Optimal Planning of Renewable based

DGs for Achieving Techno-Economic

Objectives across various Load Models

By

Habib Ur Rehman

Reg. No. 00000328570

Session 2020-22

Supervised by

Dr. Syed Ali Abbas Kazmi

US-Pakistan Center for Advanced Studies in Energy (USPCAS-E)

National University of Sciences and Technology (NUST) H-12,
Islamabad 44000, Pakistan
April 2023
Optimal Planning of Renewable based
DGs for Achieving Techno-Economic
Objectives across various Load Models

By

Habib Ur Rehman

Reg. No. 00000328570

Session 2020-22

Supervised by

Dr. Syed Ali Abbas Kazmi

A Thesis Submitted to the US-Pakistan Center for Advanced Studies

in

Energy in partial fulfilment of the requirements for the degree of

MASTER of SCIENCE

in

Electrical Engineering (Power)

US-Pakistan Center for Advanced Studies in Energy (USPCAS-E)
National University of Sciences and Technology (NUST) H-12,
Islamabad 44000, Pakistan
April 2023

i
 THESIS ACCEPTANCE CERTIFICATE

Certified that final copy of MS/MPhil thesis written by Mr. Habib Ur Rehman (Registration
No. 00000328570), of US-Pakistan Center for Advanced Studies in Energy been vetted by
undersigned, found complete in all respects as per NUST Statues/Regulations, is within the
similarity indices limit and is accepted as partial fulfilment for the award of MS/MPhil degree. It is
further certified that necessary amendments as pointed out by GEC members of the scholar have
also been incorporated in the said thesis.

Signature:

Name of Supervisor: Dr. Syed Ali Abbas Kazmi

Date:

Signature (HoD):

Date:

Signature (Dean/Principal):

Date:

ii
 Certificate

This is to certify that work in this thesis has been carried out by Mr. Habib Ur
Rehman and completed under my supervision in smart grid laboratory, US-Pakistan
Center for Advanced Studies in Energy (USPCAS-E), National University of Sciences
andTechnology, H-12, Islamabad, Pakistan.

Supervisor:
Dr. Syed Ali Abbas Kazmi
USPCAS-E
NUST, Islamabad

GEC member 1:
Dr. Hassan Abdullah Khalid
USPCAS-E
NUST, Islamabad

GEC member 2:
Dr. Muhammad Yousif
USPCAS-E
NUST, Islamabad

GEC member 3:
Dr. Muhammad Numan
USPCAS-E
NUST, Islamabad

HOD-EPE:
Dr. Hassan Abdullah Khalid
USPCAS-E
NUST, Islamabad

Dean/Principal:
Prof. Dr. Adeel Waqas Ahmad
USPCAS-E
NUST, Islamabad

iii
 Acknowledgement

I am very grateful to Allah Almighty, the Most Gracious, and the Most Merciful, for
enabling me to accomplish this study successfully. Then, I would like to say special
thanks and profound gratitude to my supervisor Dr. Syed Ali Abbas Kazmi who not
only guided me like a good mentor but also provided me with his valuable suggestions
and comments while conducting the study. Thanks a lot sir for your massive support
and sympathy shown to me all the time. God bless you. I would also like to thank the
members of my Guidance and Examination Committee, including Dr. Hassan
Abdullah Khalid, Dr. Muhamad Numan and Dr. Muhammad Yousif, for accepting to
review my thesis and for their suggestions, comments, and questions.
I also feel much indebted to my parents, whose prayers and good wishes have
supported me throughout my life and without them I’m worthless. Last words are
lacking to express my feelings and gratefulness to my affectionate family and friends
for their love and appreciations, which always supported and encouraged me mentally
during all years of my study.

iv
 Abstract
Distributed generation (DG), from the last few decades, has become the viable option
in distribution systems (DSs) to mitigate the power losses caused by substantial
increase in electricity demand and to improve the voltage profile by enhancing power
system reliability. In this study, two metaheuristic algorithms GTO and TDO are
presented to examine the utilization of DGs as well as the optimal sitting and sizing in
DS with special emphasis on maximizing voltage stability index, minimizing total
operating cost index and active power loss along with minimization of voltage
deviation. The robustness of the algorithms is examined on IEEE 33 bus and IEEE 69
bus radial distribution networks (RDNs) for PV and wind based DGs. The obtained
results are compared with the existing literature to validate the effectiveness of the
algorithms. The reduction of the active power loss is 93.15% and 96.87% of initial
value for 33 bus and for 69 bus RDN respectively while the other parameters i.e.,
operating cost index, voltage deviation and voltage stability index are also improved
that validate the efficiency of the algorithms. The proposed study is also carried out
by considering different voltage dependent load models like industrial, residential and
commercial.
Keywords: Artificial gorilla troops optimization; Distributed generation; Distributed
system; operating cost; Radial distribution network; Tasmanian devil optimization;
Voltage deviation; Voltage stability index

v
 Table of Contents
Abstract………………………. ................................................................................... v
Table of Contents ........................................................................................................ vi
List of Figures………… ...........................................................................................viii
List of Tables………................................................................................................... ix
List of Abbreviations................................................................................................... xi
List of Publications ....................................................................................................xii
Chapter 1: Introduction ............................................................................................ 1
1.1. Background ........................................................................................................ 1
1.2. Motivation ......................................................................................................... 2
1.3. Contributions ..................................................................................................... 3
1.4. Thesis Layout .................................................................................................... 3
Summary ..................................................................................................................... 5
References ................................................................................................................... 6
Chapter 2: Literature Review ................................................................................. 10
2.1 Background ....................................................................................................... 10
2.2 Analytical & heuristic Techniques ................................................................... 10
2.3 Metaheuristic Techniques ................................................................................. 11
Summary ................................................................................................................... 16
References ................................................................................................................. 17
Chapter 3: Overview of Proposed Techniques ...................................................... 23
3.1. Artificial gorilla troops optimization (GTO) ................................................... 23
3.1.1. Exploration Phase ..................................................................................... 24
3.1.2. Exploitation Phase .................................................................................... 24
3.2 Tasmanian devil optimization (TDO)............................................................... 27
3.2.1. Exploration Phase (Feeding by eating carrion) ......................................... 27
3.2.2. Exploitation Phase (Feeding by eating prey) ............................................ 27
Summary ................................................................................................................... 30
References ................................................................................................................. 31
Chapter 4: Problem Formulation ........................................................................... 32
4.1. Introduction ..................................................................................................... 32
4.2. Objective Functions ......................................................................................... 32
4.2.1 Reduction of Active Power Loss (APL) .................................................... 32
4.2.2. Reduction of Reactive Power Loss (QPL) ................................................ 32

vi
 4.2.3. Minimization of Voltage Deviation (VD)................................................. 33
4.2.4. Maximization of Voltage Stability Index (VSI) ....................................... 33
4.2.5. Minimization of total operating cost (TOC) ............................................. 33
4.3. Formulation of Multi Objective Function (MOF) ........................................... 35
4.4. Constraints ....................................................................................................... 35
4.4.1.Equality Constraints ................................................................................... 35
4.4.2.Equality Constraints ................................................................................... 36
4.5. Load Modelling ............................................................................................... 36
Summary ................................................................................................................... 38
References ................................................................................................................. 39
Chapter 5: Results and Discussion ......................................................................... 41
5.1. Introduction ..................................................................................................... 41
5.2. IEEE 33 bus test system .................................................................................. 41
5.2.1. Scenario 1: Evaluation of Single Objective Function (SOF) .................... 42
5.2.2. Scenario 2: Evaluation of Multi-Objective Function (MOF) ................... 49
5.3. IEEE 69 bus test system .................................................................................. 52
5.3.1. Scenario 1: Evaluation of Single Objective Function (SOF) .................... 52
5.3.2. Scenario 2: Evaluation of Multi-Objective Function (MOF) ................... 59
Summary ................................................................................................................... 64
References ................................................................................................................. 65
Chapter 6: Conclusion and Future Work .............................................................. 67
6.1. Conclusion ....................................................................................................... 67
6.2. Future Work ..................................................................................................... 67
Appendix…………….. .............................................................................................. 69

vii
 List of Figures
Figure 1.1: Centralized and decentralized power generation system ........................... 1
Figure 3.1: Different phases of artificial gorilla troops optimizer (GTO) ................. 23
Figure 3.2: Flow chart of GTO algorithm to solve problem of OADG ..................... 26
Figure 3.3: Flow chart of GTO algorithm to solve problem of OADG ..................... 29
Figure 5.1: Single-line diagram of 33 bus RDS ......................................................... 42
Figure 5.2: Voltage profiles for IEEE 33 bus system across various loads at different
cases for scenario 1 .................................................................................................... 47
Figure 5.3: Active power loss for 33 buses under various types of loads .................. 48
Figure 5.4: Convergence characteristics of TDO and GTO for IEEE 33 bus system at
different pf; a) Unity pf (b) 0.85 pf ............................................................................ 48
Figure 5.5: Single-line diagram of 69 bus RDS ......................................................... 53
Figure 5.6: Voltage profiles for IEEE 69 bus system for different loads at different
cases for scenario 1 .................................................................................................... 58
Figure 5.7: Active power loss for 69 buses under various types of loads .................. 58
Figure 5.8: Convergence characteristics of TDO and GTO for IEEE 69 bus system for
scenario 1 at different pf; a) Unity pf (PV type), (b) 0.85 pf (WT type) ................... 59

viii
 List of Tables
Table 2.1: Summary of optimization techniques used for OADG in RDNs .............. 14
Table 4.1: Load types and exponent values ............................................................... 37
Table 5.1: Control parameters and settings ................................................................ 41
Table 5.2: Results comparison of OADG for 33 bus RDN for APL minimization for
PV-DG in CP load model ........................................................................................... 43
Table 5.3: Results of OADG for 33 bus RDN for QPL minimization for PV-DG in CP
load model .................................................................................................................. 43
Table 5.4: Results of OADG for 33 bus RDN for APL minimization for WT-DG in
CP load model ............................................................................................................ 43
Table 5.5: Results of OADG for 33 bus RDN for QPL minimization for WT-DG in
CP load model ............................................................................................................ 44
Table 5.6: Results comparison of OADG for 33 bus RDN for APL minimization for
PV-DG in CP load model ........................................................................................... 44
Table 5.7: Results of OADG for 33 bus RDN for QPL minimization for PV-DG in VP
load model .................................................................................................................. 45
Table 5.8: Results of OADG for 33 bus RDN for APL minimization for WT-DG in
VP load model ............................................................................................................ 46
Table 5.9: Results of OADG for 33 bus RDN for QPL minimization for WT-DG in
VP load model ............................................................................................................ 46
Table 5.10: Results of OADG for 33 bus RDN for MOF in CP load model ............. 49
Table 5.11: Results of OADG for 33 bus RDN for MOF for PV-DG in VP load model
.................................................................................................................................... 50
Table 5.12: Results of OADG for 33 bus RDN for MOF for WT-DG in VP load model
.................................................................................................................................... 51
Table 5.13: Results comparison of OADG for 33 bus RDN for APL minimization for
PV-DG in CP load model ........................................................................................... 53
Table 5.14: Results of OADG for 69 bus RDN for QPL minimization for PV-DG in
CP load model ............................................................................................................ 54
Table 5.15: Results comparison of OADG for 69 bus RDN for APL minimization for
WT-DG in CP load model .......................................................................................... 54

ix
Table 5.16: Results of OADG for 69 BUS RDN for QPL minimization for WT-DG in
CP load model ............................................................................................................ 54
Table 5.17: Results comparison of OADG for 69 bus RDN for PV-DG in VP load
model .......................................................................................................................... 55
Table 5.18: Results of OADG for 69 bus RDN for QPL minimization for PV-DG in
VP load model ............................................................................................................ 56
Table 5.19: Results of OADG for 69 bus RDN for APL minimization for WT-DG in
VP load model ............................................................................................................ 56
Table 5.20: Results of OADG for 69 bus RDN for QPL minimization for WT-DG in
VP load model ............................................................................................................ 57
Table 5.21: Results of OADG for 69 bus RDN for MOF in CP load model ............. 60
Table 5.22: Results of OADG for 69 bus RDN for MOF for PV-DG in VP load model
.................................................................................................................................... 61
Table 5.23: Results of OADG for 69 bus RDN for MOF for WT-DG in VP load model
.................................................................................................................................... 62

x
 List of Abbreviations

RDN Radial Distribution Network

VD Voltage Deviation
APL Active Power Loss
QPL Reactive Power Loss
VSI Voltage Stability Index
OCI Operating Cost Index
DG Distributed Generations
DGM DG maintenance cost
DGO DG operational cost
DGINV DG investment cost
OCTDG Total operating cost of DG
GTO Artificial Gorilla Troops Optimization
TDO Tasmanian Devil Optimization
OADG Optimal Allocation of DG
Inf_R Inflation Rate
Int_R Interest Rate
PWF Power Worth Factor
PDG Active Power of DG
RDS Radial Distribution System

xi
 List of Publications
1. Rehman, H.U.; Hussain, A.; Haider, W.; Ali, S.A.; Kazmi, S.A.A.; Huzaifa,
M. Optimal Planning of Solar Photovoltaic (PV) and Wind-Based DGs for
Achieving Techno-Economic Objectives across Various Load Models.
Energies 2023, 16, 2444. https://doi.org/10.3390/en16052444 (IF = 3.252).
2. Ali, S.A.; Hussain, A.; Haider, W.; Rehman, H.U.; Kazmi, S.A.A. Optimal
Energy Management System of Isolated Multi-Microgrids with Local Energy
Transactive Market with Indigenous PV-, Wind-, and Biomass-Based
Resources. Energies 2023, 16, 1667. https://doi.org/10.3390/en16041667 (IF
= 3.252).
3. “Optimal planning approaches under uncertain and seasonal variations across
an active distribution grid encapsulating large scale electrical vehicle fleets and
renewable generation.” Muhammad Huzaifa, Arif Hussain, Waseem Haider,
Syed Ali Abbas Kazmi, Usman Ahmad, Habib Ur Rehman.
Sustainability, ISSN: 2071-1050 (Under Review). (IF = 3.889).

xii
 Chapter 1: Introduction
1.1. Background
Due to the rapidly increasing population, consumption of electricity has increased. As
a result, performance of the distribution systems (DSs) has been decreased because of
higher power losses and voltage drops due to higher R/X ratio in the existing DSs [1].
To meet the rising demand of electricity, electric utilities are constantly planning to
expand their existing electrical systems. The conventional planning approach is to
build the new grid or enlarge the existing one [2]. But this is not feasible solution in
economic and environmental aspects as most of the existing networks are radial in
nature and energized by fossil fuels [3], that affects the environment badly by the
emission of harmful gasses. Thus, an attractive substitution to meet the rising demand
is the use of distributed generation (DG) in DSs. Unlike the centralized or unified
power generation, DGs are small power sources installed near the end consumers [4]
in which flow of power can be bidirectional. This way of power generation is known
as decentralized or dispersed generation. DGs can be either renewable or non-
renewables.
Due to environmentally friendly, reliability and cost-effective benefits of renewables
DGs like solar and wind etc., are extensively used nowadays to generate the power on
a larger scale [5]-[6]. The generation of power in centralized and decentralized manner
can be seen from Figure 1.1. DGs play a vital role to enhance the DSs efficiency and
reliability by decreasing the power losses, voltage drops and enhancing the voltage
stability.

