certify that we have obtained a written
permission from the copyright owner(s) to include such material(s) in the current thesis and have
included copies of such copyright clearances to our appendix.
We declare that this is a true copy of the thesis, including any final revisions, as approved by the
thesis review committee.
We have checked write up of the present thesis using the anti-plagiarism database and it is an
allowable limit. Even though later on in case of any complaint pertaining of plagiarism, we are solely
responsible for the same and we understand that as per UGC norms, University can even revoke
Master of Engineering degree conferred to the student submitting this thesis.
Date:
Signature of Student: _____________ Signature of Guide: _____________

Name of Student: __________________ Name of Guide: _________________
Enrollment No: _________________ Institute Code: _________
6
ACKNOWLEDGEMENT
The dissertation has been one of the most significant academic challenges that I have ever
had to face. Without the support, patience, and guidance of the following people, this study would
not have been completed. It is to them that I owe my deepest gratitude.
Most importantly, I would like to thank my father, my family members and divine family
for everything that they have done for me throughout my life and the joy they have brought to me
for just having their support. Much of my success has been because of their love and encouragement.
It gives me immense pleasure to express my deepest sense of gratitude and sincere thanks
to my highly respected and esteemed supervisor assistant prof. Anandkumar G. Acharya in the
Electrical Engineering Department of SSSRGI vadasma for his revered continuous supervision
throughout my dissertation work, which made this task a pleasant job. He has always provided me
the wise advice, useful discussions, and comments. It was a real pleasure to work under his
supervision.
I am grateful to assistant prof. Anandkumar G. Acharya, Head of Electrical Engineering
Department and all the faculty members of the Electrical Engineering Department for the guidance
and suggestions whenever I needed them.
Further, I am thankful to all colleagues for their continuous support during my project work.
I would like to thank all my friends from SSSRGI vadasma, Ahmedabad for encouraging me and
co-operate throughout my post-graduation.
Patel Tarunkumar Dashrathbhai
7
TABLE OF CONTENTS
Title page 1
Certificate Page 2
Compliance Certificate 3
Paper Publication Certificate 4
Thesis Approval Certificate 5
Declaration of Originality 6
Acknowledgement 7
Table of Contents 8
List of Figures 9
List of Tables 11
Appendix 12
Abstract 13
CHAPTER 1 INTRODUCTION 15
1.1 PROBLEM STATEMENT 16
1.2 BASIC OF OPTIMIZATION 16
1.3 OBJECTIVE OF PROJECT 16
CHAPTER 2 LITERATURE REVIEW 17
CHAPTER 3 POWER SYSTEM OPTIMIZATION AND POWER LOSSES 20
3.1 CONVENTIONAL METHOD 20

3.1.1 NEWTON METHOD 20
3.1.2 GRADIENT METHOD 24
3.1.3 LIINEAR PROGRAMMING METHOD 24
3.1.4 QUADRATIC PROGRAMMING METHOD 24
3.1.5 INTERIOR POINT METHOD 25
3.1.6 DISADVANTAGE OF CONVETIONAL METHOD 25
8
3.2 INTELLIGENT METHODS 26
3.2.1 PSO METHOD 26
3.2.2 MERIT OF PSO 28
3.2.3 PSEUDO CODE 28
3.2.4 COMPARASION OF CLASSICAL AND (AI) METHOD 29
3.2.5 PSO ALGORITHM 30
CHAPTER 4 NEWTON’S RAPSHSON METHOD FOR 31
LOSS MINIMIZATIO
4.1 NEWTON RAPHSON DATA 31
4.2 COMPUTATIONAL ALGORITHMS FOR NEWTON 31
RAPHSON LOAD FLOW METHOD
4.3 FLOW CHART FOR NEWTON 33
RAPHSON LOAD FLOW METHOD
CHAPTER 5 PARTICLE SWARM OPTIMIZATION FOR 34

LOSS MINIMIZATION
5.1 MATLAB SIMULATION OF PSO FOR MINIMIZE POWER LOSS 34
5.2 SOLUTION BY PSO USING PSO ALGORITHM 35
5.3 PROBLEM FORMULATION 35
5.4 OBJECTIVE FUNCTION FOR POWER LOSSES OPTIMIZATION 36
5.4.1 EQUALITIY CONSTRAINTS 38
5.4.2 UNEQUALITY CONSTRAINTS 38
5.4.3 FORMATION OF FITNESS FUNCTION 39
5.5 MODIFICATION CONCEPT OF SEARCHING POINT BY PSO 41
5.5.1 VELOCITY UPDATE MECHANISM 41
5.5.2 CURRENT POSITION MODIFICATION 43
5.6 PSO ALGORITHM TO IEEE 14 BUS TEST SYSTEME 43
5.7 PSO FLOW CHART FOR MINIMIZE POWER LOSSE 45
9
CHAPTER 6 IEEE 14 BUS TEST SYSTEM 47
6.1 SYSTEM DIAGRAM 47
6.2 TYPE OF BUSES WITH VARIABLE 47
CHAPTER 7 RESULTS 48
NR METHOD RESULT 48
PSO RESULTS ON BENCH MARK FUNCTION 52
PSO METHOD RESULTS 55
COMPERISION BETWEEN NR METHOD AND PSO METHOD RESULTS 77
CHAPTER 8 CONCLUSION 78
REFERENCE 80
APPENDIX A SYSTEM DATA 81
APPENDIX B REVIEW CARD AND COPLIENCE REPORT 83
APPENDIX C PAPER PUBLICATION CERTIFICATE 89
10
LIST OF FIGURES
Figure Name Page No.
FIG. 1.1 SIMPLE OPTIMIZATION 16
FIG. 3.1 OPF SOLUTION METHODOLOGIES 20
FIG. 3.2 EXAMPLE OF BIRD FLOCKING IN NATURE 27
FIG. 3.3 EXAMPLE OF SCHOOL OF FISH IN NATURE 28
FIG. 3.4 FLOW CHART OF PARTICLE SWARM 30

OPTIMIZATION
FIG. 4..1 FLOW CHART OF NEWTON RAPHSON METHOD 33

FOR POWER LOSS
FIG. 5.1 MODIFICATION CONCEPT OF A SEARCHING 41

POINT BY PSO
FIG 5.2 PARTICLE SWARM OPTIMIZATION FLOW CHART 45

FOR POWER LOSS MINIMIZE
FIG.6.1 IEEE 30 BUS TEST SYSTEM 47
11
LIST OF TABLES
FIGURE NAME PAGE NO.
TABLE 3.1 COMPARISON OF GAUSS SEDIAL AND 20

NEWTON RAPHSON METHOD
TABLE 6.1 TYPE OF BUSES WITH VARIABLE 47
TABLE 7.1 PSO RESULTS ON BECH MARK FUNCTION 65
TABLE 7.2 COMPERISION BETWEEN NR METHOD AND PSO

METHOD RESULTS 77
12
Appendix A
TABLE A.1 BUS DATA
TABLE A.2 SHUNT CAPACITOR DATA
TABLE A.3 TRANSMISSION LINE DATA
13
ABSTRACT:
This dissertation presents the solution of optimal power flow (OPF) using particle swarm
optimization (PSO). The objective function is to minimize power loss by adjustment the power
control variables and at the same time satisfying the equality and inequality constraints. The
proposed particle swarm optimization (PSO) method is compared with newton’s Raphson method
(conventional method) approach on the standard IEEE 30 bus system. The analysis indicated that
the Particle Swarm Optimization (PSO) method was the most efficient method in terms of
minimizing power loss. This can be concluded that the Artificial Intelligence (AI) method such as
Particle Swarm Optimization (PSO) is the most suitable and efficient method for analyzing the
Optimal Power Flow (OPF) problem in terms of minimizing power loss.
14
CHAPTER-1
INTRODUCTION
The main task in the electrical power system is to optimize a selected objective function such as fuel
cost minimization by adjusting the state variable as well as at the same time satisfying the equality
and inequality constraints. The control variables are generator real power, generator bus voltage,
transformer tap changer and the Reactive power source such as shunt capacitor. There are many
techniques available to handle such OPF problems such as non-linear programming, quadratic
progressing, linear programming; Newton-based method sequential unconstrained minimization and
interior point method. These methods have the drawback-the convergence characteristics are
sensitive to the initial condition. Generally, the OPF problems are non-convex, non-smooth and non-
differential. So we have to develop such a new optimization technique that is efficient to overcome
their drawbacks and to handle the difficulties easily. A heuristic algorithm such as GA, gradient
method and evolutionary computation technique has been recently proposed for solving the OPF
problem. Recently some deficiencies have been identified in GA performance. A new evolutionary
technique called particle swarm optimization has been proposed and introduced by The Kennedy.
This technique is based on the sociological behavior such as fish schooling and bird flocking. In this
paper, a novel PSO based approach is proposed to solve the OPF problem. Since the PSO handle
both continuous and discrete variable easily. Therefore this method can be easily applied to mixed
integer nonlinear program. PSO is suitable for cost minimization because it can handle such
constraints easily. Here the proposed PSO approach is tested for 14 bus system and compare with
the NR method..
15
1.1 PROBLEM STATEMENT
The Optimal Power Flow has become widely known that has received much attention in power
system because it has been widely used in order to plan a power system. However, the problem of
Optimal Power Flow problem always became the main subject in the power system which one of the
objective function of the OPF problem is minimizing loss of power. In this paper, this specific
objective function will be the problem that will be analysed in this project on how to minimize the
power loss in order to optimize the power transmission in a power system. Furthermore, this problem
is a non-linear optimize so it may require a more complex formulation, since the set of equations
involved may not be linearized. Therefore, non-linear technique was proposed in this paper in order
to analyse this problem.
1.2 BASIC OF OPTIMIZATION
Set of in put process Best out put
Fig. 1.1 Simple optimization
Optimization is the process of getting the best result under the given circumstance. Work done should
be maximum in minimum time. Optimization can be finding the maximum or minimum of the
function.
Optimization help finds the input that gives the best output.
In an Optimal Power Flow, the values of some or all of the control variables need to be found so as
to optimize (minimize or maximize) a predefined objective.
OPF aims to optimize a certain objective, subject to the network power flow equations and system
and equipment operating limits. The optimal condition is attained by adjusting the available controls
to minimize an objective function subject to specified operating and security requirements.
1.3 OBJECTIVES OF PROJECT
optimize the power transmission in a power system by minimizing the power loss.
implement Particle Swarm Optimization (PSO) in Optimal Power Flow (OPF) problem.
compare the effectiveness of Particle Swarm Optimization with Newton Raphson method.
16
CHAPTER 2
LITERATURE REVIEW
1) “Particle Swarm Optimization Based Optimal Reactive Power Dispatch”, IEEE 2015,
Sundaram Pandya and Ranjit Roy.
REVIEW : Optimal reactive power dispatch is one of the major and important optimization
problems in electrical power system operation and control. This is nothing but multi-objectives,
nonlinear, minimization problem of power system optimization. This paper focuses on this
subproblem of optimal power flow. The particle swarm optimization (PSO) is one of the best
population-based intelligent technique of optimization. The basic PSO algorithm is used for optimal
reactive power dispatch. This approach applies to the IEEE-14 bus, IEEE-30 bus, IEEE57 bus, and
IEEE-39 New England bus test systems for minimization of active power loss. Simulation results
are compared with the other optimization algorithm.
2) “Optimal Power Flow Based on PSO in Electrical Power System” , International Journal of
Advances in Electrical and Electronics Engineering, ISSN: 2319-1112 Prashant Kumar ,
Vikas Dubey , Deepak Sharma.
REVIEW : This paper narrates an application of Particle Swarm Optimization (PSO) for economic
load dispatch. The main objective function is to optimize the Fuel cost and Active power loss by
considering the power flows within the allowable limits. The PSO is simple in concept and easy in
implementation. It does not require any derivative information, sure and fast convergence. Moreover,
the problem can be formulated as a mixed-integer non-linear optimization problem (MNILP).
Optimization of Fuel cost and Active power loss by adjusting the control variables such as generator
voltage, transformer tap setting and other sources of reactive power such as shunt capacitor banks.
The effectiveness of the proposed algorithm is tested on IEEE 14 bus system and their performances
are compared with the results. The result shows that PSO can converge to an optimum solution with
higher accuracy in comparison with Evolutionary Programming (EP) method. The proposed PSO
method is demonstrated and compared with Evolutionary Programming (EP) approach on the
standard IEEE 14-bus system.
3)“Transmission Loss Minimization Using Optimization Technique Based On PSO ”,IOSR
Journal of Electrical and Electronics Engineering (IOSR-JEEE) Volume 6, Issue 1 (May. - Jun.
2013), PP 01-05. Sunil Joseph P , C.DineshBalaji.
REVIEW : : Optimal power flow (OPF) is defined as the optimization of operating states of a
power system and the corresponding settings of control variables. In this paper, a particle swarm
optimization (PSO) with an aging leader and challengers (ALC-PSO) is applied for the solution of
OPF problem of power system. This study is implemented on modified IEEE 30-bus test power
system with different objectives that reflect minimization of either fuel cost or active power loss or
sum of total voltage deviation. The results presented in this paper demonstrate the potential of the
17
proposed approach and show its effectiveness and robustness for solving the OPF problems over the
other evolutionary optimization techniques surfaced in the recent state-of-theart literature.
3) “Constrained Optimal Power Flow using Particle Swarm Optimization”, International