Figure 1.1: Centralized and decentralized power generation system [7]

1
But improper allocation of DGs have adverse effects on the DSs performance by
increasing the power losses, voltage drops and reducing the voltage stability of the
network. Thus, to enhance the technical, economic and environmental benefits,
optimal allocation of DGs (OADG) in DSs needs to be studied cautiously.
The optimal sitting and sizing of DGs in DS has been a challenging problem to focus
on. Therefore, researchers have taken various studies and developed different
optimization methods that are organized into different groups like analytical &
classical [8], [9], [10] and metaheuristics [11], [12], [13], [14] to address the problem
of OADGs in DSs. These techniques are employed to achieve different objectives like
reduction of active power loss (APL) [15], [16], [17], [18], [19], minimizing average
voltage deviation (VD) [20], [21], [22], reduce the greenhouse gas (GHG) emissions
[23] and maximize the voltage stability index (VSI) [24], [25], [26] by optimizing the
single objective function (SOF) and multi-objective function (MOF). In SOF, only a
single objective is achieved while in MOF, multiple objectives are achieved
concurrently.
1.2. Motivation
To achieve maximum support from DGs, OADG in distribution networks (DNs) is
mandatory. For that purpose, various studies have been taken by researchers to
improve the voltage profile, minimize the power loss, VD and cost along with the
maximizing of VSI of the DN. Some of them only focused on optimizing the technical
objectives while others focused on optimizing the techno-economic objectives
simultaneously for constant power load model (CP). Most of the earlier studies
centered on the power flow issue, using the CP load model. However, actual loads
characteristics show how much load power depends on the bus voltage. The
distribution system's load can often be divided into three categories: industrial,
residential, and commercial loads. A few studies have been seen in literature that
consider the voltage dependent load models or variable power load model, but they
only consider the technical parameter. So, to address the problem of OADG to achieve
the techno economic objectives by considering VP load models is rare, that is explored
in this study for standard IEEE 33 bus IEEE 69 bus radial distribution networks
(RDNs).

2
1.3. Contributions
In this study, two efficient metaheuristic automatization algorithms named as artificial
gorilla troops optimization (GTO) and Tasmanian devil optimization (TDO) are
proposed to address the issue of OADG in DNs for achieving techno-economic
objectives using multi-objectivity method i.e., weighted sum method. Following are
the main contributions of the proposed study:
• Proposed optimization algorithms named GTO and TDO.
• Proposed techniques are analyzed for renewable DGs (PV and WT) at unity and
combined pf.
• The GTO and TDO techniques address OADG problem to minimize APL and QPL
with SOF.
• The GTO and TDO techniques address ODGA problem to improve APL
reduction, VD, VSI and OCI (investment, operation, and maintenance costs of DG)
concurrently with MOF.
• The robustness of the proposed algorithms is examined on two standard test bus
systems. i.e., IEEE 33 and 69 bus RDNs and compared with each other.
• The proposed techniques are further explored for voltage dependent load models
consisting of residential, commercial and industrial load models, which make the
problem more practical than CP load model.
• The proposed techniques generate better results over the existing optimization
techniques, which validate the methodology.
• The proposed GTO technique outperforms the other proposed TDO technique in
terms of convergence characteristics and optimizing SOF.
1.4. Thesis Layout
The thesis is organized as follows:
Chapter 2: Literature review
This chapter provides an insight for the reader to see which optimization techniques
had been used for OADG in DN in literature.
Chapter 3: Overview of the proposed algorithms
In this chapter, a complete overview of the proposed optimization technique and how
it works, has been presented.

3
Chapter 4: Problem Formulation
This chapter consists of the mathematical interpretation of the objective functions with
system constraints and load modeling.
Chapter 5: Results and Discussion
In this chapter, results obtained by employing the proposed algorithms are discussed
and compared with the existing optimization techniques.
Chapter 6: Conclusions and future work
Finally, the conclusion of the study is discussed in this chapter along with the future
work which describes in which possible way this study can be developed in future.

4
Summary
In this chapter, the information is provided about the challenges faced by distribution
systems to meet the growing demand of electricity along with increasing power loss
and voltage drops. The contributions of the current study are then presentedto address
these issues. Finally, the thesis organization is presented.

5
References
[1] Karunarathne, E.; Pasupuleti, J.; Ekanayake, J.; Almeida, D. Optimal Placement
and Sizing of DGs in Distribution Networks Using MLPSO Algorithm.
Energies 2020, 13, 6185, doi: 10.3390/en13236185
[2] S. Sharma, S. Bhattacharjee, and A. Bhattacharya, “Quasi-Oppositional Swine
Influenza Model Based Optimization with Quarantine for optimal allocation of
DG in radial distribution network,” Int. J. Electr. Power Energy Syst., vol. 74,
pp. 348–373, 2016, doi: 10.1016/j.ijepes.2015.07.034.
[3] W. Haider, S. J. Ul Hassan, A. Mehdi, A. Hussain, G. O. M. Adjayeng, and C.
H. Kim, “Voltage profile enhancement and loss minimization using optimal
placement and sizing of distributed generation in reconfigured network,”
Machines, vol. 9, no. 1, pp. 1–16, 2021, doi: 10.3390/machines9010020.
[4] P. Paliwal, N. P. Patidar, and R. K. Nema, “Planning of grid integrated
distributed generators: A review of technology, objectives and techniques,”
Renew. Sustain. Energy Rev., vol. 40, pp. 557–570, 2014, doi:
10.1016/j.rser.2014.07.200.
[5] K. Mahmoud, N. Yorino, and A. Ahmed, “Optimal Distributed Generation
Allocation in Distribution Systems for Loss Minimization,” IEEE Trans. Power
Syst., vol. 31, no. 2, pp. 960–969, 2016, doi: 10.1109/TPWRS.2015.2418333.
[6] X. Xiong, W. Wu, N. Li, L. Yang, J. Zhang, and Z. Wei, “Risk-based multi-
objective optimization of distributed generation based on GPSO-BFA
algorithm,” IEEE Access, vol. 7, pp. 30563–30572, 2019, doi:
10.1109/ACCESS.2019.2902886.
[7] A. Ehsan and Q. Yang, “Optimal integration and planning of renewable
distributed generation in the power distribution networks: A review of
analytical techniques,” Appl. Energy, vol. 210, no. November 2017, pp. 44–59,
2018, doi: 10.1016/j.apenergy.2017.10.106.
[8] J. A. Sa’ed, M. Amer, A. Bodair, A. Baransi, S. Favuzza, and G. Zizzo, “A
simplified analytical approach for optimal planning of distributed generation in
electrical distribution networks,” Appl. Sci., vol. 9, no. 24, 2019, doi:
10.3390/app9245446.
[9] O. Garfi and H. Aloui, “Multiple distributed generations placement and sizing
based on voltage stability index and power loss minimization,” Turkish J.

6
 Electr. Eng. Comput. Sci., vol. 27, no. 6, pp. 4567–4579, 2019, doi:
10.3906/ELK-1812-45.
[10] P. Mehta, P. Bhatt, and V. Pandya, “Optimal selection of distributed generating
units and its placement for voltage stability enhancement and energy loss
minimization,” Ain Shams Eng. J., vol. 9, no. 2, pp. 187–201, 2018, doi:
10.1016/j.asej.2015.10.009.
[11] E. S. Ali, S. M. Abd Elazim, and A. Y. Abdelaziz, “Ant Lion Optimization
Algorithm for optimal location and sizing of renewable distributed
generations,” Renew. Energy, vol. 101, pp. 1311–1324, 2017, doi:
10.1016/j.renene.2016.09.023.
[12] M. Q. Duong, T. D. Pham, T. T. Nguyen, A. T. Doan, and H. Van Tran,
“Determination of optimal location and sizing of solar photovoltaic distribution
generation units in radial distribution systems,” Energies, vol. 12, no. 1, 2019,
doi: 10.3390/en12010174.
[13] U. Raut and S. Mishra, “An improved Elitist–Jaya algorithm for simultaneous
network reconfiguration and DG allocation in power distribution systems,”
Renew. Energy Focus, vol. 30, no. September, pp. 92–106, 2019, doi:
10.1016/j.ref.2019.04.001.
[14] A. Waqar et al., “Analysis of optimal deployment of several DGs in distribution
networks using plant propagation algorithm,” IEEE Access, vol. 8, pp. 175546–
175562, 2020, doi: 10.1109/ACCESS.2020.3025782.
[15] M. C. V. Suresh and E. J. Belwin, “Optimal DG placement for benefit
maximization in distribution networks by using Dragonfly algorithm,”
Renewables Wind. Water, Sol., vol. 5, no. 1, 2018, doi: 10.1186/s40807-018-
0050-7.
[16] J. Radosavljevic, N. Arsic, M. Milovanovic, and A. Ktena, “Optimal Placement
and Sizing of Renewable Distributed Generation Using Hybrid Metaheuristic
Algorithm,” J. Mod. Power Syst. Clean Energy, vol. 8, no. 3, pp. 499–510,
2020, doi: 10.35833/MPCE.2019.000259.
[17] S. Kamel, A. Awad, H. Abdel-Mawgoud, and F. Jurado, “Optimal DG
allocation for enhancing voltage stability and minimizing power loss using
hybrid gray Wolf optimizer,” Turkish J. Electr. Eng. Comput. Sci., vol. 27, no.
4, pp. 2947–2961, 2019, doi: 10.3906/elk-1805-66.
[18] S. Essallah, A. Khedher, and A. Bouallegue, “Integration of distributed

7
 generation in electrical grid: Optimal placement and sizing under different load
conditions,” Comput. Electr. Eng., vol. 79, p. 106461, 2019, doi:
10.1016/j.compeleceng.2019.106461.
[19] V. K. Thunuguntla and S. K. Injeti, “Butterfly optimizer assisted Max–Min
based multi-objective approach for optimal connection of DGs and optimal
network reconfiguration of distribution networks,” J. Electr. Syst. Inf. Technol.,
vol. 9, no. 1, 2022, doi: 10.1186/s43067-022-00049-y.
[20] M. Shadman Abid, H. J. Apon, K. A. Morshed, and A. Ahmed, “Optimal
Planning of Multiple Renewable Energy-Integrated Distribution System With
Uncertainties Using Artificial Hummingbird Algorithm,” IEEE Access, vol. 10,
pp. 40716–40730, 2022, doi: 10.1109/ACCESS.2022.3167395.
[21] A. Rajendran and K. Narayanan, “Optimal multiple installation of DG and
capacitor for energy loss reduction and loadability enhancement in the radial
distribution network using the hybrid WIPSO–GSA algorithm,” Int. J. Ambient
Energy, vol. 41, no. 2, pp. 129–141, 2020, doi:
10.1080/01430750.2018.1451371.
[22] G. Deb, K. Chakraborty, and S. Deb, “Modified Spider Monkey Optimization-
Based Optimal Placement of Distributed Generators in Radial Distribution
System for Voltage Security Improvement,” Electr. Power Components Syst.,
vol. 48, no. 9–10, pp. 1006–1020, 2020, doi: 10.1080/15325008.2020.1829186.
[23] A. Keane et al., “State-of-the-art techniques and challenges ahead for
distributed generation planning and optimization,” IEEE Trans. Power Syst.,
vol. 28, no. 2, pp. 1493–1502, 2013, doi: 10.1109/TPWRS.2012.2214406.
[24] A. Onlam, D. Yodphet, R. Chatthaworn, C. Surawanitkun, A. Siritaratiwat, and
P. Khunkitti, “Power loss minimization and voltage stability improvement in
electrical distribution system via network reconfiguration and distributed
generation placement using novel adaptive shuffled frogs leaping algorithm,”
Energies, vol. 12, no. 3, pp. 1–12, 2019, doi: 10.3390/en12030553.
[25] A. S. Hassan, Y. Sun, and Z. Wang, “Multi-objective for optimal placement and
sizing DG units in reducing loss of power and enhancing voltage profile using
BPSO-SLFA,” Energy Reports, vol. 6, pp. 1581–1589, 2020, doi:
10.1016/j.egyr.2020.06.013.
[26] R. SELLAMI, F. SHER, and R. NEJI, “An improved MOPSO algorithm for
optimal sizing & placement of distributed generation: A case study of the

8
Tunisian offshore distribution network (ASHTART),” Energy Reports, vol. 8,
pp. 6960–6975, 2022, doi: 10.1016/j.egyr.2022.05.049.

9
 Chapter 2: Literature Review
2.1 Background
The optimal sitting and sizing of DGs in DS has been challenging problem to focus
on. Therefore, researchers have taken various studies and developed different
optimization methods including analytical, heuristic and metaheuristic approaches to
address the problem of OADG with different operating requirements for optimizing
single objective function (SOF) and multi-objective function (MOF).

2.2 Analytical & heuristic Techniques

In [2], a simplified analytical technique has been suggested for determining OADG
and optimal power factor (OPF) to enhance the active power loss (APL) reduction for
three different radial distribution systems (RDSs) i.e., IEEE 12,33 and 69 bus RDSs.
To enhance the APL reduction, voltage profile and economic benefits, an analytical
approach has been recommended in [3] for determining optimal sitting and sizing of
various types of DGs. The effectiveness of the suggested approach was validated by
testing on standard IEEE 33 and 69 bus RDSs for different loading scenarios as well.
A modified version of analytical method has been developed in [4] for OAPVDG with
the aim of enhancing economic benefits and technical benefits (APL reduction and
voltage stability margin enhancement) in DSs planning.
For defining optimal sitting of DGs line loss sensitivity factor (LSF) utilized while
particle swarm optimization (PSO) with constriction factor was employed for
ascertaining optimal DG capacity. The implied approach examined on IEEE 33 bus
RDS. The authors in [5], proposed MINLP model solved through GAMS for
determining the OADG with the aim of improving PL reduction. To validate the
robustness of the approach, two standard test systems IEEE 33 and 69 bus RDSs were
used. The proposed approach is also employed on real 27 node test DS to optimally
locate PV-DG considering a sunny day. In [6], a new VSI based LQP index was
utilized to find the optimal DG location while optimal DG capacity was ascertained
by reducing APL. The suggested approach was examined on IEEE 33 bus RDS. In [7],
voltage stability margin index (VSMI) was utilized to find the optimal placement of
DG and optimal DG capacity was ascertained by curve fitting technique (CFT). The
suggested approach was examined on IEEE 33 and 69 bus RDSs.

10
In [8], a decision-making algorithm has been developed for ascertaining OADG of
different types with the aim of minimizing total APL,QPL and improving the voltage
profile of the DSs. The suggested method was tested on IEEE 33 bus RDS. A heuristic
approach has been suggested in [9] for determining OADG and capacitor banks (CBs)
to minimize the APL and enhance the voltage profile of the DSs. The implied
technique was tested on IEEE 33, 69 and 119 test RDSs at OPF.
Analytical techniques are slow and easy to implement when DGs are placed one by
one, but they trap in local minima when multiple DGs must be integrated at a time.
For that purpose, metaheuristic techniques are best suited.

2.3 Metaheuristic Techniques

Due to robust searching abilities to calculate the optimal solution for larger DNs,
metaheuristic techniques have been widely used to solve the problem of OADG in DSs
with SOF and MOF. In SOF, only single objective is optimized while optimizing
multiple objectives simultaneously considered as MOF. Most of the metaheuristic
techniques such as multileader particle swarm optimization (MLPSO) [1], improved
stochastic fractal search algorithm (CFSA) [10], fine-tuned PSO [11], strawberry plant
propagation algorithm (SPPA) [12] and Aquila optimization (AO) [13] used SOF to
allocate multiple DGs for minimizing the APL while enhancing the voltage profile of
the system. Moreover, to minimize the APL and enhance the voltage profile, a hybrid
of binary PSO and shuffled frog leap (SLFA), (BPSO-SLFA) has been proposed in
[14] for solving the OADG problem along with network configuration. The suggested
algorithm was tested on standard IEEE 33 and 69 bus RDSs. An improved version of
forensic-based investigation (FBI), quasi oppositional FBI (QOBI) has been
recommended in [15] for solving the problem of optimal placement and sizing along
with OPF of various types of DGs with the aim of minimizing the APL and improve
the voltage profile of the system.
In [16], authors proposed the Coulomb-Franklin’s algorithm (CFA) for OADG along
with capacitor banks (CBS) in DS to minimize the APL. The above-mentioned
techniques were implemented only for constant power load model. Authors in [17]
suggested the genetic algorithm (GA) for solving the problem of OADG to minimize
the APL while enhancing the voltage profile of the system. The suggested algorithm
is validated on standard IEEE 33 bus RDS with different load states and load models
by considering the daily and yearly load profile.

11
As compared to SOF, MOF is used to optimize multiple objectives simultaneously by
converting them into SOF through various multi-objectivity methods. To optimize the
MOF based on reduction of APL, VD and VSI maximization for solving the problem
of optimal placement and sizing of multiple and various types of DGs in DN, different
multi-objectivity methods have been employed by various metaheuristic techniques
such as quasi-oppositional chaotic symbiotic organisms search (QOCSOS) [18],
quasi-oppositional differential evolution Lévy flights algorithm (QODELFA) [19],
manta rays foraging optimization (MRFO) [20] and improved grey wolf optimizer-
PSO (I-GWOPSO) [21] have used weighted sum method, the improved decomposition
based evolutionary algorithm (I-DBEA) [22] used e-constraints method, chaotic sine
cosine algorithm (CSCA) [23] and improved Harris hawks optimization (IHHO) [24]
have used grey relation analysis to find the best optimal solution from pareto optimal
solution, enhanced artificial ecosystem-based optimization (EAEO) algorithm [25]
have used Fuzzy decision making approach to find the best optimal solution from
pareto optimal solution while adaptive PSO and gravitational search algorithm (GSA)
have used pareto optimal front, e-constraints and aggregated sum methods.
Moreover, to solve MOF based on technical (APL,VD,VSI), economic (cost of CB
and cost of power produced) and environmental (pollutant emissions) objectives
improvements for aiming the OADG and CB, salp swarm Algorithm (SSA) [26] and
water cycle algorithm (WCA) [27] have used weighted sum method. A multi-objective
differential evolution (MODE) algorithm [28] have used fuzzy approach to solve MOF
for improving the technical, economic and environmental objectives. The suggested
approach was implemented on standard IEEE 33,69 bus and 62 bus real RDSs. In [29],
authors introduced multi objective firefly algorithm (MOFA) to minimize the PL, VD,
THD, cost of DG,GHG emissions and maximize the VSI as MOOP for ascertaining
the ODGA. The optimal solution was determined by using a fuzzy decision-making
approach. The recommended technique was checked on IEEE 33 bus RDN and actual
62 bus Indian distribution system.
In [30], hybrid optimization technique (GSA+GAMS) implemented for OADGs
(solar, wind and hydro) with network reconfiguration in DN to optimize the MOF
based on improvement of technical (APL) and economic (cost) benefits. In this study,
authors used GSA to find the optimal placement of DGs and for optimal capacity of
DG GAMS was applied. To check the validity and effectiveness of the algorithm,
standard IEEE 33 bus RDN and a real time varying DN were utilized. A multi-

12
objective opposition based chaotic differential evolution (MOCDE) algorithm has
been proposed on [31] to solve OADG problem for optimizing the MOF to achieve
technical objectives ( minimization of PL and reduction of VD); and economic
objective (minimization of yearly economic loss cost). The suggested algorithm is
examined on standard IEEE 33 and 69 bus RDNs. In [32], the author introduced the
artificial bee colony (ABC) optimization algorithm for optimal sitting and sizing of
DGs to minimize the APL, voltage drops and total energy costs concurrently, as multi
objective problem (MOP). The suggested algorithm is tested on standard IEEE 33 and
69 test bus RDNs.
An opposition-based chaotic differential evolution (OTCDE) is proposed in [33] for
OADG to deal with MOP consisting technical (Voltage deviation and line flow
capacity) and economic objectives. The anticipated system is tested on standard IEEE
33, 69 and 118 bus RDNs. The whale optimization algorithm (WOA) [34] and stud
krill herd algorithm (SKHA) [35] have been recommended to solve the problem of
OADG in DN for optimizing MOF based on APL, VD and OC. The MOF has been
solved by using weighted sum approach. In [36], authors proposed the ant colony
optimization (ACO) algorithm for Optimal allocation of renewable based DGs with
the aim of optimizing MOF based on APL index, VD index and OCI. The suggested
approach was tested on IEEE 33 bus and actual 85 bus Indian DS.
Authors in [37] proposed the artificial hummingbird algorithm (AHA) to solve the
problem of optimal allocation of renewables DGs considering uncertainties for
optimizing the MOF based on APL, VD, VSM and total annual energy savings by
using weighted sum method. The suggested technique was tested on standard IEEE
33, 69 bus RDS for different load states. The comprehensive review of optimization
methods used for OADG to achieve different objectives is presented in Table 2.1.

13
 Table 2.1: Summary of optimization techniques used for OADG in RDNs

Optimization Objective Functions Load Models

Ref. Year
Methods APL QPL VD VSI Cost CP VP
[38] 2018 DA ✔ ✖ ✖ ✖ ✖ ✔ ✖
[44] 2018 WOA ✔ ✖ ✔ ✔ ✔ ✔ ✖
[20] 2018 CTLBO ✔ ✖ ✔ ✔ ✖ ✔ ✖
[39] 2018 IEHO ✔ ✖ ✔ ✔ ✖ ✔ ✖
[2] 2019 Analytical ✔ ✖ ✖ ✖ ✖ ✔ ✖
[40] 2019 ASFL ✔ ✖ ✖ ✔ ✖ ✔ ✖
[41] 2019 BBO ✔ ✖ ✖ ✖ ✖ ✔ ✖
[42] 2019 MOHTLGOGWO ✔ ✖ ✖ ✖ ✖ ✔ ✖
[10] 2019 CSFS ✔ ✖ ✖ ✖ ✖ ✔ ✖
[43] 2019 LSF + GWO ✔ ✖ ✖ ✔ ✖ ✔ ✖
[17] 2019 GA ✔ ✖ ✖ ✖ ✖ ✔ ✔
[7] 2019 VSMI + CFT ✔ ✖ ✖ ✖ ✖ ✔ ✖
[9] 2019 Heuristic ✔ ✖ ✖ ✖ ✖ ✔ ✖
[19] 2019 QODELFA ✔ ✖ ✔ ✔ ✖ ✔ ✖
[1] 2020 MLPSO ✔ ✖ ✖ ✖ ✖ ✔ ✖
[18] 2020 QOCSOS ✔ ✖ ✔ ✔ ✖ ✔ ✖
[44] 2020 IRRO ✔ ✔ ✔ ✔ ✔ ✔ ✖
[23] 2020 CSCA ✔ ✖ ✔ ✔ ✖ ✔ ✖
[33] 2020 OTCDE ✔ ✖ ✔ ✖ ✔ ✔ ✖
[25] 2020 EAEO ✔ ✖ ✔ ✔ ✖ ✔ ✖
[45] 2020 APSO & MGSA ✔ ✖ ✔ ✔ ✖ ✔ ✖
[5] 2020 MINLP(GAMS) ✔ ✖ ✖ ✖ ✖ ✔ ✖
[24] 2020 IHHO ✔ ✖ ✔ ✔ ✖ ✔ ✖
[46] 2020 PSO & MOPSO ✔ ✖ ✔ ✖ ✖ ✔ ✖
[14] 2020 BPSO-SLFA ✔ ✖ ✖ ✖ ✖ ✔ ✖
[47] 2020 MSMO ✖ ✖ ✔ ✖ ✖ ✔ ✖
[20] 2020 MRFO ✔ ✖ ✔ ✔ ✖ ✔ ✖
[22] 2021 I-DBEA ✔ ✖ ✔ ✔ ✖ ✔ ✖
[32] 2021 ABC ✔ ✖ ✔ ✖ ✔ ✔ ✖
[48] 2021 BAT ✔ ✖ ✖ ✔ ✖ ✔ ✔
[11] 2021 Fine tunned PSO ✔ ✖ ✖ ✖ ✖ ✔ ✖
[12] 2021 SPPA ✔ ✖ ✖ ✖ ✖ ✔ ✖
[35] 2021 SKHA ✔ ✖ ✔ ✖ ✔ ✔ ✖
[3] 2022 Analytical ✔ ✖ ✖ ✖ ✖ ✔ ✖
[49] 2022 Improved MOPSO ✔ ✖ ✖ ✔ ✖ ✔ ✖
[29] 2022 POFA ✔ ✖ ✔ ✔ ✔ ✔ ✖
[21] 2022 I-GWOPSO ✔ ✖ ✔ ✔ ✖ ✔ ✖
[36] 2022 ACO ✔ ✖ ✔ ✖ ✔ ✔ ✖
[16] 2022 CFA ✔ ✖ ✖ ✖ ✖ ✔ ✖
[15] 2022 QOFBI ✔ ✖ ✖ ✖ ✖ ✔ ✖
[13] 2022 AO ✔ ✖ ✖ ✖ ✖ ✔ ✖
[50] 2022 Modified FPA ✔ ✖ ✖ ✖ ✖ ✔ ✔
[P] 2023 TDO, GTO ✔ ✔ ✔ ✔ ✔ ✔ ✔
Note: P – proposed; VP - variable power; CP – constant power

14
It can be seen from the above Table 2.1 that most of the research studies considered
CP load model while solving the problem of OADG in DNs for achieving technical or
economic or techno-economic objectives by optimizing SOF and MOF. However, in
[53] and [55] voltage dependent load models have been considered to solve problem
of OADG to achieve only technical objectives. Thus, in this study CP load model as
well as VP load models are considered to solve the problem of OADG in DN for
achieving the techno-economic objectives by optimizing SOF and MOF.

15
Summary
This chapter provides information about the optimization techniques that had been
used in the past to address the problem of OADG in DN for achieving different
objectives at various pf.

16
References
[1] Karunarathne, E.; Pasupuleti, J.; Ekanayake, J.; Almeida, D. Optimal
Placement and Sizing of DGs in Distribution Networks Using MLPSO
Algorithm. Energies 2020, 13, 6185, doi: 10.3390/en13236185.
[2] J. A. Sa’ed, M. Amer, A. Bodair, A. Baransi, S. Favuzza, and G. Zizzo, “A
simplified analytical approach for optimal planning of distributed generation
in electrical distribution networks,” Appl. Sci., vol. 9, no. 24, 2019, doi:
10.3390/app9245446.
[3] M. Kashyap, S. Kansal, and R. Verma, “Sizing and Allocation of DGs in A
Passive Distribution Network Under Various Loading Scenarios,” Electr.
Power Syst. Res., vol. 209, no. April, p. 108046, 2022, doi:
10.1016/j.epsr.2022.108046.
[4] O. B. Adewuyi, A. P. Adeagbo, I. G. Adebayo, H. O. R. Howlader, and Y. Sun,
“Modified analytical approach for pv-dgs integration into a radial distribution
network considering loss sensitivity and voltage stability,” Energies, vol. 14,
no. 22, 2021, doi: 10.3390/en14227775.
[5] O. D. Montoya, W. Gil-González, and L. F. Grisales-Noreña, “An exact
MINLP model for optimal location and sizing of DGs in distribution networks:
A general algebraic modeling system approach,” Ain Shams Eng. J., vol. 11,
no. 2, pp. 409–418, 2020, doi: 10.1016/j.asej.2019.08.011.
[6] O. Garfi and H. Aloui, “Multiple distributed generations placement and sizing
based on voltage stability index and power loss minimization,” Turkish J.
Electr. Eng. Comput. Sci., vol. 27, no. 6, pp. 4567–4579, 2019, doi:
10.3906/ELK-1812-45.
[7] S. Essallah, A. Khedher, and A. Bouallegue, “Integration of distributed
generation in electrical grid: Optimal placement and sizing under different load
conditions,” Comput. Electr. Eng., vol. 79, p. 106461, 2019, doi:
10.1016/j.compeleceng.2019.106461.
[8] V. Vita, “Development of a decision-making algorithm for the optimum size
and placement of distributed generation units in distribution networks,”
Energies, vol. 10, no. 9, 2017, doi: 10.3390/en10091433.
[9] A. Bayat and A. Bagheri, “Optimal active and reactive power allocation in
distribution networks using a novel heuristic approach,” Appl. Energy, vol.

17
 233–234, no. April 2018, pp. 71–85, 2019, doi:
10.1016/j.apenergy.2018.10.030.
[10] T. P. Nguyen, T. T. Tran, and D. N. Vo, “Improved stochastic fractal search
algorithm with chaos for optimal determination of location, size, and quantity
of distributed generators in distribution systems,” Neural Comput. Appl., vol.
31, no. 11, pp. 7707–7732, 2019, doi: 10.1007/s00521-018-3603-1.
[11] E. Karunarathne, J. Pasupuleti, J. Ekanayake, and D. Almeida, “Network loss
reduction and voltage improvement by optimal placement and sizing of
distributed generators with active and reactive power injection using fine-tuned
pso,” Indones. J. Electr. Eng. Comput. Sci., vol. 21, no. 2, pp. 647–656, 2020,
doi: 10.11591/ijeecs.v21.i2.pp647-656.
[12] and B. U. Mohsin Shahzad, Waseem Akram, Muhammad Arif, Uzair Khan,
“Optimal Siting and Sizing of Distributed Generators by Strawberry Plant
Propagation,” Energies, vol. 6, no. 14, pp. 1–13, 2021, [Online]. Available:
https://doi.org/10.3390/en14061744.
[13] P. Samal and D. Panigrahy, “A Novel Technique Based on Aquila Optimiser
Algorithm for Optimal Integration of Distributed Generations in the
Distribution System,” Process Integr. Optim. Sustain., no. 0123456789, 2022,
doi: 10.1007/s41660-022-00278-8.
[14] A. S. Hassan, Y. Sun, and Z. Wang, “Multi-objective for optimal placement
and sizing DG units in reducing loss of power and enhancing voltage profile
using BPSO-SLFA,” Energy Reports, vol. 6, pp. 1581–1589, 2020, doi:
10.1016/j.egyr.2020.06.013.
[15] B. K. Malika, V. Pattanaik, B. K. Sahu, and P. K. Rout, Quasi-oppositional
Forensic-Based Investigation for Optimal DG Selection for Power Loss
Minimization, no. 0123456789. Springer Nature Singapore, 2022. doi:
10.1007/s41660-022-00277-9.
[16] V. Tiwari, H. M. Dubey, and M. Pandit, “Assessment of Optimal Size and
Location of DG/CB in Distribution Systems using Coulomb–Franklin’s
Algorithm,” J. Inst. Eng. Ser. B, vol. 103, no. 6, pp. 1885–1908, 2022, doi:
10.1007/s40031-022-00811-w.
[17] F. Moaidi and M. Moaidi, “Optimal Placement and Sizing of Distributed
Generation in Microgrid for Power Loss Reduction and Voltage Profile
Improvement,” vol. 13, no. 1, pp. 26–31, 2019, doi: 10.5281/zenodo.2571760.