Journal of Emerging Technology and Advanced Engineering (ISSN 2250-2459, Volume 2,
Issue 2, February 2012), C.Kumar, Dr. Ch. Padmanabha Raju.
REVIEW : Optimal power flow is an optimizing tool for operation and planning of modern power
systems. This paper presents the solution of the optimal power flow (OPF) using particle swarm
optimization (PSO). The main goal of this paper is to verify the viability of using PSO problem
composed by different objective functions. This OPF problem involves the optimization of various
types of objective functions while satisfying a set of operational and physical constraints while
keeping the power outputs of generators, bus voltages, shunt capacitors/reactors and transformers
tap settings in their limits. The proposed PSO method is demonstrated and compared with
Evolutionary Programming (EP) approach on the standard IEEE 14-bus system. The results show
that the proposed PSO method is capable of obtaining higher quality solutions efficiently in OPF
problem.
4) “ Optimal power flow by particle swarm optimization with an aging leader and challengers”,
International Journal of Engineering, Science and Technology Vol. 7, No. 3, 2015, pp. 123-132
Rudra Pratap Singh , V. Mukherjee, S.P. Ghoshal.
REVIEW : Optimal power flow (OPF) is defined as the optimization of operating states of a
power system and the corresponding settings of control variables. In this paper, a particle swarm
optimization (PSO) with an aging leader and challengers (ALC-PSO) is applied for the solution of
OPF problem of power system. This study is implemented on modified IEEE 30-bus test power
system with different objectives that reflect minimization of either fuel cost or active power loss or
sum of total voltage deviation. The results presented in this paper demonstrate the potential of the
proposed approach and show its effectiveness and robustness for solving the OPF problems over the
other evolutionary optimization techniques surfaced in the recent state-of-theart literature.
5) “particle swarm optimization of power loss in distribution system”, IEEE 2015

𝟏𝟐𝒕𝒉 international conference on information technology – generation, Joseph B. abugri and
marc karam.
REVIEW : This research utilizes particle swarm optimization to minimize the total active power
losses in an IEEE 6 bus transmission system. newton- Raphson load flow algorithm is used in
computing power losses, with particle swarm optimization algorithm which is used to minimize
power losses. The proposed particle swarm optimization algorithm gave very satisfactory simulation
results. These successful simulation results confirm the effectiveness of particle swarm optimization
in minimizing distribution network power losses.
18
6) “power loss reduction in power system based on PSO :case study”, International journal of
computer application (0975-8887) vol 164 – no 10, April 2017, Sameer singh, vivek kumar jain,
and upendra prasad.
REVIEW: In this paper, discussed reactive power dispatch (loss reduction) with the contribution
of particle swarm optimization. The proposed algorithm has been applied to achieve the major
objective as the system loss minimization with satisfied equality and inequality constraints. The main
objective is to minimize the active power loss in the network while satisfying all the power system
operation variable constraints (equality and inequality). The particle swarm algorithm has been
coded using MATLAB software. The simulation results show that the PSO algorithm always leads
to a better result As losses are minimized, the power system will remain stable. Hence from
simulation results shown, PSO optimization technique proves well with reduced losses.
7) “ Loss power minimization using particle swarm optimization”,2006 international joint

conference neural networks July 16-21, 2018, A.A.A. Esmin, and G.Lambert – torres, member,
IEEE.
REVIEW: This paper presents a particle swarm optimization as an efficient approach for loss
reduction study. This issue can be formulated as a nonlinear optimization problem. The proposed
approach employs the PSO algorithm for optimal setting of Optimal Power Flow based on Loss
Minimization (LM) function. The proposed approach has been examined and tested on the standard
IEEE 30-bus test system. The results are promising and show the effectiveness of the proposed
approach. The beginning the PSO with 5 iterations obtained best results and get better with the
increment of iteration number. The results in PSO obtained the best result with different population
numbers. The PSO using population = (5, 10 and 15). The best results are obtained with the
population= 15.
8) “Optimal reactive power flow control for minimization of active power losses using particle
swarm optimization”, 20125 conference on power control, communication and computational
technologies for sustainable growth (PCCCTSG) December 11-12,2015, Kurnool, Andhra
Pradesh, india,
N. Tejswara Rao, Jagannath ch Yadav b. and anyapu jagannadham.
REVIEW: The main goal of this work is to determine the optimum operating conditions of the
control variables within the constraints of control variables and minimization of transmission losses
at normal operating condition of the system. For obtaining the optimum operating conditions of the
control variables and minimization of transmission losses "Particle Swarm Optimization (PSO)" is
being used. The PSO method is applied on 6-bus Ward Hale & 30-bus IEEE system. n this work,
power losses in transmission lines are calculated using power flow studies. The power losses
decreased to a better extent after applying particle swarm optimization within the limits of voltage
profile for a given power system. When comparing both load flow and particle swarm optimization
the minimum losses are obtained from particle swarm optimization.
19
CHAPTER 3
THE SOLUTION OF METHODOLOGIES
The OPF methods are broadly grouped as Conventional and Intelligent. The conventional
methodologies include the well-known techniques like Gradient method, Newton method, Quadratic
Programming method, Linear Programming method, and Interior point method. Intelligent
methodologies include the recently developed and popular methods like Genetic Algorithm, Particle
swarm optimization.
1. Conventional (classical) methods

2. Intelligent methods.
OPF Solution mythologies
Conventional method intelligent method

Gradient method Genetic algorithm
Newton method ant colony method
Linear method evolutionary programming
Quadratic programming artificial neural network
Interior method particle swarm optimization
Table 3.1 OPF Solution Methodologies (ref. book 12)
3.1 CONVENTIONAL METHOD
conventional methods are used to effectively solve OPF. The application of these methods had
been an area of active research in the recent past.
The conventional methods are based on mathematical programming approaches and used to solve
different size of OPF problems. To meet the requirements of different objective functions, types of
application and nature of constraints.
20
3.1.1 Newton Method:
In the area of Power systems, Newton’s method is well known for the solution of Power Flow. It has
been the standard solution algorithm for the power flow problem for a long time the Newton
approach is a flexible formulation that can be adopted to develop different OPF algorithms suited to
the requirements of different applications. Although the Newton approach exists as a concept entirely
apart from any specific method of implementation, it would not be possible to develop practical OPF
programs without employing special sparsity techniques. The concept and the techniques together
comprise the given approach. Other Newton-based approaches are possible.
Newton’s method is a very powerful solution algorithm because of its rapid convergence near the
solution. This property is especially useful for power system applications because an initial guess
near the solution is easily attained. System voltages will be near rated system values, generator
outputs can be estimated from historical data, and transformer tap ratios will be near 1.0 p.u
Newton’s Raphson method is solving a set of the non-linear algebraic equation. NR method is
solving power flow based problem. Has also good convergence characteristics even for a large
system it takes only two to four iteration to converge.
Power flow model of newton Raphson method:
The equation describing the performance of the network.

The bus admittance is given
I= YV
In expanding from these equations are
𝐼1 𝑌11 𝑌12 -------------------- 𝑌1𝑛 𝑉1

𝐼1 𝑌21 𝑌22 -------------------- 𝑌2𝑛 𝑉2
| | | |
| = | | = |
| | | |
𝐼𝑛 𝑌𝑁1 𝑌𝑁2 -------------------- 𝑌𝑁𝑛 𝑉𝑛
21
Typical element of the bus admittance matrix is
𝑌𝑖𝑗 = 𝑌𝑖𝑗 < 𝜃𝑖𝑗 𝐴 = | 𝑌𝑖𝑗 | 𝑐𝑜𝑠𝜃𝑖𝑗 + |𝑌𝑖𝑗 | sin𝜃𝑖𝑗 = 𝐺𝑖𝑗 + 𝑗𝐵𝑖𝑗
Voltage at atypical bus i is
𝑉𝑖 = | 𝑉𝑖 | < 𝛿𝑖 = |𝑉𝑖 |( 𝑐𝑜𝑠 𝛿𝑖 + j 𝑠𝑖𝑛 𝛿𝑖 )
The current injection in to the network at bus i is given by
𝐼𝑖 = 𝑌𝑖1 𝑉1 + 𝑌𝑖2 𝑉2 +-----------------------+ 𝑌𝑖𝑛 𝑉𝑛
𝐼𝑖 = ∑𝑁
𝑛=1 𝑌𝑖𝑛 𝑉𝑛
The complex power entering the network at bus i is given by