18
[18] K. H. Truong, P. Nallagownden, I. Elamvazuthi, and D. N. Vo, “A Quasi-
Oppositional-Chaotic Symbiotic Organisms Search algorithm for optimal
allocation of DG in radial distribution networks,” Appl. Soft Comput. J., vol.
88, p. 106067, 2020, doi: 10.1016/j.asoc.2020.106067.
[19] R. J. Mahfoud, Y. Sun, N. F. Alkayem, H. H. Alhelou, P. Siano, and M. Shafie-
Khah, “A novel combined evolutionary algorithm for optimal planning of
distributed generators in radial distribution systems,” Appl. Sci., vol. 9, no. 16,
2019, doi: 10.3390/app9163394.
[20] M. G. Hemeida, S. Alkhalaf, A. A. A. Mohamed, A. A. Ibrahim, and T. Senjyu,
“Distributed Generators Optimization Based on Multi-Objective Functions
Using Manta Rays Foraging Optimization Algorithm (MRFO),” Energies, vol.
13, no. 15, 2020, doi: 10.3390/en13153847.
[21] M. I. Akbar, S. A. A. Kazmi, O. Alrumayh, Z. A. Khan, A. Altamimi, and M.
M. Malik, “A Novel Hybrid Optimization-Based Algorithm for the Single and
Multi-Objective Achievement with Optimal DG Allocations in Distribution
Networks,” IEEE Access, vol. 10, pp. 25669–25687, 2022, doi:
10.1109/ACCESS.2022.3155484.
[22] A. Ali, M. U. Keerio, and J. A. Laghari, “Optimal Site and Size of Distributed
Generation Allocation in Radial Distribution Network Using Multi-objective
Optimization,” J. Mod. Power Syst. Clean Energy, vol. 9, no. 2, pp. 404–415,
2021, doi: 10.35833/MPCE.2019.000055.
[23] A. Selim, S. Kamel, and F. Jurado, “Efficient optimization technique for
multiple DG allocation in distribution networks,” Appl. Soft Comput. J., vol.
86, p. 105938, 2020, doi: 10.1016/j.asoc.2019.105938.
[24] A. Selim, S. Kamel, A. S. Alghamdi, and F. Jurado, “Optimal Placement of
DGs in Distribution System Using an Improved Harris Hawks Optimizer
Based on Single- And Multi-Objective Approaches,” IEEE Access, vol. 8, pp.
52815–52829, 2020, doi: 10.1109/ACCESS.2020.2980245.
[25] A. Eid, S. Kamel, A. Korashy, and T. Khurshaid, “An enhanced artificial
ecosystem-based optimization for optimal allocation of multiple distributed
generations,” IEEE Access, vol. 8, pp. 178493–178513, 2020, doi:
10.1109/ACCESS.2020.3027654.
[26] K. S. Sambaiah and T. Jayabarathi, “Optimal allocation of renewable
distributed generation and capacitor banks in distribution systems using salp

19
 swarm algorithm,” Int. J. Renew. Energy Res., vol. 9, no. 1, pp. 96–107, 2019,
doi: 10.20508/ijrer.v9i1.8581.g7567.
[27] A. A. A. El-Ela, R. A. El-Sehiemy, and A. S. Abbas, “Optimal Placement and
Sizing of Distributed Generation and Capacitor Banks in Distribution Systems
Using Water Cycle Algorithm,” IEEE Syst. J., vol. 12, no. 4, pp. 3629–3636,
2018, doi: 10.1109/JSYST.2018.2796847.
[28] J. Rajalakshmi and S. Durairaj, “Application of multi-objective optimization
algorithm for siting and sizing of distributed generations in distribution
networks,” J. Comb. Optim., vol. 41, no. 2, pp. 267–289, 2021, doi:
10.1007/s10878-020-00681-2.
[29] S. Anbuchandran, R. Rengaraj, A. Bhuvanesh, and M. Karuppasamypandiyan,
“A Multi-objective Optimum Distributed Generation Placement Using Firefly
Algorithm,” J. Electr. Eng. Technol., vol. 17, no. 2, pp. 945–953, 2022, doi:
10.1007/s42835-021-00946-8.
[30] V. V. V. S. N. Murty and A. Kumar, “Optimal DG integration and network
reconfiguration in microgrid system with realistic time varying load model
using hybrid optimisation,” IET Smart Grid, vol. 2, no. 2, pp. 192–202, 2019,
doi: 10.1049/iet-stg.2018.0146.
[31] S. Kumar, K. K. Mandal, and N. Chakraborty, “Optimal DG placement by
multi-objective opposition based chaotic differential evolution for techno-
economic analysis,” Appl. Soft Comput. J., vol. 78, pp. 70–83, 2019, doi:
10.1016/j.asoc.2019.02.013.
[32] E. A. Al-Ammar et al., “ABC algorithm based optimal sizing and placement
of DGs in distribution networks considering multiple objectives,” Ain Shams
Eng. J., vol. 12, no. 1, pp. 697–708, 2021, doi: 10.1016/j.asej.2020.05.002.
[33] S. Kumar, K. K. Mandal, and N. Chakraborty, “A novel opposition-based
tuned-chaotic differential evolution technique for techno-economic analysis by
optimal placement of distributed generation,” Eng. Optim., vol. 52, no. 2, pp.
303–324, 2020, doi: 10.1080/0305215X.2019.1585832.
[34] D. B. Prakash and C. Lakshminarayana, “Multiple DG placements in radial
distribution system for multi objectives using Whale Optimization Algorithm,”
Alexandria Eng. J., vol. 57, no. 4, pp. 2797–2806, 2018, doi:
10.1016/j.aej.2017.11.003.
[35] S. A. Chithra Devi, K. Yamuna, and M. Sornalatha, “Multi-objective

20
 optimization of optimal placement and sizing of multiple DG placements in
radial distribution system using stud krill herd algorithm,” Neural Comput.
Appl., vol. 33, no. 20, pp. 13619–13634, 2021, doi: 10.1007/s00521-021-
05992-x.
[36] N. R. Godha, V. N. Bapat, and I. Korachagaon, “Ant colony optimization
technique for integrating renewable DG in distribution system with techno-
economic objectives,” Evol. Syst., vol. 13, no. 3, pp. 485–498, 2022, doi:
10.1007/s12530-021-09416-y.
[37] M. Shadman Abid, H. J. Apon, K. A. Morshed, and A. Ahmed, “Optimal
Planning of Multiple Renewable Energy-Integrated Distribution System With
Uncertainties Using Artificial Hummingbird Algorithm,” IEEE Access, vol.
10, pp. 40716–40730, 2022, doi: 10.1109/ACCESS.2022.3167395.
[38] M. C. V. Suresh and E. J. Belwin, “Optimal DG placement for benefit
maximization in distribution networks by using Dragonfly algorithm,”
Renewables Wind. Water, Sol., vol. 5, no. 1, 2018, doi: 10.1186/s40807-018-
0050-7.
[39] N. K. Meena, S. Parashar, A. Swarnkar, N. Gupta, and K. R. Niazi, “Improved
Elephant Herding Optimization for Multiobjective der Accommodation in
Distribution Systems,” IEEE Trans. Ind. Informatics, vol. 14, no. 3, pp. 1029–
1039, 2018, doi: 10.1109/TII.2017.2748220.
[40] A. Onlam, D. Yodphet, R. Chatthaworn, C. Surawanitkun, A. Siritaratiwat, and
P. Khunkitti, “Power loss minimization and voltage stability improvement in
electrical distribution system via network reconfiguration and distributed
generation placement using novel adaptive shuffled frogs leaping algorithm,”
Energies, vol. 12, no. 3, pp. 1–12, 2019, doi: 10.3390/en12030553.
[41] M. Q. Duong, T. D. Pham, T. T. Nguyen, A. T. Doan, and H. Van Tran,
“Determination of optimal location and sizing of solar photovoltaic
distribution generation units in radial distribution systems,” Energies, vol. 12,
no. 1, 2019, doi: 10.3390/en12010174.
[42] S. Arabi Nowdeh et al., “Fuzzy multi-objective placement of renewable energy
sources in distribution system with objective of loss reduction and reliability
improvement using a novel hybrid method,” Appl. Soft Comput. J., vol. 77, pp.
761–779, 2019, doi: 10.1016/j.asoc.2019.02.003.
[43] S. Kamel, A. Awad, H. Abdel-Mawgoud, and F. Jurado, “Optimal DG

21
 allocation for enhancing voltage stability and minimizing power loss using
hybrid gray Wolf optimizer,” Turkish J. Electr. Eng. Comput. Sci., vol. 27, no.
4, pp. 2947–2961, 2019, doi: 10.3906/elk-1805-66.
[44] S. Nagaballi and V. S. Kale, “Pareto optimality and game theory approach for
optimal deployment of DG in radial distribution system to improve techno-
economic benefits,” Appl. Soft Comput. J., vol. 92, p. 106234, 2020, doi:
10.1016/j.asoc.2020.106234.
[45] A. Eid, “Allocation of distributed generations in radial distribution systems
using adaptive PSO and modified GSA multi-objective optimizations,”
Alexandria Eng. J., vol. 59, no. 6, pp. 4771–4786, 2020, doi:
10.1016/j.aej.2020.08.042.
[46] S. Kamel, A. Selim, W. Ahmed, and F. Jurado, “Single- and multi-objective
optimization for photovoltaic distributed generators implementation in
probabilistic power flow algorithm,” Electr. Eng., vol. 102, no. 1, pp. 331–347,
2020, doi: 10.1007/s00202-019-00878-7.
[47] G. Deb, K. Chakraborty, and S. Deb, “Modified Spider Monkey Optimization-
Based Optimal Placement of Distributed Generators in Radial Distribution
System for Voltage Security Improvement,” Electr. Power Components Syst.,
vol. 48, no. 9–10, pp. 1006–1020, 2020, doi:
10.1080/15325008.2020.1829186.
[48] T. Yuvaraj et al., “Optimal integration of capacitor and distributed generation
in distribution system considering load variation using bat optimization
algorithm,” Energies, vol. 14, no. 12, 2021, doi: 10.3390/en14123548.
[49] R. SELLAMI, F. SHER, and R. NEJI, “An improved MOPSO algorithm for
optimal sizing & placement of distributed generation: A case study of the
Tunisian offshore distribution network (ASHTART),” Energy Reports, vol. 8,
pp. 6960–6975, 2022, doi: 10.1016/j.egyr.2022.05.049.
[50] A. Ramshanker, J. R. Isaac, B. E. Jeyeraj, J. Swaminathan, and R. Kuppan,
“Optimal DG Placement in Power Systems Using a Modified Flower
Pollination Algorithm,” Energies, vol. 15, no. 22, 2022, doi:
10.3390/en15228516.

22
 Chapter 3: Overview of Proposed
Techniques
In this chapter, working of the proposed metaheuristic optimization techniques to
address the problem of optimal placement and sizing of DGs in distribution systems
is presented. The proposed metaheuristic techniques are as follows:
• Artificial gorilla troops optimization (GTO)
• Tasmanian devil optimization (TDO)
3.1. Artificial Gorilla Troops Optimization (GTO)
In 2021 [1], Abdollahzadeh and Miralilli introduced the novel bio-inspired
metaheuristic algorithm, named gorilla troops optimizer (GTO), based on gorilla
group behavior in nature for finding food and living. Each troop has a silverback
gorilla that serves as the troop's leader which is responsible to ensure the safety of the
troop by taking the crucial decisions. Silverback gorilla is followed by young-male
gorillas (known as black-backs) to provide the backup protection for the group.
Like other optimization techniques, for optimization operation two phases are used by
GTO algorithm named as Exploration and Exploitation. The explanation of the phases
can be seen from Figure 3.1.

Figure 3.1: Different phases of artificial gorilla troops optimizer (GTO)

23
3.1.1. Exploration Phase
In GTO, three different strategies are used for exploration phase. In the first strategy,
the gorilla moves toward an unknown place while in the second gorilla moves in the
direction of another gorilla and in third strategy it moves to a known place. The X and
GX represent the position of gorilla and the silverback respectively. The mathematical
formulas used in this phase are formulated in Equation (3.1):

𝐺𝑋(𝑡 + 1) = (𝑢𝑏 − 𝑙𝑏) ∗ 𝑅1 + 𝑙𝑏, 𝑟𝑎𝑛𝑑 < 𝑃 (3.1a)

𝐺𝑋(𝑡 + 1) = (𝑅2 − 𝐶) ∗ 𝑋𝑟(𝑡) + 𝐿 ∗ 𝐻, 𝑟𝑎𝑛𝑑 ≥ 0.5 (3.1a)
𝐿 ∗ (𝑋(𝑡) − 𝐺𝑋𝑟(𝑡)) +
𝐺𝑋(𝑡 + 1) = 𝑋(𝑖) − 𝐿 ∗ ( ), 𝑟𝑎𝑛𝑑 < 0.5 (3.1c)
𝑅3 ∗ (𝑋(𝑡) − 𝐺𝑋𝑟(𝑡))

In Equation (3.1), ub and lb represent the upper and lower bound respectively. R1, R2
and R3 are random numbers in the range of [0,1], whereas t represents the current
iteration. p is a predetermined value in the interval of [0,1] which is used to select the
movement of the gorilla to unknown site as in the above strategies. rand denotes the
random value in the interval of 0-1. GXr and Xr are the randomly selected solutions
from the population. Whereas, other parameters are calculated using Equations (3.2),
(3.4) and (3.5).

𝑖𝑡
𝐶 = 𝐹 ∗ (1 − ) (3.2)
𝑚𝑎𝑥𝑖𝑡

(3.3)
𝐹 = cos(2 ∗ 𝑅4) + 1

(3.4)
𝐿 =𝐶∗𝑙
(3.5)
𝐻 = 𝑍 ∗ 𝑋(𝑡)

(3.6)
𝑍 = [−𝐶, 𝐶]

Where, it and maxit denote the current and max number of iterations respectively. R4
is the random parameter in the range of 0-1. The value of l is in the range of [-1,1].
3.1.2. Exploitation Phase
In exploitation phase, two different mechanisms are used, either gorilla’s troop follows
the silverback gorilla decision or competitions for the adult females. The probability
of selecting the mechanism is based on the value of C. If C >= W, the gorilla’s troop
follow the silverback instructions. This behavior is shown in Equation (3.7):

24
 𝐺𝑋(𝑡 + 1) = 𝐿 ∗ 𝑀 ∗ (𝑋(𝑡) − 𝑋𝑠𝑖𝑙𝑣𝑒𝑟𝑏𝑎𝑐𝑘 ) + 𝑋(𝑡) (3.7)
1
𝑁 𝑔 𝑔
1 (3.8)
𝑀 = (| ∑ 𝐺𝑋𝑖 (𝑡)| )
𝑁
𝑖=1

𝑔 = 2𝐿 (3.9)

Where, silverback position is represented by 𝑋𝑠𝑖𝑙𝑣𝑒𝑟𝑏𝑎𝑐𝑘 .

If C < W, competition for the adult females is selected and this behavior is expressed
in Equation (3.10):
𝐺𝑋(𝑖) = 𝑋𝑠𝑖𝑙𝑣𝑒𝑟𝑏𝑎𝑐𝑘 − (𝑋𝑠𝑖𝑙𝑣𝑒𝑟𝑏𝑎𝑐𝑘 ∗ 𝑄 − 𝑋(𝑡) ∗ 𝑄 ) ∗ 𝐴 (3.10)

𝑄 = 2 ∗ 𝑅5 − 1 (3.11)

𝐴= 𝛽∗𝐸 (3.12)

𝑁 , 𝑟𝑎𝑛𝑑 ≥ 0.5
𝐸= { 1 (3.13)
𝑁2 , 𝑟𝑎𝑛𝑑 < 0.5

Where Q resembles impact force, R5 is a random number in the interval [0,1] and β is
predefined variable. The value of rand is between 0 and 1The "E" will be equal to
random values in the problem matrix and gaussian distribution when the rand >=0.5
and contrary to this "E" will be from just gaussian distribution, when the rand < 0.5.
The process of GTO technique to address OADG in DN is depicted in Figure 3.2.

25
 Start

Initialize GTO parameters N,

T, p and 

Random locations are

generated for gorillas

t=0

Calculate the objective fitness

function for each gorilla

C and L are calculated using eq.

(3.2) and (3.4)

Gorilla location is updated using

eq. (3.1b)

yes No Yes
rand  rand < p

No

Gorilla location is updated using Gorilla location is updated using

eq. (3.1c) eq. (3.1a)

Calculate the objective fitness

function for new position of gorillas

t=t+1
No Yes
Gorilla location is updated using Gorilla location is updated using
C W
eq. (3.10) eq. (3.7)

Calculate the objective fitness function

for the updated position of gorillas

Yes
t Maxit

No

Get the optimal

solution

End

Figure 3.2: Flow chart of GTO algorithm to solve problem of OADG

26
3.2 Tasmanian Devil Optimization (TDO)
In 2022, Dehghani introduced the novel bio-inspired metaheuristic algorithm [2],
named Tasmanian devil optimizer (TDO), based on the Tasmanian devil’s behavior in
nature during feeding. Two different approaches/strategies are used by Tasmanian
devil during feeding process. The first one is eating carrion and the second one is
eating prey. The mechanism of these approaches is discussed below.
3.2.1. Exploration Phase (Feeding by eating carrion)
In TDO algorithm, the Tasmanian devil approach of feeding by eating carrion is
mathematically designed using Equations (3.14) to (3.15). The process of choosing
carrion is mimicked in (3.14).
𝐶𝑖 = 𝑋𝑘 , 𝑖 = 1, 2, 3, . . . . , 𝑁, 𝑘 𝜖 {1, 2, . . . , 𝑁|𝑘 ≠ 𝑖} (3.14)

Where, 𝐶𝑖 denotes the carrion selected by ith Tasmanian devil and X represents the
Tasmanian devils population. A new position for the Tasmanian devil in the search
space is determined based on the chosen carrion. The updated location of the
Tasmanian devil can be formulated using Equation (3.15).

𝑥𝑖 ,𝑗 + 𝑟 ∗ (𝑐𝑖 ,𝑗 − 𝑙 ∗ 𝑥𝑖 ,𝑗 ), 𝐹𝐶𝑖 < 𝐹𝑖 (3.15)

𝑥𝑖𝑛𝑒𝑤
,𝑗
,𝑆1
= {
𝑥𝑖 ,𝑗 + 𝑟 ∗ (𝑥𝑖 ,𝑗 − 𝑐𝑖 ,𝑗 ), 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒

𝑋𝑖𝑛𝑒𝑤 ,𝑆1 , 𝐹𝑖𝑛𝑒𝑤 ,𝑆1 < 𝐹𝑖

𝑋𝑖 = { (3.16)
𝑋𝑖 , 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒

In Equation (3.15), 𝑥𝑖 ,𝑗 represents the candidate value from the candidate solution for
the jth variables, r denotes a random number in the range of [0,1], l is a predefined
parameter of value 1 or 2 whereas, 𝐹𝐶𝑖 is a value of objective fitness function for
selected carrion.
In Equation (3.16), 𝑋𝑖𝑛𝑒𝑤 ,𝑆1 denotes the new position of ith Tasmanian devil and
𝑥𝑖𝑛𝑒𝑤
,𝑗
,𝑆1
represents its value for jth variable and 𝐹𝑖𝑛𝑒𝑤 ,𝑆1 represents its objective
function value.
3.2.2. Exploitation Phase (Feeding by eating prey)
In this approach, there are two steps that Tasmanian devil follows during the attacking
behavior. In the first step, it chooses the prey and launches an attack on it after
scanning the surroundings. After reaching the prey, it chases it in the second step to
stop it and begin feeding. Therefore, the first step is designed in the same manner as

27
designed for the first approach of Tasmanian devil feeding. The process of choosing
prey is mimicked in Equation (3.17).
𝑃𝑖 = 𝑋𝑘 , 𝑖 = 1, 2, 3, . . . . , 𝑁, 𝑘 𝜖 {1, 2, . . . , 𝑁|𝑘 ≠ 𝑖} (3.17)
Where, 𝑃𝑖 denotes the prey selected by the ith Tasmanian devil. A new position for the
Tasmanian devil in the search space is determined based on the chosen prey. The
updated location of the Tasmanian devil can be formulated using Equation (3.18).

𝑥𝑖 ,𝑗 + 𝑟 ∗ (𝑝𝑖 ,𝑗 − 𝑙 ∗ 𝑥𝑖 ,𝑗 ), 𝐹𝑝𝑖 < 𝐹𝑖 (3.18)

𝑥𝑖𝑛𝑒𝑤
,𝑗
,𝑆2
= {
𝑥𝑖 ,𝑗 + 𝑟 ∗ (𝑥𝑖 ,𝑗 − 𝑝𝑖 ,𝑗 ), 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒

𝑋𝑖𝑛𝑒𝑤 ,𝑆2 , 𝐹𝑖𝑛𝑒𝑤 ,𝑆2 < 𝐹𝑖

𝑋𝑖 = { (3.19)
𝑋𝑖 , 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒

In Equation (3.18), 𝑥𝑖 ,𝑗 represents the candidate value from the candidate solution for
the jth variables, r denotes a random number in the range of [0,1], l is a predefined
parameter of value 1 or 2 whereas, 𝐹𝑝𝑖 is a value of objective fitness function for
selected prey.
In Equation (3.19), 𝑋𝑖𝑛𝑒𝑤 ,𝑆1 denotes the new position of ith Tasmanian devil and
𝑥𝑖𝑛𝑒𝑤
,𝑗
,𝑆1
represents its value for jth variable and 𝐹𝑖𝑛𝑒𝑤 ,𝑆1 represents its objective
function value.
The second step during the attacking behavior by Tasmanian devil is, chasing process,
which is designed using Equations (3.20) to (3.22). At this point, the Tasmanian devil
location is regarded as the hub of the area where the process of pursuing prey is
occurring. The Tasmanian devil follows its victim over a range that corresponds to this
neighborhood's radius, which is computed using Equation (3.20).
𝑡
𝑅 = 0.01 (1 − ) (3.20)
𝑇
Here, R represents the attack location’s immediate surrounding, t denotes the current
iteration and T represents the maximum iterations. A new position for the Tasmanian
devil is determined based on the chasing task. The updated location of the Tasmanian
devil can be formulated using Equation (3.21).
(3.21)
𝑥𝑖𝑛𝑒𝑤
,𝑗 = 𝑥𝑖 ,𝑗 + (2 ∗ 𝑟 − 1) ∗ 𝑅 ∗ 𝑥𝑖 ,𝑗

𝑋𝑖𝑛𝑒𝑤 , 𝐹𝑖𝑛𝑒𝑤 < 𝐹𝑖

𝑋𝑖 = { (3.22)
𝑋𝑖 , 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒

28
In Equation (3.22), 𝑋𝑖𝑛𝑒𝑤 denotes the new position of ith Tasmanian devil in the
district of 𝑋𝑖 and 𝑥𝑖𝑛𝑒𝑤
,𝑗 represents its value for jth variable and 𝐹𝑖
𝑛𝑒𝑤
represents its
objective function value. The process of TDO technique to address OADG in DN is
depicted in Figure 3.3.

Start

Initialize TDO
Yes
parameters N and T Pr > 0.5 selecting Prey
Calculate the objective
fitness function for
Random locations are No
Calculate the objective chasing process
generated for tasmanian
devils fitness function for each

t = t + 1
selecting carrion
chosen Prey Update the position of
Tasmanian devil based
t=1 Calculate the objective Update the position of on fitness value
fitness function for each Tasmanian devil based
chosen carrion on fitness value
No
Calculate the objective t Maxit
fitness function for each Update the position of
tasmanian devil Update the R
Tasmanian devil based
Yes
on fitness value
Get optimal solution
containg optimal DG
location and DG size

End

Figure 3.3: Flow chart of GTO algorithm to solve problem of OADG

29
Summary
In this chapter, an overview of the optimization approaches GTO and TDO is
presented along with the flow chart for solving the problem of optimal sitting and
sizing of DGs in distribution network using the optimization approaches.

30
References
[1] B. Abdollahzadeh, F. Soleimanian Gharehchopogh, and S. Mirjalili, “Artificial
gorilla troops optimizer: A new nature-inspired metaheuristic algorithm for
global optimization problems,” Int. J. Intell. Syst., vol. 36, no. 10, pp. 5887–
5958, 2021, doi: 10.1002/int.22535.
[2] M. Dehghani, S. Hubalovsky, and P. Trojovsky, “Tasmanian Devil
Optimization: A New Bio-Inspired Optimization Algorithm for Solving
Optimization Algorithm,” IEEE Access, vol. 10, pp. 19599–19620, 2022, doi:
10.1109/ACCESS.2022.3151641.

31
 Chapter 4: Problem Formulation
4.1. Introduction
In the planning of the DSs, two things need to be addressed carefully i.e., to find the
location where DG installed and the second one is to find the correct DG capacity to
be installed at that location. The selection of DG sitting and capacity in such a way
that it minimizes the system losses and improves the voltage profile of the system
along with voltage drops while satisfying the system constraints.
4.2. Objective Functions
The main objective of this study is to optimally allocate the DGs to solve SOF and
MOF. The SOF includes reduction of total power losses i.e., APL and QPL while MOF
includes reduction of APL, minimization of VD, maximization of VSI and
minimization of total OC including investment, operation and maintenance costs of
DGs. The mathematical modelling for SOF and MOF is presented as:
4.2.1 Reduction of Active Power Loss (APL)
In the RDN, the APL is computed by using the Equation (4.1) [1], [2]:
𝑀𝑏𝑟
2
𝐴𝑃𝐿 = ∑|𝐼𝑗 | . 𝑅𝑗 (4.1)
𝑗=1

Here, j is a branch number, total number of branches is represented by 𝑀𝑏𝑟 , |𝐼𝑗 |

represents the absolute current passing through the branch of resistance 𝑅𝑗 . The first
objective function (𝑂𝐹1 ) is represented in Equation (4.2):
𝑂𝐹1 = 𝑀𝑖𝑛(𝐴𝑃𝐿) (4.2)

4.2.2. Reduction of Reactive Power Loss (QPL)

In the RDN, the QPL is computed by using the Equation (4.3):
𝑀𝑏𝑟
2
𝑄𝑃𝐿 = ∑|𝐼𝑗 | . 𝑋𝑗 (4.3)
𝑗=1

Here, j is a branch number, total number of branches is represented by 𝑀𝑏𝑟 , the

absolute current |𝐼𝑗 | represents the absolute current passing through the branch of
inductive reactance 𝑋𝑗 . The second objective function 𝑂𝐹2 is represented in Equation
(4.4):
𝑂𝐹2 = 𝑀𝑖𝑛(𝑄𝑃𝐿) (4.4)

32
4.2.3. Minimization of Voltage Deviation (VD)
The lifetime and performance of the equipment are affected by the deviation of the
voltages. So, keep in mind that the reason minimization of voltage deviation is taken
as OF. VD is calculated by using the Equation (4.5) [3]:
𝑚

𝑉𝐷 = ∑ (𝑉𝑟𝑒𝑓 − 𝑉𝑗 )2 (4.5)
𝑗=1

Where, 𝑉𝑟𝑒𝑓 denotes the reference voltage of value 1.0 p.u. The third OF (𝑂𝐹3 ) is
represented in Equation (4.6):
𝑂𝐹3 = 𝑀𝑖𝑛 (𝑉𝐷) (4.6)
4.2.4. Maximization of Voltage Stability Index (VSI)
Along with VD, VSI is another crucial component to take into account when
determining the security level of DN. Due to a variety of factors, when a bus in DN
exceeds permitted voltage limits, the entire system may experience voltage instability,
or VSI. All the buses in a DN must maintain VSI at a stable limit for steady operation.
Equation (4.7) is used to compute the VSI [4].
2
𝑉𝑆𝐼(𝑘+1) = |𝑉𝑘 |4 − 4 (𝑃(𝑘+1) 𝑥𝑘 − 𝑄(𝑘+1) 𝑟𝑘 )
(4.7)
− 4 |𝑉𝑘 |2 (𝑃(𝑘+1) 𝑟𝑘 − 𝑄(𝑘+1) 𝑥𝑘 )

In the equation (4.7), 𝑃(𝑘+1) and 𝑄(𝑘+1) represents the real and reactive load demands
respectively, whereas the line resistance is denoted by 𝑟𝑘 and the line reactance is
represented by 𝑥𝑘 . For secure and stable operation of the RDN, VSI must be greater
than zero. The fourth OF (𝑂𝐹4 ) is represented in Equation (4.8):
𝑂𝐹4 = 𝑀𝑎𝑥(𝑀𝑖𝑛 (𝑄𝑃𝐿)) (4.8)
4.2.5. Minimization of total operating cost (TOC)
The total price to obtain electricity from the grid consists of the price of power loss
and the price of power supplied to the customer. The total operating cost before the
DG integration is computed through the following equation [5]:
𝑇𝑝

𝑂𝐶𝑏𝑎𝑠𝑒 = ∑ 𝑃𝑊𝐹 𝑡 × (𝑃𝐿 + 𝑃𝑇𝑙𝑜𝑎𝑑 ) × 𝑇 × 𝐶𝑒 (4.9)

𝑡=1

Where, 𝑃𝐿 denotes the power loss in the base case and 𝑃𝑇𝑙𝑜𝑎𝑑 denotes the supplying
power to the customer, T refers to the time period (hr./year), 𝑇𝑝 refers to the
planning years and the cost of energy price 𝐶𝑒 is 49 ($/MWh). 𝑃𝑊𝐹 is a present
worth factor and it is evaluated by using the Equation (4.10):

33
 1 + 𝐼𝑛𝑓_𝑅
𝑃𝑊𝐹 = (4.10)
1 + 𝑖𝑛𝑡_𝑅
Here, inf _𝑅 is the inflation rate having the value of 9% and 𝑖𝑛𝑡_𝑅 is the interest rate
having value of 12.5%.
The annual operational cost can be broken down into the following four elements: The
first element is the maintenance cost of DG that is determined using the Equation
(4.11):
𝑇𝑝 𝑁𝑑𝑔

𝐷𝐺𝑚 = (∑ ∑ 𝑃𝑊𝐹 𝑡 × 𝑃𝐷𝐺,𝑖 × 𝑇 × 𝐾𝑀,𝐷𝐺 ) (4.11)

𝑡=1 𝑖=1

Where, 𝐾𝑀,𝐷𝐺 denotes the DG maintenance cost and its value is taken 7 ($/MWh),
𝑃𝐷𝐺,𝑖 represents the power of installed DG, 𝑁𝑑𝑔 represents the number of DG installed
units.
The second element is the DG operational cost which is computed through the
Equation (4.12):
𝑇𝑝 𝑁𝑑𝑔

𝐷𝐺𝑂 = (∑ ∑ 𝑃𝑊𝐹 𝑡 × 𝑃𝐷𝐺,𝑖 × 𝑇 × 𝐾𝑂,𝐷𝐺 ) (4.12)

𝑡=1 𝑖=1

Where, 𝐾𝑂,𝐷𝐺 denotes the DG operation cost and its value is taken 29 ($/MWh).
The third component is the DG installation cost, that is, computed through the
Equation (4.13):
𝑁𝑑𝑔

𝐷𝐺𝐼𝐶 = 𝐾𝐼𝐶,𝐷𝐺 ∑ 𝑃𝐷𝐺,𝑖 (4.13)