𝑃𝑖 + 𝑗𝑄𝑖 = 𝑉𝑖 𝐼𝑖
Bus power equation can be obtained from above two equation
𝑃𝑖 − 𝑗𝑄𝑖 = 𝑉𝑖 𝐼𝑖 = 𝑉𝑖 * ∑𝑁
𝑛=1 𝑌𝑖𝑛 𝑉𝑛
= | 𝑉𝑖 | < −𝛿𝑖 ∑𝑁
𝑛=1 | 𝑌𝑖𝑛 |∠ 𝜃𝑖𝑛 |𝑉𝑛 | < 𝛿𝑛
𝑃𝑖 − 𝑗𝑄𝑖 = ∑𝑁
𝑛=1 |𝑉𝑖 |𝑉𝑛 | | 𝑌𝑖𝑛 | ∠ 𝜃𝑖𝑛 + 𝛿𝑛 −𝛿𝑖
Separating the real and imaginary parts, we obtain
𝑃𝑖 = ∑𝑁
𝑛=1 |𝑉𝑖 |𝑉𝑛 | | 𝑌𝑖𝑛 | cos( 𝜃𝑖𝑛 + 𝛿𝑛 −𝛿𝑖 )
𝑄𝑖 = - ∑𝑁
𝑛=1 |𝑉𝑖 |𝑉𝑛 | | 𝑌𝑖𝑛 | cos( 𝜃𝑖𝑛 + 𝛿𝑛 −𝛿𝑖 )
The real and reactive power obtained from above two equations calculated power during the power
flow calculations, their value depends on the latest bus voltage.
Finally, the non linear equations to be solved in power flow analysis are
∑𝑁
𝑛=1 |𝑉𝑖 |𝑉𝑛 | | 𝑌𝑖𝑛 | cos( 𝜃𝑖𝑛 + 𝛿𝑛 −𝛿𝑖 ) = 𝑃𝐼𝑖
- ∑𝑁
𝑛=1 |𝑉𝑖 |𝑉𝑛 | | 𝑌𝑖𝑛 | sin( 𝜃𝑖𝑛 + 𝛿𝑛 −𝛿𝑖 ) = 𝑄𝐼𝑖
The N total number of buses in the power system , let the number of P -Q buses be N1, and P-V
buses be N2. Then N = N1 + N2
The basic problem is to find the

1) Unknown phase angle 𝛿 at the N1 and N2 number of PQ and PV buses and
2) Unknown voltage magnitude |v| at N1 number of P-Q buses.
Thus, total number of unknown variable is
2N1+N2
We can write N1+N2 real power specification equation and
22
N1 is reactive power specification equation
Thus total number of equations = 2N1+N2
Therefore, number of equation = number of variable = 2N1 +N2
Thus in power flow study, we need to solve the equations

∑𝑁
𝑛=1 |𝑉𝑖 |𝑉𝑛 | | 𝑌𝑖𝑛 | cos( 𝜃𝑖𝑛 + 𝛿𝑛 −𝛿𝑖 ) = 𝑃𝐼𝑖
For i=1,2---------------------N
- ∑𝑁
𝑛=1 |𝑉𝑖 |𝑉𝑛 | | 𝑌𝑖𝑛 | sin( 𝜃𝑖𝑛 + 𝛿𝑛 −𝛿𝑖 ) = 𝑄𝐼𝑖
For i=1,2---------------------N
For unknown variable
𝛿𝑖 = For i=1,2---------------------N
| 𝑣𝑖 | = For i=1,2---------------------N
The unknown variables are also called as state variables
COMPARISON OF GAUSS-SEIDEL AND NEWTON-RAPHSON METHOD
PROPERTIES GS METHOD NR METHOFD

Accuracy Lesser Higher
Convergence rate Slow Fast
Number of iterations iterations improves with the Remains fixed
size of power system
Type of convergence Linear Quadratic Stability

Stability For small power system For large power system
Effect of slack bus selection More dependent Less dependent

Guarantee of convergence Not guaranteed Guaranteed
Acceleration External Internal
Time for each iteration Less More
Table 3.1 comparison of gauss- sedial and newton Raphson method

23
OTHER CONVENTIONAL METHOD ARE :
3.1.2 Gradient Method:

The Generalized Reduced Gradient is applied to the OPF problem with the main motivation being
the existence of the concept of the state and control variables, with load flow equations providing a
nodal basis for the elimination of state variables. With the availability of good load flow packages,
the sensitivity information needed is provided. This, in turn, helps in obtaining a reduced problem in
the space of the control variables with the load flow equations and the associated state variables
eliminated.
3.1.3 Linear Programming Method :
Linear Programming (L.P) method treats problems having constraints and objective functions
formulated in the linear form with non-negative variables. Basically, the simple method is well
known to be very effective for solving LP problems.
The Linear Programming approach has been advocated on the grounds that
(a) The L.P solution process is completely reliable.
(b) L.P solutions can be very fast.
(c) The accuracy and scope of the linearized model is adequate for most engineering purposes.
It may be noted that point (a) is certainly true while point (b) depends on the specific algorithms and
problem formulations. The observation (c) is frequently valid since the transmission network is
quasi-linear, but it needs to be checked out for any given system and application.
3.1.4 Quadratic Programming Method:

Quadratic Programming (QP) is a special form of NLP. The objective function of QP optimization
model is quadratic and the constraints are in linear form. Quadratic Programming has higher
accuracy than LP – based approaches. Especially the most often used objective function is a
quadratic.
The NLP having the objective function and constraints described in Quadratic form is having a lot
of practical importance and is referred to as quadratic optimization. The special case of NLP where
the objective function is quadratic (i.e. is involving the square, cross product of one or more
variables) and constraints described in linear form is known as quadratic programming. Derivation
of the sensitivity method is aimed at solving the NLP on the computer. Apart from being a common
form for many important problems, Quadratic Programming is also very important because many of
the problems are often solved as a series of QP or Sequential Quadratic Programming (SQP)
problems.
Quadratic Programming based optimization is involved in power systems for maintaining a desired
voltage profile, maximizing power flow and minimizing generation cost. These quantities are
24
generally controlled by complex power generation which is usually having two limits. Here
minimization is considered as maximization can be determined by changing the sign of the objective
function. Further, the quadratic functions are characterized by matrices and vectors.
3.1.5 Interior Point Method :
It has been found that, the projective scaling algorithm for linear programming proposed by N.
Karmarkar is characterized by significant speed advantages for large problems reported to be as
much as 12:1 when compared to the simplex method Further, this method has a polynomial bound
on worst-case running time that is better than the ellipsoid algorithms. Karmarkar’s algorithm is
significantly different from Dantzig’s simplex method. Karmarkar’s interior point rarely visits too
many extreme points before an optimal point is found. In addition, the IP method stays in the interior
of the polytope and tries to position a current solution as the “center of the universe” in finding a
better direction for the next move. By properly choosing the step lengths, an optimal solution is
achieved after a number of iterations. Although this IP approach requires more computational time
in finding a moving direction than the traditional simplex method, the better moving direction is
achieved resulting in less iteration. In this way, the IP approach has become a major rival of the
simplex method and has attracted attention in the optimization community. Several variants of
interior points have been proposed and successfully applied to optimal power flow.
The Interior Point Method is one of the most efficient algorithms. The IP method classification is a
relatively new optimization approach that was applied to solve power system optimization problems;
it solves a large scale linear programming problem by moving through the interior, rather than the
boundary as in the simple method, of the feasible reason to find an optimal solution. The IP method
was originally proposed to solve linear programming problems; however, later it was implemented
to efficiently handle quadratic programming problems.
3.1.6 disadvantages of conventional method:
In most cases, mathematical formulations have to be simplified to get the solutions because of the
extremely limited capability to solve real-world large scale power system problems. They are weak
in handling qualitative constraints. They have poor convergence, may get stuck at local optimum,
they can find only a single optimized solution in a single simulation run, they become too slow if the
number of variables is large and they are computationally expensive for a solution of a large system.
25
3.2 Intelligent Methods:
The major advantage of the intelligent method is that they are relatively versatile for handling various
qualitative constraints. These methods can find multiple optimal solutions in a single simulation run.
So they are quite suitable for solving the multi-objective optimization problem. In most case, they
can find the global optimum solution.
The main advantages of Intelligent methods are:
Possesses learning ability, fast, appropriate for non-linear modelling.
3.2.1 PARTICLE SWARM OPTIMIZATION METHOD
The PSO is a relatively new and powerful intelligence evolution algorithm for solving optimization
problems. It is a population-based approach.
Particle swarm optimization (PSO) is a population-based stochastic optimization technique inspired
by social behavior of bird flocking or fish schooling. was first developed in 1995 by Dr. James
Kennedy and Dr. Russell Eberhart.
In PSO, the search for an optimal solution is conducted using a population of particles, each of which
represents a candidate solution to the optimization problem. Particles change their position by flying
around a multidimensional space by following current optimal particles until a relatively unchanged
position has been achieved or until computational limitations are exceeded. Each particle adjusts its
trajectory towards its own previous best position and towards the global best position attained till
them.
PSO is easy to implement and provides fast convergence for many optimization problems and has
gained a lot of attention in power system applications recently. The system is initialized with a
population of random solutions and searches for optima by updating generations. However, unlike
GA, PSO has no evolution operators such as crossover and mutation. In PSO, the potential solutions,
called particles, fly through the problem space by following the current optimum particles.In PSO,
each particle makes its decision using its own experience together with it is a neighbor’s experience.
The aim of this method is to find the best perfxorming individual in the whole group. It is used to
solve wider ranges of complex different optimization problems such as function minimization and
maximization.
26
Fig 3.2 Example on the flock of bird in nature (ref. book 12)
Fig 3.3 Example of school of fish in nature (ref. book 12)
27
Particle: An individual within the swarm.
Position: A particle’s dimensional coordinates representing the solution to the Problem.
Swarm: Total population of particles.
Fitness: A function that gives the interface between the optimization problem and that physical
problem, and amplifies the accuracy of the solution within the position in the solution space.
𝑷𝒃𝒆𝒔𝒕 : The position in the parameter space of the best fitness returned for a specific particle.
𝑮𝒃𝒆𝒔𝒕 : The position in the parameter space of the best fitness returned for the entire swarm
𝑽𝒎𝒂𝒙 : The maximum
𝑽𝒎𝒊𝒏 : The minimum allowable value for the velocity in a given direction.
3.2.2 Merits of particle swarm optimization:

1. PSO is one of the modern heuristic algorithms capable to solve large-scale non-convex
optimization problems like OPF.
2. The main advantages of PSO algorithms are a simple concept, easy implementation, relative
robustness to control parameters and computational efficiency.
3. The prominent merit of PSO is its fast convergence speed.
4. PSO algorithm can be realized simply for less parameter adjusting.
5. PSO can easily deal with nondifferentiable and nonconvex objective functions.
6. PSO has the flexibility to control the balance between the global and local exploration of the
search space.
3.2.3 PSEUDOCODE:
For each particle
Initialize particle
Do until maximum interaction or minimize error critical
For each Particle
Calculate data fitness value If the fitness value is better than p best
28
Set p best = current fitness value
If p best is better than g best
Set g best =p best
For each particle

Calculate particle velocity
Us e g best and velocity to update particle data
3.2.4 Comparison between classical and Artificial Intelligence (AI) method :
In the classical method, they had made excellent advancements. However, they suffer from the
following disadvantages which are, in most cases, mathematical formulations have to be simplified
in order to get the solutions because of the extremely limited capability to solve real-world large-
scale power system problems. They are weak in handling qualitative constraints and also have poor
convergence. Because of these disadvantages, the classical method may get stuck at local optimum
which they can find only a single optimized solution in a single simulation run. Besides, they become
too slow if the number of variables is large and they are computationally expensive for the solution
of a large system.
The major advantage of the AI methods is that they are relatively versatile for handling various
qualitative constraints. AI methods can find multiple optimal solutions in a single simulation run. So
they are quite suitable for solving multi-objective optimization problems. In most cases, they can
find the global optimum solution.
29
3.2.5 PARTICLE SWARM OPTIMIZATION ALGORITHM:
Initial population
Calculate fitness value for each

particle
Is current fitness value better then

p best?
Yes no
Current fitness value as Keep previous p best

new p best
Assign best particle p best value to g

best
Calculate velocity for each particle
Use each particles velocity value to

update it data values
no
Target or maximize
reached?
yes
stop
FIG 3.4 FLOW CHART OF PARTICLE SWARM OPTIMIZATION
30
CHAPTER : 4
NEWTON’S RAPHSON METHOD (CONVETIONAL METHOD) FOR