𝑖=1

Where, 𝐾𝐼𝐶,𝐷𝐺 is the DG installation cost and its value is taken 400,000 ($/MW).
The last element is the cost of operation after DG integration which is calculated using
the Equation (4.14):
𝑇𝑝

𝑂𝐶𝐷𝐺 = ∑ 𝑃𝑊𝐹 𝑡 × (𝑃𝐿 + 𝑃𝑇𝑙𝑜𝑎𝑑 ) × 𝑇 × 𝐶𝑒 (4.14)

𝑡=1

Where, 𝑃𝐿 denotes the power loss after integrating DG and 𝑃𝑇𝑙𝑜𝑎𝑑 denotes the
supplying power to the customer. Thus, the annually operating cost can be constructed
as:
𝑂𝐶𝑇𝐷𝐺 = 𝐷𝐺𝑀 + 𝐷𝐺𝑂 + 𝐷𝐺𝐼𝐶 + 𝑂𝐶𝐷𝐺 (4.15)

34
The fifth OF (𝑂𝐹5 ) is represented as:
𝑂𝐹5 = 𝑀𝑖𝑛 (𝑂𝐶𝑇𝐷𝐺 ) (4.16)
4.3. Formulation of Multi Objective Function (MOF)
Unlike SOF, in MOF two or more conflicting objective functions are deal
simultaneously. To get TEOF, each of the objective functions is divided by its base
value and corresponding weight. For MOF, the weighted sum method is used. The
TEOF is formulated in Equation (4.17):
𝑅𝑃𝐿 𝑇𝑉𝐷 𝑉𝑆𝐼 −1
𝑇𝐸𝑂𝐹 = 𝑀𝑖𝑛 (𝛼1 × + 𝛼2 × + 𝛼3 × −1
𝑅𝑃𝐿𝑏𝑎𝑠𝑒 𝑇𝑉𝐷𝑏𝑎𝑠𝑒 𝑉𝑆𝐼𝑏𝑎𝑠𝑒
(4.17)
𝑂𝐶𝑇𝐷𝐺
+ 𝛼4 × )
𝑂𝐶𝑏𝑎𝑠𝑒
Also,
4

∑ 𝛼𝑖 = 1; 𝛼 ∈ [0,1] (4.18)
𝑖=1

−1
Where, 𝑅𝑃𝐿𝑏𝑎𝑠𝑒 , 𝑇𝑉𝐷𝑏𝑎𝑠𝑒 , 𝑉𝑆𝐼𝑏𝑎𝑠𝑒 and 𝑂𝐶𝑏𝑎𝑠𝑒 are the base values of real power
losses, total voltage deviation, voltage stability index and operating cost respectively.
𝛼1 , 𝛼2 , 𝛼3 and 𝛼4 are the different weights given to the individual objective function.
Their values are 0.5, 0.35, 0.15 and 0.1 respectively. The weights shown preference is
given to the technical objectives over economic objective.
4.4. Constraints
The MOF that is in Equation (4.17) is subjected to the following constraints:
4.4.1.Equality Constraints
The flow of power in the distribution system must be balanced. In other words, power
drawn from the substation and the power generated from the DG units should be
sufficient to meet the following requirements for load demand and power losses [5].
𝑁𝑑𝑔

𝑃𝑠 + ∑ 𝑃𝐺 (𝑖) = 𝑃𝐷 + 𝐴𝑃𝐿 (4.19)

𝑖=1

𝑁𝑑𝑔

𝑄𝑠 + ∑ 𝑄𝐺 (𝑖) = 𝑄𝐷 + 𝑄𝑃𝐿 (4.20)

𝑖=1

Where, 𝑃𝑠 and 𝑄𝑠 are the active and reactive powers drawn from the substation
respectively. 𝑃𝐺 and 𝑄𝐺 are the incoming powers that generated from the installed DG

35
units. 𝑃𝐷 , 𝑄𝐷 and 𝑅𝑃𝐿 and 𝑄𝑃𝐿 are the active and reactive load demands and real and
reactive power losses respectively.
4.4.2.Equality Constraints
4.4.2.1. Voltage Limits
The voltage magnitude of the bus in the network should be within permissible limits,
that is, expressed in Equation (4.21) [1],[6]:
0.95𝑝. 𝑢 ≤ 𝑉𝑖 ≤ 1.05𝑝. 𝑢 (4.21)
4.4.2.2. Branch Current Limits
The maximum limit of a branch's current should not be exceeded [1], as stated by:
𝐼𝑗 ,𝑖 ≤ 𝐼𝑗𝑚𝑎𝑥
,𝑖 (4.22)
4.4.2.3. DG Capacity Limits
The range of the output power of DG units is expressed as [5],[7]:
𝑃𝐺𝑚𝑖𝑛 ≤ 𝑃𝐺 ≤ 𝑃𝐺𝑚𝑎𝑥 (4.23)

𝑄𝐺𝑚𝑖𝑛 ≤ 𝑄𝐺 ≤ 𝑄𝐺𝑚𝑎𝑥 (4.24)

4.4.2.4. DG Power Factor Limits
The following range of power factors can be used to operate DG units [1],[8]:
𝑚𝑖𝑛 𝑚𝑎𝑥
𝑝. 𝑓𝐷𝐺 ≤ 𝑝. 𝑓𝐷𝐺 ≤ 𝑝. 𝑓𝐷𝐺 (4.25)

4.5. Load Modelling

Most of the earlier studies focused on the power flow issue using constant load models,
or constant active and reactive powers. However, the characteristics of actual loads
show how much power demands depend on system voltage. The distribution system's
load can often be divided into three categories: industrial, residential, and commercial
load. In this study, real life problems are considered by modelling voltage dependent
load models such as: industrial, residential and commercial load models. The
mathematical formulation of voltage dependent load models is expressed in Equation
(4.26) and (4.27):
𝑃 = 𝑃𝑜 𝑉𝑛𝛼 (4.26)
𝛽
𝑄 = 𝑄𝑜 𝑉𝑛 (4.27)
Here 𝑃 and 𝑄 represent the values of active and reactive powers respectively; 𝑃𝑜 and
𝑄𝑜 refer to the values of active and reactive power at nominal voltage respectively; 𝑉𝑛
denotes the magnitude of the voltage whereas α and β are the values of the active and

36
reactive power exponents. The values of the α and β used in this study for various
types of loads are shown in Table 4.1 taken from [6],[9]
Table 4.1: Load types and exponent values

Load types α β
Constant 0 0
Residential 0.92 4.04
Commercial 1.51 3.40
Industrial 0.18 6.0

37
Summary
In this chapter, mathematical modeling of the objective functions consisting of APL
and QPL minimization, minimization of VD and TOC along with the VSI
maximization is described. Also, the load modelling of the VP load models is
formulated.

38
References
[1] M. I. Akbar, S. A. A. Kazmi, O. Alrumayh, Z. A. Khan, A. Altamimi, and M.
M. Malik, “A Novel Hybrid Optimization-Based Algorithm for the Single and
Multi-Objective Achievement with Optimal DG Allocations in Distribution
Networks,” IEEE Access, vol. 10, pp. 25669–25687, 2022, doi:
10.1109/ACCESS.2022.3155484.
[2] A. Ehsan and Q. Yang, “Optimal integration and planning of renewable
distributed generation in the power distribution networks: A review of
analytical techniques,” Appl. Energy, vol. 210, no. November 2017, pp. 44–59,
2018, doi: 10.1016/j.apenergy.2017.10.106.
[3] T. P. Nguyen, T. A. Nguyen, T. V. H. Phan, and D. N. Vo, “A comprehensive
analysis for multi-objective distributed generations and capacitor banks
placement in radial distribution networks using hybrid neural network
algorithm,” Knowledge-Based Syst., vol. 231, p. 107387, 2021, doi:
10.1016/j.knosys.2021.107387.
[4] S. Arabi Nowdeh et al., “Fuzzy multi-objective placement of renewable energy
sources in distribution system with objective of loss reduction and reliability
improvement using a novel hybrid method,” Appl. Soft Comput. J., vol. 77, pp.
761–779, 2019, doi: 10.1016/j.asoc.2019.02.003.
[5] N. R. Godha, V. N. Bapat, and I. Korachagaon, “Ant colony optimization
technique for integrating renewable DG in distribution system with techno-
economic objectives,” Evol. Syst., vol. 13, no. 3, pp. 485–498, 2022, doi:
10.1007/s12530-021-09416-y.
[6] S. Kamel, A. Awad, H. Abdel-Mawgoud, and F. Jurado, “Optimal DG
allocation for enhancing voltage stability and minimizing power loss using
hybrid gray Wolf optimizer,” Turkish J. Electr. Eng. Comput. Sci., vol. 27, no.
4, pp. 2947–2961, 2019, doi: 10.3906/elk-1805-66.
[7] S. Nagaballi and V. S. Kale, “Pareto optimality and game theory approach for
optimal deployment of DG in radial distribution system to improve techno-
economic benefits,” Appl. Soft Comput. J., vol. 92, p. 106234, 2020, doi:
10.1016/j.asoc.2020.106234.
[8] A. Eid, “Allocation of distributed generations in radial distribution systems
using adaptive PSO and modified GSA multi-objective optimizations,”

39
 Alexandria Eng. J., vol. 59, no. 6, pp. 4771–4786, 2020, doi:
10.1016/j.aej.2020.08.042.
[9] M. C. V. Suresh and E. J. Belwin, “Optimal DG placement for benefit
maximization in distribution networks by using Dragonfly algorithm,”
Renewables Wind. Water, Sol., vol. 5, no. 1, 2018, doi: 10.1186/s40807-018-
0050-7.

40
 Chapter 5: Results and Discussion
5.1. Introduction
To evaluate the validity and effectiveness of the proposed algorithms, they have been
implemented on two IEEE benchmark RDSs including IEEE 33 bus and IEEE 69 bus
systems to optimize the SOF and MOF considering CP load model as well as VP loads
models consisting of residential, commercial and industrial load models. The
reduction of APL and QPL by optimal placement and sizing of DG units is
accomplished by optimizing SOF while minimization of APL, VD, OCI and
maximization of VSI simultaneously is achieved with optimizing MOF. The control
parameter settings for the suggested methods are demonstrated in Table 5.1.
In this study, PV-DG and WT-DGs are considered. PV-DG having unity pf while WT-
DG having combined pf (0.85). The period considered for this study is 20 years. The
proposed algorithms are simulated using MATLAB 2021b software and Windows
2010Pro with Intel(R) Core (TM) i5-4210U processors and 8 GB of RAM. To verify
the robustness of the algorithms, following scenario and cases are considered for
studied systems:
• Scenario 1: Optimizing SOF by integrating 3 DGs
o Case 1: Minimizing APL for both CP and VP load models
o Case 2: Minimizing QPL for both CP and VP load models
• Scenario 2: Optimizing MOF by integrating 3 DGs
o Case 1: Minimizing APL, VD, OCI and maximizing VSI for both CP
and VP load models
Table 5.1: Control parameters and settings

Algorithms Parameters
GTO population=50, max-iter = 200, DG-size (MVA) = 0-2000, β=3,
p=0.03, W=0.8
TDO population=50, max-iter = 200, DG-size (MVA) = 0-2000

5.2. IEEE 33 bus test system

The proposed algorithms are first tested on standard IEEE 33 bus RDS. The single
line diagram of the 33 bus RDS is shown in Figure 5.1. It consists of 33 buses and 32
branches and the information about the line and load data of the system is obtained
from [1]. The total load demand of the system is (3.715 + j2.300) MVA with the base
voltage and base MVA of 12.66 KV and 100 MVA respectively.

41
 23 24 25 26 27 28 29 30 31 32 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

19 20 21 22

Figure 5.1: Single-line diagram of 33 bus RDS

5.2.1. Scenario 1: Evaluation of Single Objective Function (SOF)

In this scenario, TDO and GTO are used to solve the problem of optimal sitting and
sizing of DGs to reduce the total active and reactive power losses as SOF for CP load
model and VP load model i.e., industrial, residential and commercial load models for
three PV-DG (unity pf) and three WT-DG (0.85 pf).
5.2.1.1. CP load Model
The total power loss of the system without integration of the DGs are calculated by
using Forward-backward power flow method is (210.07KW + j142.44KVAR), the
minimum voltage is 0.904 p.u at bus 18, the VD is 0.1328 p.u and the VSI is 0.6672
p.u. Table 5.2 represents the results of OADG in CP load model to optimize the SOF
at unity pf (PV-DG). It can be seen from the table that by using both TDO and GTO
algorithms, total APL are reduced to 70.64 KW that is 66.37% reduction with respect
to base value. The results are compared with other existing optimization techniques.
The result of APL reduction is better than other BAT [6], SOS-NNA [7], IHHO [5],
GAMS [2], CFA [4] and QOFBI [3] mentioned in the table 5.2.
Likewise, Table 5.3 shows the results of OADG at unity pf (PV-DG) for QPL
minimization. It can be seen from Table 5.3 that with the integration of the DGs, QPL
is reduced to 49.22 KVAR for both algorithms, that is 65.44% reduction with respect
to base value.
Moreover, the results of OADG for APL minimization at 0.85 pf (WT-DG) are
presented in Table 5.4. By employing the TDO and GTO, the APL is reduced to 14.39
KW that is 93.15% reduction, that can be seen from the Table 5.4. The obtained results

42
are compared with other optimization approach I-DBEA [30]. The purposed
algorithms give better results in terms of APL reduction.
Table 5.2: Results comparison of OADG for 33 bus RDN for APL minimization for
PV-DG in CP load model

APL (KW)
Methods Location DG size (KW) VD (p.u) VSI (p.u)
APLR
Base case - - 210.07 0.1328 0.6672
72.78
BAT [6] 13,25,30 380, 490, 990 - 0.8652
(65.5%)
SOS-NNA 801.8, 1091.3, 72.785
13,24,30 0.015113 1.1358
[7] 1053.6 (65.5%)
775.54, 1080.83, 72.79
IHHO [5] 14,24,30 - -
1066.69 (65.50%)
770.9, 1096.9, 72.79
GAMS [2] 14,24,30 - -
1065.8 (65.50%)
1059.32, 1090.16, 72.79
CFA [4] 30,24,13 - -
801.88 (65.50%)
1091.33, 1053.64, 72.78
QOFBI [3] 24,30,13 - -
801.71 (65.5%)
70.64
TDO [P] 30,14,24 1213, 866, 1186 0.011541 0.8940
(66.37%)
70.64
GTO [P] 24,30,14 1186, 1213, 866 0.011541 0.8940
(66.37%)

Table 5.3: Results of OADG for 33 bus RDN for QPL minimization for PV-DG in
CP load model

QPL (KVAR)
Methods Location DG Size (KW) VD (p.u) VSI (p.u)
QPLR
Base case - - 142.44 0.1328 0.6672
49.22
TDO [P] 13, 24, 30 926, 1135, 1160 0.011766 0.8901
(65.44%)
49.22
GTO [P] 24, 30, 13 1135, 1160, 926 0.011766 0.8901
(65.44%)

Table 5.4: Results of OADG for 33 bus RDN for APL minimization for WT-DG in
CP load model

DG size APL (KW) VD VSI

Methods Location
KW KVAR APLR (p.u) (p.u)
Base case - - - 210.07 0.1328 0.6672
749.1, 1042, 14.57
I-DBEA [8] 13,24,30 - 0.0002 0.9733
1239.5 (92.81%)
1333, 1147, 826,640, 14.39
TDO [P] 30,24,13 0.000604 0.9669
836 486 (93.15%)
1147, 1333, 640,826, 14.39
GTO [P] 24,30,13 0.000604 0.9669
836 486 (93.15%)