LOSS MINIMIZATION
4.1 newton Raphson method data
Total bus 30
Total transmission line 37
Total transformer 4
Total PV data 5
Total PQ data 21
Tolerance 0.001
Jacobian matrix 22*22
4.2 COMPUTATIONAL ALGORITHMS FOR NEWTON RAPHSON LOAD FLOW

METHOD
Step 1: Form the nodal admittance matrix (𝒀𝒊𝒋 ).
Step 2: Assume an initial set of bus voltage and set bus n as the reference bus as:
𝑉𝑖 =𝑉𝑖 , spec∠ 00 (at all PV buses)
𝑉𝑖 = 1∠00 (at all PQ buses)
Step 3: Calculate the real Power 𝑷𝒊 using the load flow equation:
|𝑉|2 𝐺𝑖𝑖 + ∑𝑁𝑛=1 |𝑉𝑖 | |𝑉𝑛 ||𝑌𝑖𝑛 | cos(𝜃𝑖𝑛 + 𝛿𝑛 − 𝛿𝑖 ) = 𝑃𝑖
𝑛=≠𝑖
Step 4: Calculate the real Power 𝑸𝒊 using the load flow equation:
−|𝑉|2 𝐵𝑖𝑖 + ∑𝑁𝑛=1 |𝑉𝑖 | |𝑉𝑛 ||𝑌𝑖𝑛 | sin(𝜃𝑖𝑛 + 𝛿𝑛 − 𝛿𝑖 ) = 𝑄𝑖

𝑛=≠𝑖
Step 5: from the Jacobian matrix sub matrix H, N, K, and L can be calculated as follows:
𝐻𝑖𝑖 = −𝑄𝑖 − |𝑉𝑖 |2 𝐵𝑖𝑖
𝑁𝑖𝑖 = 𝑃𝑖 + |𝑉𝑖 |2 𝐺𝑖𝑖
𝑀𝑖𝑖 = 𝑃𝑖 − |𝑉𝑖 |2 𝐺𝑖𝑖
31
𝐿𝑖𝑖 = 𝑄𝑖 − |𝑉𝑖 |2 𝐵𝑖𝑖
𝐻𝑖𝑗 = −|𝑉𝑖 ||𝑉𝑗 ||𝑌𝑖𝑗 | sin(𝜃𝑖𝑗 + 𝛿𝑗 − 𝛿𝑖 )
𝑁𝑖𝑗 = |𝑉𝑖 ||𝑉𝑗 ||𝑌𝑖𝑗 | cos(𝜃𝑖𝑗 + 𝛿𝑗 − 𝛿𝑖 )
𝑀𝑖𝑗 = −|𝑉𝑖 ||𝑉𝑗 ||𝑌𝑖𝑗 | cos(𝜃𝑖𝑗 + 𝛿𝑗 − 𝛿𝑖 )
𝐿𝑖𝑗 = −|𝑉𝑖 ||𝑉𝑗 | |𝑌𝑖𝑗 |sin(𝜃𝑖𝑗 + 𝛿𝑗 − 𝛿𝑖 )
The dimension of the sub matrix are as follows:

H (𝑁1 + 𝑁2 ) × (𝑁1 + 𝑁2 )
N (𝑁1 + 𝑁2 ) × 𝑁1
M 𝑁1 × (𝑁1 + 𝑁2 ) and
L 𝑁1 × 𝑁1
Where 𝑁1 is the number of P-Q buses 𝑁2 is P-V buses .
Step 6:Find the power differences Δ𝑷𝒊 and Δ𝑸𝒊 for all i=1, 2, 3… (n-1);
Δ𝑃𝑖 = 𝑃𝑖,𝑠𝑝𝑒𝑐, - 𝑃𝑖,𝑐𝑎𝑙,
Δ𝑄𝑖 = 𝑄𝑖,𝑠𝑝𝑒𝑐, - 𝑄𝑖,𝑐𝑎𝑙,
Step 7: Choose the tolerance values.

Step 8: Stop the iteration if all Δ𝑷𝒊 and Δ𝑸𝒊 are within the tolerance values.
Step 9: Update the values of 𝑽𝒊 and Δ𝜹𝒊 using the equation 𝑿𝒌+𝟏 =𝑿𝒌 + Δ𝑿𝒌 .
32
4.2 FLOW CHART FOR NEWTON RAPHSON LOAD FLOW METHOD:
In context to various steps involved in carrying out load flow studies with Newton Raphson
method, following detailed flow chart has been designed:
start
↓
Read load data: primitive matrix, slack bus voltage, real and reactive bus power etc.
↓
Formulate Y-matrix using load flow data
↓
Make initial assumption i.e, Vi and 𝜃i for i=1,2,……,n
↓
Set iteration count r=0
↓
Compute pi and qi for i=1,2,…..n using power flow equations
↓
Evaluate power mismatches i.e Δ𝑃𝑖 and Δ𝑞𝑖
Are
Max Δ𝑃𝑖 >𝜖

r=r+1
Max Δ𝑄𝑖 >𝜖
Calculate line flow

and
↓ Line loss
Evaluate Jacobian matrix elements
↓
ΔP 𝑗1 𝑗2 Δ𝜃 Output load flow
solution. i.e,
ΔQ = 𝑗3 𝑗4 Δ𝑉 voltage magnitude,
phase angle, etc.
↓ ↓
Evaluate increments and update variable stop
ΔQ
Fig 4.1 flow chart of newton Raphson method for power loss (ref. book 12)
33
CHAPTER :5
PARTICLE SWARM OPTIMIZATION ( ARTIFICIAL INTELLIGENCE

METHOD) FOR LOSS MINIMIZATION
5.1 MATLAB SIMULATION OF PSO FOR MINIMIZE POWER LOSS
The basic Optimal Power flow (OPF) problem can be described mathematically as a minimization
of the problem of Minimizing the total fuel cost of all committed plants subject to the constraints.
Minimize∑𝑛𝑖=1 𝐹𝑖 (𝑃𝑖 ) (A1)
F( Pi) is the fuel cost equation of the 𝑖 𝑡ℎ plant. It is the variation of fuel cost ($ or Rs) with
generated power (MW).Normally it is expressed as continuous quadratic equation.
𝐹𝑖 (𝑃𝑖 ) = 𝑎𝑖 𝑃𝑖2 + 𝑏𝑖 𝑃𝑖 + 𝑐𝑖 , 𝑃𝑖𝑚𝑖𝑛 ≤ 𝑃𝑖 ≤ 𝑃𝑖𝑚𝑎𝑥 (A2)
The total generation should meet the total demand and transmission loss. The transmission loss can
be determined from power flow.
∑𝑛𝑖=1 𝑝𝑖 = D +𝑃𝑖 (A3)
𝑃𝑖 = real (∑𝑛𝑗 𝑉𝑖 𝑌𝑖𝑗∗ 𝑉𝑖 ), i=1,2…..n (A4)
𝑄𝑖 = image (∑𝑛𝑗 𝑉𝑖 𝑌𝑖𝑗∗ 𝑉𝑖 ), i=1,2…..n (A5)
𝑉𝑖𝑚𝑖𝑛 ≤ 𝑉𝑖 ≤ 𝑉𝑖𝑚𝑎𝑥 (A6)

𝐿𝐹𝑖𝑗 ≤ Line flow limits (A7)
34
5.2 SOLUTION BY PSO USING PSO ALGORITHM
1. Collect the bus data, line data and cost coefficients, and their limits.
2. Convert the constrained optimization problem as an unconstrained problem by the penalty
function method.
Minimize
∑𝑛𝑖=1 𝐹𝑖 (𝑃𝑖 ) +1000 * abs( ∑𝑛𝑖=1 𝑃𝑖 – D- 𝑃𝑖 )
3. The file of Data will be run for Data load in MATLAB Simulation and then run PSO file.
5.3 THE PROBLEM FORMULATION:
The LM loss mode of the OPF problem can be formulated as follows

Minimize f(x, u)
subject to g(x, u) = 0, equality constraint
and h(x) ≤ 0, inequality constraint (ref. paper 4)
where,
f is the objective function to be minimized.
x is the vector of the dependent variable consisting of
1) Generator active power output stack bus 𝑃𝐺1
2) Load bus voltage 𝑉𝐿
3) Generator reactive power output 𝑄𝐺
4) Transmission Line Loadings 𝑆𝑙
Hence, x can be expressed as given below
𝑋 𝑇 =[ 𝑃𝐺1 , 𝑉𝐿1 − − − 𝑉𝐿𝑁𝐿 , 𝑄𝐺1 − − − 𝑄𝐺𝑁𝐺 , 𝑆𝑙 − − − 𝑠𝑙𝑛𝑙 ,]
Where,
NL,NG and NL are number of load buses, number of generators and number of transmission line
respectively
u is the vector of independent variable consisting of
1) Generator bus voltage 𝑉𝐺
2) Generator active power output at PG at PV buses except at the stack bus 𝑃𝐺1
3) Transformer Tap setting T
35
4) Shunt VAR compensation 𝑄𝐶
Hence, u can be expressed as given below
𝑢𝑇 = [ 𝑉𝐺1 − − − 𝑉𝐺𝑁𝐺 , 𝑃𝐺1 − − − 𝑃𝐺𝑁𝐺 , 𝑇1 − − − 𝑇𝑁𝑇 , , 𝑄𝐶1 − − − 𝑄𝐶𝑁𝐶 ]
Where NT and NC are the number of the regulating transformers and shunt compensators,
respectively
g is the equality constraint represents the typical load flow equation

𝑃𝐺𝑖 - 𝑃𝐷𝑖 - ∑𝑁𝐵
𝑗=1 |𝑉𝑖 | |𝑉𝑗 | |𝑉𝑖𝑗 | cos(𝜃𝑖𝑗 -𝛿𝑖 + 𝛿𝑗 ) ) = 0 (ref. paper 4)
𝑄𝐺𝑖 - 𝑄𝐷𝑖 - ∑𝑁𝐵

𝑗=1 |𝑉𝑖 | |𝑉𝑗 | |𝑉𝑖𝑗| sin(𝜃𝑖𝑗 -𝛿𝑖 + 𝛿𝑗 ) ) = 0 (ref. paper 4)
The solution process of the LM consists of optimizing the objective function and satisfying the
following constraints:
1) Power flow equations;
2) Branch flow limits;
3) Bus voltage limits;
4) Control variable limits.
Stopping criteria :
The stopping criterion mentioned in the aforementioned algorithm depends on the type of problem
being solved. Usually, the algorithm is run for a fixed number of iterations (objective function
evaluations) or until a specified error bound is reached.
5.4 OBJECTIVE FUNCTION FOR POWER LOSSES OPTIMIZATION
The optimal reactive power dispatch problem is a complex optimization problem where a specific
objective function is minimized while satisfying a number of constraints.
The objective is basically to minimize the system’s total active power transmission losses by
optimally adjusting various control variables while satisfying a given set of constraints.
Objective Function:
The objective functions to be minimized are given by the sum of line losses
𝑃𝐿 = ∑𝑁𝑇𝐿
𝑘=1 𝑃𝑘 𝑙𝑜𝑠𝑠
Individual line losses 𝑃𝑘 𝑙𝑜𝑠𝑠 can be expressed in terms of voltages and phase angles as
36
2 2
𝑃𝑘 𝑙𝑜𝑠𝑠 = ∑𝑁𝑇𝐿
𝑘=1 𝐺𝐾 (𝑣𝑖 + 𝑣𝑗 - 2𝑣𝑖 𝑣𝑗 cos𝜃𝑖𝑗 )
The objective function can now be written as
2 2
Min 𝑃𝐿 = Min ∑𝑁𝑇𝐿 𝑁𝑇𝐿
𝑘=1 𝑃𝑘 𝑙𝑜𝑠𝑠 = ∑𝑘=1 𝐺𝐾 (𝑣𝑖 + 𝑣𝑗 - 2𝑣𝑖 𝑣𝑗 cos𝜃𝑖𝑗 ) (ref. paper 7)
Where,
𝑣𝑖 , 𝑣𝑗 : voltage of the i-th and the j-th bus,
NTL : number of transmission lines
Constraints for Objective Function of Power Loss Minimization:
There are many alternatives available for reducing losses at the distribution level: reconfiguration,
capacitor installation, load balancing, and the introduction of higher voltage levels.
The controllable system quantities are generator MW, controlled voltage magnitude, reactive power
injection from reactive power sources and transformer tapping.
The objective use herein is to minimize the power transmission loss function by optimizing the
control variables within their limits. Therefore, no violation of other quantities (e.g. MVA flow of
transmission lines, load bus voltage magnitude, generator MVAR) occurs in normal system
operating conditions.
The OPF problem solution aims at optimizing specific objective functions such as loss of power by
adjusting the power control variables and at the same time satisfying the equality and the inequality
constraints. The inequality constraints are the upper and the lower limits at the control and some
state variables, while the equality constraints are the power flow equations.
These are system constraints to be formed as equality and inequality constraints as shown below:
a) Equality Constraints
b) Inequality constraints
37
5.4.1 Equality Constraints:
System constraints are generator MW, controlled voltage, reactive power injection from
reactive power sources and transformer tapping. The objective is to minimize the power transmission
loss function by optimizing the control variables within their limits. Hence there will be no violation
of other quantities (e.g. MVA flow of transmission lines, load bus voltage magnitude, generator
MVAR) occurs in normal system operating conditions. These are system constraints to be formed as
equality and inequality constraints as follows.
𝑃𝐺𝑖 - 𝑃𝐷𝑖 - ∑𝑁𝐵

𝑗=1 |𝑉𝑖 | |𝑉𝑗 | |𝑉𝑖𝑗 | cos(𝜃𝑖𝑗 -𝛿𝑖 + 𝛿𝑗 ) ) = 0
𝑄𝐺𝑖 - 𝑄𝐷𝑖 - ∑𝑁𝐵

𝑗=1 |𝑉𝑖 | |𝑉𝑗 | |𝑉𝑖𝑗| sin(𝜃𝑖𝑗 -𝛿𝑖 + 𝛿𝑗 ) ) = 0 (ref. paper 7)
Where,
𝑃𝐺𝑖 = is the real power generation at bus i
𝑃𝐷𝑖 = is the real power demand at bus j
𝑄𝐺𝑖 = is the reactive power generation at bus i
𝑄𝐷𝑖 =is the reactive power demand at bus i
NB = is the total number of buses
𝜃𝑖𝑗 = is the angle of bus admittance element i,j
𝑉𝑖𝑗 = is the magnitude of bus admittance element i; j
5.4.2 Inequality Constraints

The inequality constraints are the system operating constraints. They are grouped into control
variables and state variables. The control variables are self-restricted and include the generator bus
voltages, the reactive power generated by the capacitor and the transformer tap settings. The state
variables are restricted by adding a quadratic penalty term to the objective function and include
reactive power generation, load bus voltages, active power generation at the slack bus, and line flow
limit. These constraints are given by
i. Generators real power outputs

𝑚𝑖𝑛 𝑚𝑎𝑥
𝑃𝐺𝑖 ≤ 𝑃𝐺𝑖 ≤ 𝑃𝐺𝑖 i = 1,2,L, ,, NG
ii. The voltage magnitude limits for each bus
𝑣𝑖𝑚𝑖𝑛 ≤ 𝑣𝑖 ≤ 𝑣𝑖𝑚𝑎𝑥 , i = 1,2,L, ,, NB

where 𝑁𝐵 is the total number buses
iii. The reactive power generation limit for each generator bus is given by
𝑄𝐺𝑖 ≤ 𝑄𝐺𝑖 ≤ 𝑄𝐺𝑖 i = 1,2,L, ,, NG
38
where 𝑁𝑔 is the total number of generator buses
iv. The reactive power output of compensators

𝑄𝑐𝑖 ≤ 𝑄𝑐𝑖 ≤ 𝑄𝑐𝑖 i = 1,2,L, ,, NC
v. The transformer tap-setting constraint
𝑇𝑘𝑚𝑖𝑛 ≤ 𝑇𝑘 ≤ 𝑇𝑘𝑚𝑎𝑥
vi. The power flow limit constraint of each transmission line
𝑠𝑙 ≤ 𝑠𝑙𝑚𝑎𝑥 i = 1,2,L, ,, NTL (ref. paper 7)

Where,
𝑣𝑖𝑚𝑖𝑛 𝑎𝑛𝑑 𝑣𝑖𝑚𝑎𝑥 = upper and lower limits of voltage magnitude of bus i
𝑇𝑘𝑚𝑖𝑛 𝑎𝑛𝑑 𝑇𝑘𝑚𝑎𝑥 = are upper and lower limit of tap positions of transformer k
𝑄𝑐𝑖 𝑎𝑛𝑑 𝑄𝑐𝑖 = upper and lower limits of reactive power source i
NG , NL : number of generator and load buses respectively,

NTL : number of transmission lines
NT : number of regulating transformers,
NC : number of shunt compensators,
The scripts “min” and “max” denote the corresponding lower and upper limits, respectively.
5.4.3 Formation of the Fitness Function :
The fitness function FP is given by

𝑓𝑝 = ∑𝐾∈𝑁𝐸 𝑃𝑘 𝑙𝑜𝑠𝑠 + Penalty Function
where the Penalty Function is given by

𝑁 𝑁
𝐺
𝑘1 × ∑𝑖=1 𝑓 (𝑄𝑔𝑖 ) + 𝑘2 × ∑𝑁 𝐿
𝑖=1 𝑓 (𝑉𝑖 ) + 𝑘3 × ∑𝑚=1 𝑓 (𝑆𝑙𝑚 )
and 𝑘1, 𝑘2, 𝑘3 are constants called penalty factor.
39
0 if 𝑥 𝑚𝑖𝑛 ≤ 𝑥 ≤ 𝑥 𝑚𝑎𝑥
f (x) = ( 𝑥 − 𝑥 𝑚𝑖𝑛 )2 if 𝑥 > 𝑥 𝑚𝑎𝑥
( 𝑥 − 𝑥 𝑚𝑖𝑛 )2 if 𝑥 > 𝑥 𝑚𝑖𝑛
In x represents all control variables, 𝑥 𝑚𝑖𝑛 and 𝑥 𝑚𝑎𝑥 represent the minimum and maximum limits
for all the control variables. The penalty function guarantees that, if there is violation of the system
in case the control variables exceed their limits, the fitness function satisfies the inequality
constraints.
5.4 PARTICLE SWARM OPTIMIZATION CHARACTERISTICS :

1. Objective and constraints function can be non-differentiable
2. Constraints non-linear (non-linear mean high complexity)
3. Discrete/discontinuous search space.
4. Mixed variables (integer, real, Boolean, etc.)
5. A large number of constraints and variable.
6. The objective function can be multimodal.
(multimodal function have more than one optima)
7. Computationally expensive objective functions and constraints
Eg. If the objective function is 1 and constraints are 45, it may have lots of computationally
Steps.
If we satisfied all constraints than we can find objective functions.
40
5.5 MODIFICATION CONCEPT OF SEARCHING POINT BY PSO:
The basic concept of PSO lies in accelerating each particle toward its 𝑃𝑏𝑒𝑠𝑡 and the 𝐺𝑏𝑒𝑠𝑡 locations.
with a random weighted acceleration at each time step as shown in the figure.
Fig 5.1 modification concept of a searching point by PSO (ref. paper 2)

Where,
𝑆 𝐾 = current searching point
𝑆 𝐾+1 = modified searching point
𝑉 𝐾 = current velocity
𝑉 𝐾+1 = modified velocity
𝑉𝑃𝑏𝑒𝑠𝑡 = velocity based on 𝑃𝑏𝑒𝑠𝑡
𝑉𝑔𝑏𝑒𝑠𝑡 = velocity based on 𝑔𝑏𝑒𝑠𝑡
5.5.1 VELOCITY UPDATE MACHANIM
Each particle tries to modify its position using the following information:
The current positions,

The current velocity,
The distance between the current position and 𝑃𝑏𝑒𝑠𝑡
41
The distance between the current position and 𝐺𝑏𝑒𝑠𝑡
The modification of the particle’s position can be mathematically modelled according the following
equation:
𝑉𝒊𝒌+𝟏 = w𝑉𝒊𝒌 + 𝑐1 𝑟𝑎𝑛𝑑1 (.) × (𝑝𝑏𝑒𝑠𝑡𝑖 - 𝑠𝑖𝑘 ) + 𝑐2 𝑟𝑎𝑛𝑑2 (.) × (𝑔𝑏𝑒𝑠𝑡𝑖 - 𝑠𝑖𝑘 ) (ref. paper 7)
Where,
𝑉𝒊𝒌+𝟏 = The velocity of 𝑖 𝑡ℎ particle at 𝑘 + 1𝑡ℎ iteration,
𝑉𝒊𝒌 = The velocity of 𝑖 𝑡ℎ particle at 𝑘 𝑡ℎ iteration

𝑐1 , 𝑐2 =Positive constants having values between [0, 2.5]
rand = uniformly distributed random number between 0 and 1,
w = Inertia weight of the particle,
𝑠𝑖𝑘 = The position of 𝑖 𝑡ℎ particle at 𝑘 𝑡ℎ iteration
The following weighting function is usually utilized:
𝑣max − 𝑤𝑚𝑖𝑛
w = 𝑤𝑚𝑎𝑥 - × iteration (ref. paper 1)
𝑖𝑡𝑒𝑟𝑚𝑎𝑥
A large inertial weight (w) facilitates a global search while a small inertia weight facilitates a local
search. By linearly decreasing the inertia weight from a relatively large value to a small value
through the course of the PSO run gives the best PSO performance compared with inertial weight
settings.
Large w --------------------- greater global search ability

Smaller w ------------------- greater local search ability.
Where,
𝑤𝑚𝑎𝑥 = initial weight
𝑤𝑚𝑖𝑛 = final weight
Max. iteration = maximum iteration number
Iteration = current iteration number
5.5.2 CURRENT POSITION CAN MODIFEIED
42
𝑆𝒊𝒌+𝟏 = 𝑆𝑖𝑘 + 𝑉𝒊𝒌+𝟏 (ref. paper 2)
Where,
𝑆 𝐾 = current searching point
𝑆 𝐾+1 = modified searching point
𝑉𝒊𝒌+𝟏 = new updated velocity of agent i at iteration k+1,
5.6 PSO ALGORITHM TO IEEE 30-BUS TEST SYSTEM
1. Define the control variables within their allowable range as well as the population size and the
number of iterations. Include the PSO parameters and define the data of the 14-bus test system.
2. Let iter = 0
3. Generate randomly the population of the particles as well as their velocities.
4. Run NR load flow for each particle to find out losses.
5. For each particle, calculate the fitness function using the equation
𝐹𝑝 = ∑𝑘𝜖𝑁𝐸 𝑃𝑙𝑜𝑠𝑠 + Penalty Function