43
 Table 5.5: Results of OADG for 33 bus RDN for QPL minimization for WT-DG in
CP load model

DG Size QPL (KVAR)

Methods Location VD (p.u) VSI (p.u)
(KW) (KVAR) QPLR
Base case - - - 142.44 0.1328 0.6672
1104,867, 599,494 11.68
TDO [P] 24,13,30 0.000631 0.9671
1298 804 (91.80%)
1102,869, 599,492, 11.68
GTO [P] 24,13,30 0.000631 0.9671
1299 805 (91.80%)

Similarly, the results of OADG for QPL minimization at 0.85 pf (WT-DG) are
presented in Table 5.5. The QPL is reduced to 11.68 KVAR for both GTO and TDO
algorithms after integration of DGs, which is 91.80% reduction with respect to base
value.
5.2.1.2. VP load Model
This section extends the OADG issue for practical non-linear loads to show how
strongly power demands rely on voltage of the network. In the base case, the APL for
industrial is 163.22KW, for residential is 158.76KW and for commercial is 152.32
KW. Table 5.6 shows the results of SOF based on APL minimization for optimum
placement and sizing of DGs at unity pf (PV-DG).
Table 5.6: Results comparison of OADG for 33 bus RDN for APL minimization for
PV-DG in CP load model
Methods Parameters Industrial Residential Commercial
APL (KW) 163.22 158.76 152.32
Base case VD (p.u) 0.098637 0.097821 0.094018
VSI (p.u) 0.70741 0.70953 0.71526
Location 13, 25, 30 13, 25, 30 13, 25, 30
DG size (KW) 790, 850, 1020 720, 830, 980 710, 820, 940
BAT [6] APL (KW) 36.01 43.70 46.42
APLR (%) 77.99 72.53 69.57
VD (p.u) - - -
VSI (p.u) 0.9069 0.8974 0.8959
Location 14, 24, 30 24, 30, 14 24, 30, 14
DG size (KW) 843, 1187, 1163 1158, 1108, 793 1132, 1065, 756
TDO [P] APL (KW) 34.55 42.05 44.80
APLR (%) 78.83 73.51 70.59
VD (p.u) 0.005841 0.007153 0.007595
VSI (p.u) 0.9245 0.9164 0.9139
Location 14, 24, 30 24, 14, 30 30, 24, 14
DG size (KW) 843, 1187, 1163 1158, 793, 1108 1064, 1136, 755
APL (KW) 34.55 42.05 44.80
GTO [P] APLR (%) 78.83 73.51 70.59
VD (p.u) 0.005841 0.007152 0.007599
VSI (p.u) 0.9245 0.9164 0.9139

44
It can be seen from the table that by employing TDO and GTO algorithms, APL is
reduced to 34.55 KW (78.83%) for industrial, 42.05 KW (73.51%) for residential and
44.80 KW (70.59%) for commercial load models. The results are compared with other
optimization methods, BAT [10a] mentioned in Table 5.6. The results obtained by
employing the purposed algorithms are better in terms of APL reduction.
Table 5.7: Results of OADG for 33 bus RDN for QPL minimization for PV-DG in
VP load model

Parameters Industrial Residential Commercial

QPL (KVAR) 110.42 107.19 102.70

Base case VD (p.u) 0.098637 0.097821 0.094018

VSI (p.u) 0.70741 0.70953 0.71526
QPL (KVAR) 24.54 29.48 31.24
QPLR (%) 77.77 72.49 65.68
Location 13, 30, 24 30, 24, 13 13. 24, 30
TDO [P] DG Size (KW) 901, 1112, 1130 1059, 1105, 849 808,1085,1016
DG Size (KVAR) - - -
VD (p.u) 0.005964 0.007305 0.007765
VSI (p.u) 0.9208 0.9127 0.9103
QPL (KVAR) 24.54 29.48 31.24
QPLR (%) 77.77 72.49 65.68
Location 13, 30, 24 24, 30, 13 13, 30, 24
GTO [P]
DG Size (KW) 901, 1112. 1130 1105, 1059, 849 808,1016,1085
DG Size (KVAR) - - -
VD (p.u) 0.005964 0.007305 0.007765
VSI (p.u) 0.9208 0.9127 0.9103

Likewise, the QPL in the base case for industrial load model is 110.42 KVAR, for
residential is 107.19 KVAR and for commercial is 102.70 KVAR. The results of SOF
based on QPL minimization for optimum placement and sizing of multiple DGs at
unity pf (PV-DG) are presented in Table 5.7. It can be seen that, after the integration
of DGs, the QPL is reduced to 24.54 KVAR (77.77%) for industrial load model, for
residential APL is reduced to 29.48 KVAR (72.49%) and for commercial it is reduced
to 31.24 KVAR (65.68%) for both TDO and GTO algorithms.
Furthermore, Table 5.8 represents the results of OADG at 0.85 pf (WT-DG) for APL
minimization for different load models. The APL is reduced to 10.53 KW (93.55),
10.38 KW (93.46%) and 10.33 KW (93.22%) for industrial, residential and
commercial load models respectively for both TDO and GTO algorithms.

45
 Table 5.8: Results of OADG for 33 bus RDN for APL minimization for WT-DG in
VP load model

Parameters Industrial Residential Commercial

APL (KW) 163.22 158.76 152.32
VD (p.u) 0.098637 0.097821 0.094018
Base case
VSI (p.u) 0.70741 0.70953 0.71526
APL (KW) 10.53 10.38 10.33
APLR (%) 93.55 93.46 93.22
Location 24, 13, 30 30, 13, 24 30, 24, 13
TDO [P] DG Size (KW) 1181, 869, 1129 1137, 798, 1139 1123, 1111, 750
DG Size (KVAR) 477, 250, 700 704, 321, 531 696, 553, 349
VD (p.u) 0.000517 0.000507 0.000486
VSI (p.u) 0.9694 0.9698 0.9705
APL (KW) 10.53 10.38 10.33
APLR (%) 93.55 93.46 93.22
Location 24, 13, 30 24, 13, 30 30, 24, 13
GTO [P] DG Size (KW) 1182, 868, 1129 1140, 798, 1137 1123, 1111, 750
DG Size (KVAR) 477, 249, 700 532, 321, 705 696, 553, 350
VD (p.u) 0.000516 0.000505 0.000486
VSI (p.u) 0.9694 0.9698 0.9705

Table 5.9: Results of OADG for 33 bus RDN for QPL minimization for WT-DG in
VP load model

Parameters Industrial Residential Commercial

QPL (KVAR) 110.42 107.19 102.70
Base case VD (p.u) 0.098637 0.097821 0.094018
VSI (p.u) 0.70741 0.70953 0.71526
QPL (KVAR) 8.66 8.50 8.41
QPLR (%) 92.15 92.07 91.81
Location 24, 13, 30 24, 30, 13 13, 24, 30
TDO [P] DG Size (KW) 1128,895,1097 1092, 1105, 825 777, 1065, 1092
DG Size (KVAR) 457, 259, 680 505, 685, 329 356, 523, 677
VD (p.u) 0.000542 0.000533 0.000518
VSI (p.u) 0.9694 0.9699 0.9705
QPL (KVAR) 8.66 8.50 8.41
QPLR (%) 92.15 92.07 91.81
Location 30, 13, 24 24, 13, 30 30, 24, 13
GTO [P] DG Size (KW) 1097,895,1128 1091, 1106, 825 1092, 1065, 777
DG Size (KVAR) 680, 259, 457 504, 685, 329 677, 522, 356
VD (p.u) 0.000541 0.000535 0.000517
VSI (p.u) 0.9695 0.9699 0.9705

Similarly, the results of OADG for QPL reduction at 0.85 pf (WT-DG) for different
load models are described in Table 5.9. The QPL for industrial load model is reduced
to 8.66 KVAR (92.15%), for residential load model QPL is reduced to 8.50 KVAR
(92.07%) and for commercial load models QPL is reduced to 8.41 KVAR (91.81%)
for both TDO and GTO algorithms.

46
5.2.1.3. Voltage profiles for 33 bus RDS
Figure 5.2 illustrates the impact of the installation of various types of DGs on the
voltage profile for the 33 bus RDS under various types of loads i.e., constant,
residential, commercial, and industrial. In Figure 5.2, “const” stands for constant,
“Res” denotes the residential, “Comm” denotes the commercial and” Ind” stands for
Industrial loads. The outer values of the graphs represent the “bus number” and the
inner values represent “ bus voltage” corresponding to the particular bus number.

Figure 5.2: Voltage profiles for IEEE 33 bus system across various loads at different
cases for scenario 1
5.2.1.4. Active power losses for 33 bus RDS
Figure 5.3 depicts the APL information for each bus of 33 bus RDS after integration
of three PV-DG (unity pf) and three WT-DG (0.85 pf) under various types of loads
i.e., constant, residential, commercial and industrial under scenario 1.
5.2.1.5. Convergence characteristics for 33 bus RDS
Figures 5.4(a) and 5.4(b) display the convergence characteristics of GTO and TDO at
various power factors (unity and 0.85pf) for CP load model . It is apparent from the
below figure that GTO has higher efficiency than TDO.

47
 Figure 5.3: Active power loss for 33 buses under various types of loads

(a)

(b)
Figure 5.4: Convergence characteristics of TDO and GTO for IEEE 33 bus system at
different pf; a) Unity pf (b) 0.85 pf

48
5.2.2. Scenario 2: Evaluation of Multi-Objective Function (MOF)
In this scenario, TDO and GTO algorithms are employed to optimize the MOF based
on APL, VD, VSI and OCI to solve the problem of OADG in RDS considering CP
load model and VP load models for three PV (unity pf) and three WT (0.85 pf).
5.2.2.1. CP load Model
The total power loss of the system in the base case is (210.07KW + j142.44KVAR),
the VD is 0.1328 p.u, VSI is 0.6672 and operating cost is 16.837972 (M$). Table 5.10
presents the results of MOF in terms of APL reduction, VD minimization, VSI
maximization and minimization of operating costs index for optimum placement and
sizing of DGs for three PV (unity pf) and three WT (0.85 pf) for CP load model.
It can be seen from Table 5.10 that at unity pf (PV-DG), the APL is reduced to 71.74
KW (65.85%) for TDO and 72.02 KW (65.72%) for GTO. The VD obtained by
employing TDO and GTO is 0.007390 p.u and 0.006894 p.u respectively while the
value of VSI computed for TDO and GTO is 0.9160 p.u and 0.9189 p.u respectively
and the operating cost index for both TDO and GTO is 8084 and 8075 respectively.
Furthermore, the cost savings after 20 years at unity pf (PV-DG) are 19.16% and
19.25% for GTO and TDO respectively.
Table 5.10: Results of OADG for 33 bus RDN for MOF in CP load model

PV-DG WT-DG
Parameters Base case
TDO [P] GTO [P] TDO [P] GTO [P]
APL (KW) 210.07 71.74 72.02 14.65 14.66
APLR (%) - 65.85 65.72 93.03 93.02
Location - 30, 24, 14 14, 30, 24 13, 30, 24 13, 30, 24
DG Size 1371, 1282, 922, 1393, 884, 1368, 874, 1367,
-
(KW) 913 1273 1205 1239
DG Size 504, 848, 509, 847,
- -
(KVAR) 660 643
VD (p.u) 0.1328 0.007390 0.006894 0.000287 0.000291
VSI (p.u) 0.6672 0.9160 0.9189 0.9753 0.9753
OCDG (M$) 16.837972 0.946949 0.853764 1.169609 1.071008
DGM (M$) - 2.185373 2.198856 2.118574 2.132669
DGO (M$) - 9.053689 9.109545 8.776950 8.835345
DGIC (M$) - 1.4264 1.4352 1.3828 1.3920
OCTDG (M$) 16.837972 13.612412 13.597365 13.447933 13.431022
OCI 1.0 0.8084 0.8075 0.7987 0.7977

Similarly, at 0.85 pf (WT-DG), the APL is reduced to 14.65 KW (93.03%) for TDO
and 14.66 KW (93%) for GTO. The VD obtained by employing GTO and TDO is
0.000287 p.u and 0.000291 p.u respectively while the value of VSI computed for both

49
TDO and GTO is 0.9753 p.u and the operating cost index for TDO and GTO is 7987
and 7977 respectively. Moreover, the cost savings after 20 years at 0.85 pf (WT-DG)
for TDO and GTO are 20.13% and 20.23% respectively.
It can be concluded that the results obtained at 0.85 pf (WT-DG) in terms of APL
reduction, minimization of VD, minimization of OCI and maximization of VSI are
improved than the results obtained at unity pf (PV-DG). Furthermore, the saving of
cost is maximum at 0.85 pf (PV-DG).
5.2.2.2. VP load Model
The results of MOF in terms of APL reduction, VD minimization, VSI maximization
and minimization of operating cost index for optimum placement and sizing of DGs
for VP load models for three PV (unity pf) and three WT (0.85 pf) are exhibited in
Table 5.11 and Table 5.12.
Table 5.11: Results of OADG for 33 bus RDN for MOF for PV-DG in VP load
model
Parameters Industrial Residential Commercial
APL (KW) 163.22 158.76 152.32
VD (p.u) 0.098637 0.097821 0.094018
Base case
VSI (p.u) 0.70741 0.70953 0.71526
OC (M$) 16.484648 15.859454 15.371032
Parameters TDO [P] GTO [P] TDO [P] GTO [P] TDO [P] GTO [P]
APL (KW) 35.09 35.09 42.69 42.70 45.50 45.50
APLR (%) 78.50% 78.50% 73.11% 73.10% 70.13% 70.13%
Location 14, 30, 24 24, 30, 14 14, 24, 30 14, 24, 30 30, 14, 24 24, 13, 30
DG Size 893, 1255, 1257, 829, 1224, 827, 1231, 1199, 1197, 1199, 843,
(KW) 1261 1256, 894 1230 1232 788 1161
DG Size - - -
(KVAR)
With DG VD (p.u) 0.00374 0.00373 0.00472 0.00472 0.00498 0.00485
VSI (p.u) 0.9385 0.9386 0.9331 0.9333 0.9319 0.9302
OCDG (M$) 1.310889 1.319478 1.277943 1.24798 1.253954 1.172408
DGM (M$) 2.089158 2.087932 2.011941 2.016231 1.951270 1.962914
DGO (M$) 8.655083 8.650005 8.335183 8.352955 8.083833 8.132071
DGIC (M$) 1.3636 1.3628 1.3132 1.316 1.2736 1.2812
OCTDG (M$) 13.418730 13.420216 12.938267 12.933171 12.562656 12.548593
OCI 0.8141 0.8141 0.8158 0.8155 0.8173 0.8164

It can be seen from the Table 5.11 that at unity pf (PV-DG), by employing TDO, the
APL is reduced to 35.09 KW (78.50%) for industrial load model, 42.69 KW (73.11%)
for residential and 45.50 KW (70.11%) for commercial while by employing GTO, the
reduction of APL for industrial is 35.09 KW (78.50%), for residential is 42.70 KW
(73.10%) and for commercial is 45.50 KW (70.13%). The obtained VD from TDO