𝑁 𝑁
𝐺
𝑘1 × ∑𝑖=1 𝑓 (𝑄𝑔𝑖 ) + 𝑘2 × ∑𝑁 𝐿
𝑖=1 𝑓 (𝑉𝑖 ) + 𝑘3 × ∑𝑚=1 𝑓 (𝑆𝑙𝑚 )
43
1. For all particles, find out from their fitness, their 𝑃𝑏𝑒𝑠𝑡 and 𝑔𝑏𝑒𝑠𝑡 particle.
2. Let iter = iter+1.
3. calculate each particle velocity and adjust it, if there is violation of its limits
4. Calculate the new position of each particle.
5. Run NR load flow for each particle to find out losses.
6. For each particle, calculate the fitness function using equation 𝐹𝑝 = ∑𝑘𝜖𝑁𝐸 𝑃𝑙𝑜𝑠𝑠 + Penalty Function

𝑁 𝑁
𝐺
𝑘1 × ∑𝑖=1 𝑓 (𝑄𝑔𝑖 ) + 𝑘2 × ∑𝑁 𝐿
𝑖=1 𝑓 (𝑉𝑖 ) + 𝑘3 × ∑𝑚=1 𝑓 (𝑆𝑙𝑚 )
7. If current fitness p for each particle is better than 𝑃𝑏𝑒𝑠𝑡 then set 𝑃𝑏𝑒𝑠𝑡 = p.
8. Let 𝑔𝑏𝑒𝑠𝑡 be set as best of 𝑃𝑏𝑒𝑠𝑡
9. Repeat at step 7 until the maximum number of iterations is completed
10. The optimized values of the control variables are given by the coordinates of 𝑔𝑏𝑒𝑠𝑡 particle and the
minimized value of the losses by the corresponding fitness function.
44
5.7 PARTICLE SWARM OPTIMIZATION FLOW CHART FOR MINIMIZE POWER LOSS
start
↓
Define the objective function to be
minimized
↓
Initialize the parameters of PSO, n,c1 ,c2, ITERmax, w, e
↓
Randomly allocate active power to the N units satisfying the equality and inequality constraints
↓
Iteration
↓ ITER =0
↓
Calculate the objective function
↓
Update the 𝐺𝑏𝑒𝑠𝑡 and 𝑝𝑏𝑒𝑠𝑡 values
↓
Update the position and velocity of the particles
45
↓
𝑉𝒊𝒌+𝟏 = w𝑉𝒊𝒌 + 𝑐1 𝑟𝑎𝑛𝑑1 (.) × (𝑝𝑏𝑒𝑠𝑡𝑖 - 𝑠𝑖𝑘 ) + 𝑐2 𝑟𝑎𝑛𝑑2 (.) × (𝑔𝑏𝑒𝑠𝑡𝑖 - 𝑠𝑖𝑘 )
↓
𝑆𝒊𝒌+𝟏 = 𝑆𝑖𝑘 + 𝑉𝒊𝒌+𝟏
↓
Iteration ITER =ITER+1
Check the stopping criteria

according the
no
error limit
yes ↓
𝐺𝑏𝑒𝑠𝑡 of PSO is the solution of problem
End
FIG 5.2 PARTICLE SWARM OPTIMIZATION FLOW CHART
(ref. book 12)
46
CHAPTER -6
IEEE 30 BUS TEST SYSTEM
The IEEE 30-bus test system is used in this work. In the IEEE 30-bus system, there are 30 buses, out
of which 5 are generator buses. Bus 1 is the slack bus and remaining are PQ load buses. The network
has 37 total transmission line and the total transformer is 4.
6.1 SYSTEM DIAGRAM
Fig 6.1 IEEE 30 Bus test system
6.2 TYPE OF BUSES WITH VARIABLE
TYEPE OF BUS KNOW VARIABLE UNKNOW VARIABLE
LOAD BUS PQ V𝛿
SLACK BUS V𝛿 PQ
GENERATOR BUS PV 𝑄𝛿
Table 6.1 Type of bus with variable (ref. paper 7)
47
CHAPTER -7
NR METHOD RESULTS
Detailed Report of Load flow (NR method)

------------------------------------------------------------
Iter. No = 1 Max. Reactive power mismatch at bus = 12, Max. mismatch = 0.07172308
Iter. No = 2 Max. Reactive power mismatch at bus = 2, Max. mismatch = 0.05997805
____________________________________________________________________________________
Converged load flow iteration No = 3, Max. mismatch = 0.00011139
-------------------------------------------------------------------------------------------------------------------------------
Q-limits at PV-buses accounted
------------------------------------------------------------------------------------------------------------------------------
Please wait...
Solution Converged !!!
Load flow results:

-------------------------------------------------------------------------------------------------------------------------------
Buse No V del PG QG PL QL
-------------------------------------------------------------------------------------------------------------------------------
1 1.060000 0.000000 2.609497 -0.167940 0.000000 0.000000
2 1.043138 -5.351888 0.400000 0.500111 0.217000 0.127000
3 1.020743 -7.531967 0.000000 0.000000 0.024000 0.012000
4 1.011766 -9.284072 0.000000 0.000000 0.076000 0.016000
5 1.010000 -14.165716 0.000000 0.368477 0.942000 0.190000
6 1.010258 -11.064566 0.000000 0.000000 0.000000 0.000000
7 1.002378 -12.865066 0.000000 0.000000 0.228000 0.109000
8 1.010000 -11.813251 0.000000 0.371421 0.300000 0.300000
9 1.050913 -14.108881 0.000000 0.000000 0.000000 0.000000
48
10 1.045128 -15.699603 0.000000 0.000000 0.058000 0.020000
11 1.082000 -14.108881 0.000000 0.161713 0.000000 0.000000
12 1.057120 -14.943262 0.000000 0.000000 0.112000 0.075000
13 1.071000 -14.943262 0.000000 0.106182 0.000000 0.000000
14 1.042281 -15.835401 0.000000 0.000000 0.062000 0.016000
15 1.037683 -15.927333 0.000000 0.000000 0.082000 0.025000
16 1.044391 -15.526291 0.000000 0.000000 0.035000 0.018000
17 1.039903 -15.861323 0.000000 0.000000 0.090000 0.058000
18 1.028154 -16.541643 0.000000 0.000000 0.032000 0.009000
19 1.025653 -16.715379 0.000000 0.000000 0.095000 0.034000
20 1.029738 -16.518817 0.000000 0.000000 0.022000 0.007000
21 1.032727 -16.142315 0.000000 0.000000 0.175000 0.112000
22 1.033258 -16.128079 0.000000 0.000000 0.000000 0.000000
23 1.027183 -16.318019 0.000000 0.000000 0.032000 0.016000
24 1.021584 -16.494598 0.000000 0.000000 0.087000 0.067000
25 1.017339 -16.066726 0.000000 0.000000 0.000000 0.000000
26 0.999661 -16.486383 0.000000 0.000000 0.035000 0.023000
27 1.023250 -15.542338 0.000000 0.000000 0.000000 0.000000
28 1.006818 -11.688353 0.000000 0.000000 0.000000 0.000000
29 1.003411 -16.772279 0.000000 0.000000 0.024000 0.009000
30 0.991936 -17.655102 0.000000 0.000000 0.106000 0.019000
-------------------------------------------------------------------------------------------------------------------------------
49
Line flows:
----------------------------------------------------------------------------------------------------------------------------
Line flows Line flows
_____________________ _______________________
From To P-flow Q-flow From To P-flow Q-flow
----------------------------------------------------------------------------------------------------------------------------
1 2 1.7323 -0.2134 2 1 -1.6804 0.3103
1 3 0.8772 0.0455 3 1 -0.8461 0.0242
2 4 0.4362 0.0395 4 2 -0.4261 -0.0475
3 4 0.8221 -0.0362 4 3 -0.8135 0.0521
2 5 0.8238 0.0182 5 2 -0.7943 0.0617
2 6 0.6034 0.0050 6 2 -0.5839 0.0146
4 6 0.7215 -0.1632 6 4 -0.7152 0.1761
5 7 -0.1477 0.1168 7 5 0.1494 -0.1331
6 7 0.3812 -0.0296 7 6 -0.3774 0.0241
6 8 0.2957 -0.0808 8 6 -0.2946 0.0754
9 11 -0.0000 -0.1571 11 9 0.0000 0.1617
9 10 0.2772 0.0591 10 9 -0.2772 -0.0511
12 13 0.0000 -0.1048 13 12 -0.0000 0.1062
12 14 0.0786 0.0240 14 12 -0.0778 -0.0225
12 15 0.1790 0.0680 15 12 -0.1768 -0.0637
12 16 0.0725 0.0335 16 12 -0.0719 -0.0324
14 15 0.0158 0.0065 15 14 -0.0158 -0.0064
16 17 0.0369 0.0144 17 16 -0.0369 -0.0141
15 18 0.0602 0.0160 18 15 -0.0598 -0.0152
18 19 0.0278 0.0062 19 18 -0.0278 -0.0061
19 20 -0.0672 -0.0279 20 19 0.0674 0.0283
10 20 0.0902 0.0371 20 10 -0.0894 -0.0353
10 17 0.0533 0.0442 17 10 -0.0531 -0.0439
10 21 0.1579 0.1001 21 10 -0.1567 -0.0977
50
10 22 0.0762 0.0460 22 10 -0.0757 -0.0449
21 22 -0.0183 -0.0143 22 21 0.0183 0.0143
15 23 0.0504 0.0291 23 15 -0.0501 -0.0285
22 24 0.0574 0.0306 24 22 -0.0569 -0.0299
23 24 0.0181 0.0125 24 23 -0.0180 -0.0124
24 25 -0.0120 0.0202 25 24 0.0121 -0.0200
25 26 0.0354 0.0237 26 25 -0.0350 -0.0230
25 27 -0.0476 -0.0037 27 25 0.0478 0.0041
27 29 0.0619 0.0167 29 27 -0.0610 -0.0151
27 30 0.0709 0.0166 30 27 -0.0693 -0.0136
29 30 0.0370 0.0061 30 29 -0.0367 -0.0054
8 28 -0.0054 -0.0040 28 8 0.0054 -0.0394
6 28 0.1867 -0.0003 28 6 -0.1861 -0.0109
6 9 0.2772 -0.0817 9 6 -0.2772 0.0980
6 10 0.1584 0.0016 10 6 -0.1584 0.0112
4 12 0.4421 0.1426 12 4 -0.4421 -0.0957
28 27 0.1806 0.0503 27 28 -0.1806 -0.0375
---------------------------------------------------------------------------------------------------------------------------
Total real power losses in the system = 0.175516
Total reactive power losses in the system = 0.077964
51
RESULTS OF PSO
PSO RESULTS ON BENCH MARK FUNCTION
PSO DATA
number of population size 401

Number of iteration 100
Number of decision variable 2
Inertia weight factor(w)
𝑊𝑚𝑎𝑥 1
𝑊𝑚𝑖𝑛 0.3
C1=C2 1.05
Table 5.1 PSO data
1. Benchmark function :
Sphere function
2
F(x)=∑𝑛𝑖=1 𝑥𝑖
𝑔𝑏𝑒𝑠𝑡 =
0.0020 -0.4534
Answer =
5.1615 11.0071 0.0000
52
2. Benchmark function :
Rosenbrock function
F(x , y)=∑𝑛−1 2 2 2
𝑖=1 [100 (𝑥𝑖+1 -𝑥𝑖 ) +(1-𝑥𝑖 ) ]
𝑔𝑏𝑒𝑠𝑡 =
0.9173 0.8690
Answer =
1.0000 1.0000 0.0000
53
54
SIMULATION RESULTS OF PSO O IEEE 30 BUS SYSTEM
This is the simulation result of IEEE 30 bus system.
SO: 1/200 iterations, 𝐺𝑏𝑒𝑠𝑡 = 805.09421750869865.

PSO: 20/200 iterations, 𝐺𝑏𝑒𝑠𝑡 = 801.89517133289905.
P1 =
48.8556
21.5005
21.6274
12.1322
12.0000
F1 =
801.8437
P=
55
176.6549 48.8556 21.5005 21.6274 12.1322 12.0000
VV =
Columns 1 through 12
1.0600 1.0430 1.0254 1.0171 1.0100 1.0148 1.0050 1.0100 1.0530 1.0467 1.0820 1.0599
1.0710 1.0450 1.0402 1.0471 1.0415 1.0304 1.0277 1.0317 1.0345 1.0350 1.0296 1.0237
1.0203 1.0027 1.0269 1.0128 1.0071 0.9957
TL =
9.3706
56
VV =
Columns 1 through 6
1.060000000000000 1.043000000000000 1.025413783394172 1.017131509026298 1.010000000000000

1.014769664042682
1.005038164105417 1.010000000000000 1.052965909519290 1.046719608823060 1.082000000000000

1.059937203976502
1.071000000000000 1.045026686493674 1.040245314359665 1.047141864538727 1.041528634572232

1.030400240859670
1.027704952651282 1.031677290240933 1.034456181153072 1.035022169486857 1.029553721518365

1.023697392194044
1.020347779790547 1.002724681831473 1.026853788152576 1.012841471195776 1.007092030932707

0.995661809041243
TL=
9.376133573332481
57
Simulation Results:-
58
59
60
61
P1 =
48.8126
21.4592
21.6837
12.0892
12.0000
F1 = 801.8436
P = 176.7310 48.8126 21.4592 21.6837 12.0892 12.0000
62
VV = Columns 1 through 14
1.0600 1.0430 1.0254 1.0171 1.0100 1.0148 1.0050 1.0100 1.0530 1.0467 1.0820
1.0599 1.0710 1.0450
1.0402 1.0471 1.0415 1.0304 1.0277 1.0317 1.0345 1.0350 1.0296 1.0237 1.0203
1.0027 1.0269 1.0128
1.0071 0.9957
TL =
9.3756
PSO Calculation:-

63
P1 =
167.2494
260.3704
130.2239
171.5403
80.0989
F1 =
1.5116e+004
P=
64
441.4061 167.2494 260.3704 130.2239 171.5403 80.0989
VV =
1.0250 1.0200 1.0450 1.0500 1.0450 1.0006 0.9953 0.9983 1.0103 0.9908 0.9979
0.9942 1.0225 1.0082
0.9990 0.9903 0.9823 1.0074 1.0045 0.9830 0.9770 0.9795 0.9780 0.9690 0.9750 1.0150
TL =
-12.1111
65
66
Penalty Function Calculation:-
tol1 = 10e-5; % tolerance or accuracy
tol2 = 10e-5;
R0 = 0.1; % initial Penalty parameter
x0 =[myrand(0,2) myrand(0,2)];
% x0 = [0 0];
c = input('enter 0 for interior penalty, 1 for exterior penalty = ');
if c==0
c = myrand(0,1); % c to update penalty parameter
else
c = myrand(0,10);
end
t = 1;
x(t,:)= x0;
R(t)= R0;
x1 = x(1);
x2 = x(2);
s = ((x1-5).^2 +x2.^2 -26); % constraint violation
if s < 0
for t = 1
x1 = x(t,1);
x2 = x(t,2);
s = ((x1-5).^2 +x2.^2 -26);
k = R.*s^2;
F = penfun(x(t,:),0);
P(1,t) = penfun(x(t,:),R(1,t));
G(t,:) = runevolution(x(t,:),R(1,t)); % to perform unconstrained search for

given R
x(t+1,:) = G(t,:);
t = t+1;
P(1,t) = penfun(G(t-1,:),R(1,t-1));
err = abs(P(1,t));
R(1,t) = c*R(1,t-1);
x1 = x(t,1);
x2 = x(t,2);
s = ((x1-5).^2 +x2.^2 -26);
67
end
else
x1 = x(t,1);
x2 = x(t,2);
k = 0;
F = penfun(x(t,:),0);
P(1,t) = penfun(x(t,:),0);
G(t,:) = runevolution(x(t,:),0);
t = t+1;
P(1,t) = penfun(G(t-1,:),0);
err = abs(P(1,t));
R(1,t) = c*R(1,t-1);
x(t,:) = G(t-1,:);
end
a(:,1) = G(t-1,:)';
while err>tol2
G(t,:) = runevolution(x(t,:),R(1,t));
x(t+1,:) = G(t,:);
t = t+1;
R(1,t) = c*R(1,t-1);
P(1,t) = penfun(G(t-1,:),R(1,t-1));
err = abs(P(1,t)-P(1,t-1));
x1 = x(t,1);
x2 = x(t,2);
s = ((x1-5).^2 +x2.^2 -26);
a(:,1) = G(t-1,:)';
if abs(s) <tol1
break;
elseif (abs(x(t,1)-x(t-1,1))<10e-5 && abs(x(t,2)-x(t-1,2))<10e-5)
break;
end
end
fval = penfun(x(t,:),0);
fprintf('The minimum point is (%4f,%4f) with a function value of %.4f \n',a,fval)
68
Voltage Losses Mitigation and Voltage Profile Improvement
val1 =
0.8936
index1 =
18
val2 =
0.6393
index2 =
18
Elapsed time is 0.723467 seconds.
ans =
1.0e+003 *
69
Columns 1 through 7
3.2645 - 1.6615i -3.2645 + 1.6641i 0 0 0 0 0
-3.2645 + 1.6641i 5.0854 - 3.1300i -0.6107 + 0.3111i 0 0 0 0
0 -0.6107 + 0.3111i 2.0064 - 1.1215i -0.8227 + 0.4190i 0 0 0
0 0 -0.8227 + 0.4190i 1.6128 - 0.8214i -0.7901 + 0.4024i 0 0
0 0 0 -0.7901 + 0.4024i 1.0554 - 0.6314i -0.2653 + 0.2290i 0
0 0 0 0 -0.2653 + 0.2290i 1.9183 - 1.5459i -0.1698 + 0.5614i
0 0 0 0 0 -0.1698 + 0.5614i 0.3155 - 0.6666i
0 0 0 0 0 0 -0.1457 + 0.1051i
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 -1.2102 + 1.1548i 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 -0.5729 + 0.3915i 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 -1.4832 + 0.7555i 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
70
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
-0.1457 + 0.1051i 0 0 0 0 0 0
0.3885 - 0.2796i -0.2428 + 0.1745i 0 0 0 0 0
-0.2428 + 0.1745i 0.4849 - 0.3467i -0.2421 + 0.1722i 0 0 0 0
0 -0.2421 + 0.1722i 1.9808 - 0.7471i -1.7387 + 0.5749i 0 0 0
0 0 -1.7387 + 0.5749i 2.6517 - 0.8768i -0.9130 + 0.3019i 0 0
0 0 0 -0.9130 + 0.3019i 1.0725 - 0.4274i -0.1595 + 0.1255i 0
0 0 0 0 -0.1595 + 0.1255i 0.4158 - 0.4628i -0.2562 + 0.3373i
0 0 0 0 0 -0.2562 + 0.3373i 0.6142 - 0.6559i
0 0 0 0 0 0 -0.3580 + 0.3186i
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
71
0 0 0 0 0 0 0
0 0 0 0 -1.2102 + 1.1548i 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
-0.3580 + 0.3186i 0 0 0 0 0 0
0.6894 - 0.5606i -0.3314 + 0.2420i 0 0 0 0 0
-0.3314 + 0.2420i 0.4371 - 0.3832i -0.1057 + 0.1412i 0 0 0 0
0 -0.1057 + 0.1412i 0.4265 - 0.3927i -0.3208 + 0.2515i 0 0 0
0 0 -0.3208 + 0.2515i 0.3208 - 0.2515i 0 0 0
0 0 0 0 1.3493 - 1.2802i -0.1391 + 0.1254i 0
0 0 0 0 -0.1391 + 0.1254i 0.5307 - 0.5828i -0.3916 + 0.4575i
0 0 0 0 0 -0.3916 + 0.4575i 0.5862 - 0.7148i
0 0 0 0 0 0 -0.1946 + 0.2574i
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
72
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 -0.5729 + 0.3915i 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 -1.4832 + 0.7555i 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
-0.1946 + 0.2574i 0 0 0 0 0 0
0.1946 - 0.2574i 0 0 0 0 0 0
0 0.8330 - 0.5969i -0.2601 + 0.2054i 0 0 0 0
0 -0.2601 + 0.2054i 0.5226 - 0.4108i -0.2625 + 0.2054i 0 0 0
0 0 -0.2625 + 0.2054i 0.2625 - 0.2054i 0 0 0
0 0 0 0 2.5428 - 1.2950i -1.0596 + 0.5395i 0
0 0 0 0 -1.0596 + 0.5395i 1.2611 - 0.7171i -0.2015 + 0.1776i
0 0 0 0 0 -0.2015 + 0.1776i 0.4695 - 0.4112i
0 0 0 0 0 0 -0.2681 + 0.2335i
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
73
0 0 0 0 0
0 0 0 0 0
0 0 0 0 0
0 0 0 0 0
0 0 0 0 0
0 0 0 0 0
0 0 0 0 0
0 0 0 0 0
0 0 0 0 0
0 0 0 0 0
0 0 0 0 0
0 0 0 0 0
0 0 0 0 0
0 0 0 0 0
0 0 0 0 0
0 0 0 0 0
0 0 0 0 0
0 0 0 0 0
0 0 0 0 0
0 0 0 0 0
0 0 0 0 0
0 0 0 0 0
0 0 0 0 0
0 0 0 0 0
0 0 0 0 0
0 0 0 0 0
0 0 0 0 0
-0.2681 + 0.2335i 0 0 0 0
0.8613 - 0.5357i -0.5933 + 0.3022i 0 0 0
-0.5933 + 0.3022i 0.7901 - 0.4968i -0.1969 + 0.1946i 0 0
0 -0.1969 + 0.1946i 0.7147 - 0.7981i -0.5178 + 0.6035i 0
0 0 -0.5178 + 0.6035i 0.8432 - 1.1095i -0.3254 + 0.5059i
0 0 0 -0.3254 + 0.5059i 0.3254 - 0.5059i
74
75
76
COMPERISION BETWEEN NR METHOD AND PSO METHOD
RESULTS
NR Method PSO Method
Total real power losses in the system = 0.175516
Total Loss = 9.3706

Total reactive power losses in the system = 0.077964
77
CHAPTER-7
CONCLUSION:
Particle swarm optimization for OPF problem has been proposed. The proposed approach has been
tested and examined on the standard IEEE 30 bus system. The simulation results show the effectiveness
and robustness of the proposed algorithm to solve the OPF problem. As a result, the PSO method
proves that it can find a place among the effective search method in order to find a global solution.
when comparing both the Newton Raphson method and particle swarm optimization method the
minimum losses are obtained from the particle swarm optimization method.
78
SYSTEM DATA
For the above case we will take reference IEEE 30 case data of the below models:
Bus data
Shunt capacitor data
Line data
All the data are shown with details in APPENDIX-A section 1.
79
References
1. “Particle Swarm Optimization Based Optimal Reactive Power Dispatch”, IEEE 2015,
Sundaram Pandya and Ranjit Roy.
2. “Optimal Power Flow Based on PSO in Electrical Power System”, International Journal of
Advances in Electrical and Electronics Engineering, ISSN: 2319-1112 Prashant Kumar,
Vikas Dubey, Deepak Sharma.
3. “Transmission Loss Minimization Using Optimization Technique Based On PSO ”,
IOSR Journal of Electrical and Electronics Engineering (IOSR-JEEE) Volume 6, Issue 1
(May. - Jun. 2013), PP 01-05. Sunil Joseph P, C.DineshBalaji.
4. “Constrained Optimal Power Flow using Particle Swarm Optimization”, International Journal
of Emerging Technology and Advanced Engineering (ISSN 2250-2459, Volume 2, Issue 2,
February 2012), C.Kumar, Dr. Ch. Padmanabha Raju.
5. “ Optimal power flow by particle swarm optimization with an aging leader and challengers”,
International Journal of Engineering, Science and Technology Vol. 7, No. 3, 2015, pp. 123-
132 Rudra Pratap Singh, V. Mukherjee, S.P. Ghoshal.
6. “particle swarm optimization of power loss in distribution system”, IEEE 2015 12^th
international conference on information technology – generation, Joseph B. abugri, and Marc
Karam.
7. “power loss reduction in a power system based on PSO: case study”, International Journal of
computer application (0975-8887) vol 164 – no 10, April 2017, Sameer Singh, Vivek Kumar
Jain, and Upendra Prasad.
8. “ Loss power minimization using particle swarm optimization”,2006 international joint
conference neural networks July 16-21, 2018, A.A.A. Esmin, and G.Lambert – Torres,
member, IEEE.
9. “Optimal reactive power flow control for minimization of active power losses using particle
swarm optimization”, 20125 conference on power control, communication and computational
technologies for sustainable growth (PCCCTSG) December 11-12,2015, Kurnool, Andhra
Pradesh, India, N. Tejswara Rao, Jagannath ch Yadav b. and anyapu jagannadham.
references of book
10. Artificial intelligent, Author: Elaine rich
11. 11.Modern power system analysis, Author: D.P. Kothari,I.J. Nagrath
12. A thesis of comparison between Newton Raphson method and particle swarm optimization
13. in the analysis of economic load dispatch, researched by rihab Hassan abdelghafour Hassan
80
APPENDIX-A section 1
IEEE 30 BUS DATA
TABLE A1 LOAD BUS DATA
Bus Load Bus no Load

no P(MW) Q(MW) P(MW) Q(MW)
1 0.00 0.00 16 3.50 1.80
2 21.7 12.7 17 9.00 5.80
3 2.40 1.20 18 3.20 0.90
4 7.60 1.60 19 9.50 3.40
5 94.2 19.0 20 2.20 0.70
6 0.00 0.00 21 17.5 11.2
7 22.8 10.9 22 0.00 0.00
8 30.0 30.0 23 3.20 1.60
9 0.00 0.00 24 8.70 6.70
10 5.80 2.00 25 0.00 0.00
11 0.00 0.00 26 3.50 2.30
12 11.2 7.50 27 0.00 0.00
13 0.00 0.00 28 0.00 0.00
14 6.20 1.60 29 2.40 0.90
15 8.20 2.50 30 10.6 1.90
TABLE A2 SHUNT CAPACITOR DATA
Bus no Susceptance
10 19
24 4
81
TABLE A3 TRANSMISSION LINE DATA
Line From To Series impedance (p.u) Half line Tap MVA

no bus bus R X charging setting Rating
susceptance
(p.u)
1 1 2 0.01920 0.05750 0.02640 0.300
2 1 3 0.04520 0.18520 0.02040 0.300
3 2 4 0.05700 0.17370 0.01840 0.300
4 3 4 0.01320 0.03790 0.00420 0.300
5 2 5 0.04720 0.19830 0.02090 0.300
6 2 6 0.05810 0.17630 0.01870 0.300
7 4 6 0.01190 0.04140 0.00450 0.300
8 5 7 0.04600 0.11600 0.01020 0.300
9 6 7 0.02670 0.08200 0.00850 0.300
10 6 8 0.01200 0.04200 0.00450 0.300
11 6 9 0.00000 0.20800 0.00000 1.0155 0.300
12 6 10 0.00000 0.55600 0.00000 0.9629 0.300
13 9 11 0.00000 0.20800 0.00000 0.300
14 9 10 0.00000 0.11000 0.00000 0.300
15 4 12 0.00000 0.25600 0.00000 1.0129 0.650
16 12 13 0.00000 0.14000 0.00000 0.650
17 12 14 0.12310 0.25590 0.00000 0.320
18 12 15 0.06620 0.13040 0.00000 0.320
19 12 16 0.09450 0.19870 0.00000 0.320
20 14 15 0.22100 0.19970 0.00000 0.160
21 16 17 0.08240 0.19320 0.00000 0.160
22 15 18 0.10700 0.21850 0.00000 0.160
23 18 19 0.06390 0.12920 0.00000 0.160
24 19 20 0.03400 0.06800 0.00000 0.320
25 10 20 0.09360 0.20900 0.00000 0.320
26 10 17 0.03240 0.08450 0.00000 0.320
27 10 21 0.03480 0.07490 0.00000 0.300
28 10 22 0.07270 0.14990 0.00000 0.300
29 21 22 0.01160 0.02360 0.00000 0.300
30 15 23 0.10000 0.20200 0.00000 0.160
31 22 24 0.11500 0.17900 0.00000 0.300
32 23 24 0.13200 0.27000 0.00000 0.160
33 24 25 0.18850 0.32920 0.00000 0.300
34 25 26 0.25440 0.38000 0.00000 0.300
35 25 27 0.10930 0.20870 0.00000 0.300
36 28 27 0.00000 0.36900 0.00000 0.9581 0.300
37 27 29 0.21980 0.41530 0.00000 0.300
38 27 30 0.32020 0.60270 0.00000 0.300
39 29 30 0.29990 0.45330 0.00000 0.300
40 8 28 0.06360 0.20000 0.02140 0.300
41 6 28 0.1690 0.05990 0.00650 0.300
82
APPENDIX-B
Review Card & Compliance Report
83
84
85
86
87
Compliance report
Internal Review
Comment 1:- understanding of optimization techniques done.
Comment 2:- understanding of reference test system done.
Comment 3 basic load flow programming on MATLAB done.
Comment 4:- Detail study of particle swarm optimization done.
Dissertation Phase -1
Comment 1:- Modification in a title is required “inertia study” is misleading done.
Comment 2:- fundamental of power flow should be improved improved
Comment 3:- implementation of PSO and comparison with conventional
methods are required to be shown improved.
Mid Semester Review

Comment 1:- Final objective is still not clear clear
Comment 2:- Poor concept of the optimization function improved
Comment 3: Results should be verified by leading your algorithm on some done
standard benchmark function
Comment 4: clearly define decision variables and pure penalty functions clear
88
APPENDIX-C
Paper Publication Certificate
89
90

DP 2 Tarun PDF

Diunggah oleh

Informasi Dokumen

Judul Asli

Hak Cipta

Format Tersedia

Bagikan dokumen Ini

Bagikan atau Tanam Dokumen

Opsi Berbagi

Apakah menurut Anda dokumen ini bermanfaat?

Apakah konten ini tidak pantas?

Hak Cipta:

Format Tersedia

DP 2 Tarun PDF

Diunggah oleh

Hak Cipta:

Format Tersedia

MINIMIZE POWER LOSSES USING

PARTICLE SWARM OPTIMIZATION

Under the guidance of

Electrical Engineering Department

Shri satsangi saketdham ram ashram group of institutions, vadasma

Signature and Name of Signature and Name of

Guide Head of department

Assistant prof. A.G. Acharya Assistant prof. A.G. Acharya

SSSRGI Vadasma SSSRGI Vadasma

Signature and Name of Signature and Name of

Guide Head of department

Assistant prof. A.G. Acharya Assistant prof. A.G. Acharya

SSSRGI Vadasma SSSRGI Vadasma

Candidate Name Signature and Name of Guide

Patel Tarunkumar D. Assistant prof. A.G. Acharya

Enrollment No.( 170910707018 ) SSSRGI Vadasma

Examiner’s Sign and Name:

Signature of Student: _____________ Signature of Guide: _____________

Patel Tarunkumar Dashrathbhai

CHAPTER 2 LITERATURE REVIEW 17

CHAPTER 3 POWER SYSTEM OPTIMIZATION AND POWER LOSSES 20

3.1 CONVENTIONAL METHOD 20

CHAPTER 5 PARTICLE SWARM OPTIMIZATION FOR 34

APPENDIX A SYSTEM DATA 81

APPENDIX B REVIEW CARD AND COPLIENCE REPORT 83

APPENDIX C PAPER PUBLICATION CERTIFICATE 89

Figure Name Page No.

FIG. 1.1 SIMPLE OPTIMIZATION 16

FIG. 3.1 OPF SOLUTION METHODOLOGIES 20

FIG. 3.2 EXAMPLE OF BIRD FLOCKING IN NATURE 27

FIG. 3.3 EXAMPLE OF SCHOOL OF FISH IN NATURE 28

FIG. 3.4 FLOW CHART OF PARTICLE SWARM 30

FIG. 4..1 FLOW CHART OF NEWTON RAPHSON METHOD 33

FIG. 5.1 MODIFICATION CONCEPT OF A SEARCHING 41

FIG 5.2 PARTICLE SWARM OPTIMIZATION FLOW CHART 45

FIG.6.1 IEEE 30 BUS TEST SYSTEM 47

FIGURE NAME PAGE NO.

TABLE 3.1 COMPARISON OF GAUSS SEDIAL AND 20

TABLE 6.1 TYPE OF BUSES WITH VARIABLE 47

TABLE 7.1 PSO RESULTS ON BECH MARK FUNCTION 65

TABLE 7.2 COMPERISION BETWEEN NR METHOD AND PSO

TABLE A.1 BUS DATA

TABLE A.2 SHUNT CAPACITOR DATA

TABLE A.3 TRANSMISSION LINE DATA

1.2 BASIC OF OPTIMIZATION

Set of in put process Best out put

Fig. 1.1 Simple optimization

1.3 OBJECTIVES OF PROJECT

3) “Constrained Optimal Power Flow using Particle Swarm Optimization”, International

5) “particle swarm optimization of power loss in distribution system”, IEEE 2015

7) “ Loss power minimization using particle swarm optimization”,2006 international joint

THE SOLUTION OF METHODOLOGIES

1. Conventional (classical) methods

OPF Solution mythologies

Conventional method intelligent method

Table 3.1 OPF Solution Methodologies (ref. book 12)

3.1 CONVENTIONAL METHOD

Power flow model of newton Raphson method:

The equation describing the performance of the network.

In expanding from these equations are

𝐼1 𝑌11 𝑌12 -------------------- 𝑌1𝑛 𝑉1

The complex power entering the network at bus i is given by

Separating the real and imaginary parts, we obtain

Signature of Student: _ Signature of Guide: _