50
algorithm for industrial load model is 0.00374 p.u, for residential is 0.00472 p.u and
for commercial is 0.00498 p.u while VD obtained from GTO is 0.00373 p.u for
industrial, 0.00472 p.u for residential and 0.00485 p.u for commercial load models.
The VSI computed for industrial load model is 0.9385 p.u, for residential is 0.9331 p.u
and for commercial is 0.9319 p.u by employing TDO while by employing GTO, VD
for industrial load model is 0.9386 p.u, for residential is 0.9333 p.u and for commercial
is 0.9302 p.u.
The operating cost index obtained from TDO and GTO for industrial load is 0.8141,
for residential 0.8158 and 0.8155 and for commercial 0.8173 and 0.8164 respectively.
Besides, the cost saving after 20 years for industrial load will be 18.59% for both TDO
and GTO, for residential load 18.42% and 18.45% for TDO and GTO, and for
commercial 18.27% and 18.36% for TDO and GTO
Table 5.12: Results of OADG for 33 bus RDN for MOF for WT-DG in VP load
model
Parameters Industrial Residential Commercial
APL (KW) 163.22 158.76 152.32
VD (p.u) 0.098637 0.097821 0.094018
Base case
VSI (p.u) 0.70741 0.70953 0.71526
OC (M$) 16.484648 15.859454 15.371032
Parameters TDO [P] GTO [P] TDO [P] GTO [P] TDO [P] GTO [P]
APL (KW) 10.69 10.69 10.50 10.51 10.44 10.45
APLR (%) 93.45% 93.45% 93.39 % 93.38% 93.15% 93.14%
Location 30, 24, 13 30, 24, 13 30, 24, 13 13, 24, 30 13, 30, 24 30, 24, 13
DG Size 1151,1255 1160,1225 1160,1200 831, 1204, 769, 1149, 1151, 1170,
(KW) 901 894 824 1157 1166 771
DG Size 706, 484, 708, 489, 719, 531, 331, 532, 365, 712, 713, 544,
(KVAR) 265 267 337 717 539 360
With DG VD (p.u) 0.000291 0.000288 0.000294 0.000294 0.000294 0.000293
VSI (p.u) 0.9755 0.9756 0.9755 0.9755 0.9755 0.9756
OCDG (M$) 1.643765 1.622348 1.564572 1.530298 1.532499 1.498209
DGM (M$) 2.026649 2.029713 1.951270 1.956173 1.889986 1.894889
DGO (M$) 8.396116 8.408811 8.083833 8.104144 7.829943 7.850254
DGIC (M$) 1.3228 1.3248 1.2736 1.2768 1.2336 1.2368
OCTDG (M$) 13.389330 13.385671 12.873275 12.867414 12.486029 12.480152
OCI 0.8122 0.8120 0.8117 0.8113 0.8123 0.8119

Moreover, the results of OADG for MOF at 0.85 pf (WT-DG) are shown in Table
5.12. By employing TDO, the APL is reduced to 10.69 KW (93.45%) for industrial
load model, 10.50 KW (93.39%) for residential and 10.44 KW (93.15%) for
commercial while by employing GTO, the APL is reduced to 10.69 KW (93.45%),
10.51 KW (93.38%) and 10.45 KW (93.14%) for industrial, residential and
51
commercial load models respectively. The obtained VD from TDO algorithm for
industrial, residential and commercial is 0.000291 p.u, 0.000294 p.u and 0.000294 p.u
respectively while VD obtained from GTO is 0.000288 p.u for industrial, 0.000294
p.u for residential and for commercial is 0.000293 p.u. The VSI computed by
employing TDO for industrial load model is 0.9755 p.u, for residential is 0.9755 p.u
and for commercial is 0.9755 p.u while by employing GTO, VSI computed for
industrial load model is 0.9756 p.u, for residential is 0.9755 p.u and for commercial is
0.9756 p.u. The operating cost index obtained from TDO and GTO for industrial load
is 0.8122 and 0.8120, for residential is 0.8117 and 0.8113 and for commercial is 0.8123
and 0.8119 respectively. Besides, the cost saving after 20 years for industrial load will
be 18.78% and 18.80% for TDO and GTO, for residential load 18.83% and 18.87%
for TDO and GTO and for commercial will be 18.77% and 18.81% for TDO and GTO.
It is concluded that as compared to unity pf (PV-DG), the results of MOF in terms of
APL reduction, minimization of VD and OCI, and maximization of VSI are improved
at 0.85 pf (WT-DG). Furthermore, maximum saving of cost is achieved at 0.85 pf
(WT-DG) for industrial, residential and commercial load models.
5.3. IEEE 69 bus test system
In this section, the conclusions of standard IEEE 69 bus RDS are achieved by the
suggested techniques. Figure 5.5 displays the 69 bus DS single line diagram. It has 69
buses and 68 branches, and the information about the line and load data of the system
is obtained from [9]. The base voltage and base MVA of the system are 12.66 KV and
100 MVA, respectively, while the total load demand is (3.8 + j2.69) MVA.
5.3.1. Scenario 1: Evaluation of Single Objective Function (SOF)
In this scenario, TDO and GTO are used to solve the problem of optimal sitting and
sizing of DGs to reduce the total active and reactive power losses as SOF for CP load
model and VP load model for three PV (unity pf) and three WT (0.85 pf).

52
 36 37 38 39 40 41 42 43 44 45 46

47 48 49 50 51 52 66 67 68 69

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27

28 29 30 31 32 33 34 35 53 54 55 56 57 58 59 60 61 62 63 64 65

Figure 5.5: Single-line diagram of 69 bus RDS

5.3.1.1. CP load Model

The Forward-backward load flow method is used to determine the power flow
solution. The total power loss of the system in the base case is (225.60KW +
j101.99KVAR) and the minimum voltage is 0.9102 p.u at bus 65. The VD and VSI in
the base case are 0.09803 p.u and 0.6855 p.u respectively. The results of the proposed
algorithms for OADG at unity pf (PV-DG ) in CP load model are presented in Table
5.13.
Table 5.13: Results comparison of OADG for 33 bus RDN for APL minimization for
PV-DG in CP load model

APL (KW)
Methods Location DG size (KW) VD (p.u) VSI (p.u)
APLR
Base case - - 224.60 0.09803 0.68548
68.97
BAT [6] 12,19,61 535, 340, 1693 - 0.9113
(69.34%)
SOS-NNA 69.4284
11,18,61 526.8, 380.3, 1719 0.005201 1.0887
[7] (69.14%)
527.2, 382.5, 69.41
IHHO [5] 11,17,61 - -
1719.4 (69.15%)
526.8, 1718.97, 69.3972
QOFBI [3] 11,61,18 - -
380.06 (69.15%)
2148.7, 471.7, 78.347
I-DBEA [8] 61,19,11 0.0002 0.9772
712.6 (65.17%)
68.68
TDO [P] 18,61,11 431, 1931, 525 0.003613 0.9356
(69.42%)
68.68
GTO [P] 11,18,61 525, 431, 1931 0.003613 0.9356
(69.42%)

53
It can be seen from Table 5.13 that the total APL is reduced to 68.68 KW for both
TDO and GTO, that is 69.42% reduction with respect to base value. The obtained
results are compared which other optimization approaches BAT [53], SOS-NNA [57],
IHHO [29], I-DBEA [27] and QOFBI [20] presented in the table. The purposed
algorithms perform well in terms of APL reduction.
Likewise, Table 5.14 shows the results of OADG at unity pf (PV-DG) for QPL
minimization in CP load model. It can be seen from the table that with the integration
of the DGs, QPL is reduced to 31.63 KVAR for both TDO and GTO algorithms, that
is 68.99% reduction with respect to the base value.
Table 5.14: Results of OADG for 69 bus RDN for QPL minimization for PV-DG in
CP load model
DG Size QPL (KVAR)
Methods Location VD (p.u) VSI (p.u)
(KVAR) QPLR
Base case - - 101.99 0.09803 0.68548
800, 617, 31.63
TDO [P] 50, 17, 61 0.003804 0.9371
2000 (68.99%)
617, 2000, 31.63
GTO [P] 17, 61, 50 0.003804 0.9371
800 (68.99%)

Table 5.15: Results comparison of OADG for 69 bus RDN for APL minimization for
WT-DG in CP load model
DG size APL (KW) VD VSI
Methods Location
KW KVAR APLR (p.u) (p.u)
Base case - - - 224.60 0.09803 0.68548
1500,370, 7.966
I-DBEA [8] 61,59,16 - 0.000266 0.9774
575 (96.4%5)
428,1700, 265,1054 7.03
TDO [P] 18,61,11 0.000718 0.9587
687 ,426 (96.87%)
1700,429, 1054,266 7.03
GTO [P] 61,18,11 0.000711 0.9588
689 ,427 (96.87%)

Table 5.16: Results of OADG for 69 BUS RDN for QPL minimization for WT-DG
in CP load model

DG Size QPL (KVAR)

Methods Location VD (p.u) VSI (p.u)
(KW) (KVAR) QPLR
Base case - - - 101.99 0.09803 0.68548
830,1076, 503,6671 3.96
TDO [P] 50,12,61 0.00176 0.9577
700 054 (96.12%)
830,10761 503,667, 3.96
GTO [P] 50,12,61 0.00176 0.9577
700 1054 (96.12%)

Moreover, the results of OADG for APL minimization at 0.85 pf (WT-DG) are
presented in table 5.15. By employing TDO and GTO, the APL is reduced to 7.03 KW,

54
that is 96.87% reduction which is better than other optimization approach I-DBEA
[30] mentioned in Table 5.15. Similarly, the QPL is reduced to 3.96 KVAR for both
GTO and TDO algorithms after the integration of DGs, which is 96.12% reduction
with respect to the base value, that can be seen from Table 5.16.
5.3.1.2. VP load Model
This section extends the OADG issue for practical non-linear load models to show
how strongly power demands rely on voltage of the network. In the base case, the APL
for industrial load model is 171.72KW, for residential is 165.31KW and for
commercial is 157.39KW. Table 5.17 shows the results of SOF based on APL
minimization for optimum placement and sizing of DGs at unity pf (PV-DG).
Table 5.17: Results comparison of OADG for 69 bus RDN for PV-DG in VP load
model

Methods Parameters Industrial Residential Commercial

APL (KW) 171.7224 165.3079 157.3927
Base case VD (p.u) 0.079855 0.076462 0.072316
VSI (p.u) 0.71505 0.7218 0.72986
Location 12, 19, 61 12, 19, 61 12, 19, 61
DG size (KW) 460, 320, 1670 450, 310, 1570 440, 300, 1480
BAT [53] APL (KW) 28.81 37.73 41.33
APLR (%) 83.19 77.11 73.66
VD (p.u) - - -
VSI (p.u) 0.9431 0.9357 0.9314
Location 61, 11, 18 18, 61, 11 11, 61, 18
DG size (KW) 1854, 543, 422 421, 1761, 424 517, 1656, 410
TDO [P] APL (KW) 28.71 37.54 41.07
APLR (%) 83.28 77.29 73.91
VD (p.u) 0.002009 0.002428 0.002575
VSI (p.u) 0.9523 0.9477 0.9463
Location 18, 11, 61 18, 11, 61 18, 11, 61
DG size (KW) 424, 538, 1855 412, 523, 1748 402, 512, 1663
APL (KW) 28.71 37.54 41.07
GTO [P] APLR (%) 83.28 77.29 73.91
VD (p.u) 0.00201 0.002429 0.002575
VSI (p.u) 0.9523 0.9477 0.9463

It can be seencthat the APL is reduced to 28.71 KW (83.28%), 37.61 KW (77.25%)

and 41.07 KW (73.91%) for industrial, residential, and commercial load models
respectively for both TDO and GTO. The obtained results are compared with other
optimization method BAT [53] mentioned in Table 5.17. The purposes algorithms
show better results in term of APL reduction. Likewise, the QPL in the base case for
industrial load model is 79.23 KVAR, for residential is 76.52 KVAR and for

55
commercial is 73.13 KVAR. The results of SOF based on QPL minimization for
optimum placement and sizing of multiple DGs at unity pf (PV-DG) are presented in
Table 5.18. After the integration of DGs, the QPL is reduced to 14.45 KVAR (81.76%)
for industrial load model, 18.23 KVAR (76.17%) for residential and for commercial
QPL is reduced to 19.73 KVAR (73.02%) for both TDO and GTO algorithms.
Table 5.18: Results of OADG for 69 bus RDN for QPL minimization for PV-DG in
VP load model

Parameters Industrial Residential Commercial

QPL (KVAR) 79.23 76.52 73.13
Base case VD (p.u) 0.079855 0.076462 0.072316
VSI (p.u) 0.71505 0.7218 0.72986
QPL (KVAR) 14.45 18.23 19.73
QPLR (%) 81.76 76.17 73.02
Location 61, 50, 17 50, 17, 61 17, 50, 61
DG Size (KW) 1943, 798, 611 796, 594, 1833 581, 793,1748
TDO [P]
DG Size (KVAR) - - -
VD (p.u) 0.002071 0.002477 0.002617
VSI (p.u) 0.9559 0.9512 0.9498
QPL (KVAR) 14.45 18.23 19.73
QPLR (%) 81.76 76.17 73.02
Location 61, 50, 17 17, 50, 61 50, 61, 17
GTO [P] DG Size (KW) 1942, 799, 612 594, 796, 1833 793, 1748, 581
DG Size (KVAR) - - -
VD (p.u) 0.002068 0.002477 0.002618
VSI (p.u) 0.9558 0.9512 0.9498

Table 5.19: Results of OADG for 69 bus RDN for APL minimization for WT-DG in
VP load model

Parameters Industrial Residential Commercial

APL (KW) 171.7224 165.3079 157.3927
Base case VD (p.u) 0.079855 0.076462 0.072316
VSI (p.u) 0.71505 0.7218 0.72986
APL (KW) 3.89 3.87 3.83
APLR (%) 97.73 97.66 97.57
Location 11, 17, 61 11, 61, 18 61, 18, 11
TDO [P] DG Size (KW) 546, 418, 1830 544, 1711, 402 1628, 400, 517
DG Size (KVAR) 329, 217, 777 337, 927, 239 988, 248, 318
VD (p.u) 0.000118 0.000118 0.000118
VSI (p.u) 0.9778 0.9777 0.9777
APL (KW) 3.89 3.87 3.82
APLR (%) 97.73 97.66 97.57
Location 11, 61, 18 18, 61, 11 61, 11, 18
GTO [P] DG Size (KW) 546, 1830, 418 403, 1712, 542 1622, 539, 394
DG Size (KVAR) 329, 777, 217 239, 927, 336 983, 334, 244
VD (p.u) 0.000118 0.000118 0.000116
VSI (p.u) 0.9778 0.9777 0.9777

56
Table 5.20: Results of OADG for 69 bus RDN for QPL minimization for WT-DG in
VP load model
Parameters Industrial Residential Commercial
QPL (KVAR) 79.23 76.52 73.13
Base case VD (p.u) 0.079855 0.076462 0.072316
VSI (p.u) 0.71505 0.7218 0.72986
QPL (KVAR) 2.24 2.17 2.12
QPLR (%) 97.17 97.16 97.11
Location 18, 50, 61 17, 50, 61 50, 61, 17
TDO [P] DG Size (KW) 627, 836, 1837 602, 824, 1755 828, 1709, 582
DG Size (KVAR) 347, 517, 790 362, 511, 959 513, 1034, 360
VD (p.u) 0.000353 0.000283 0.000230
VSI (p.u) 0.9808 0.9811 0.9818
QPL (KVAR) 2.24 2.17 2.12
QPLR (%) 97.17 97.16 97.11
Location 18, 50, 61 17, 50, 61 50, 61, 17
GTO [P] DG Size (KW) 627, 836, 1837 602, 824, 1755 828, 1709, 582
DG Size (KVAR) 347, 517, 790 362, 511, 959 513, 1034, 360
VD (p.u) 0.000353 0.000283 0.000230
VSI (p.u) 0.9808 0.9811 0.9818

Moreover, Table 5.19 represents the results of OADG at 0.85 pf (WT-DG) for APL
minimization for different load models. The APL is reduced to 3.89 KW (97.73) for
industrial load model, 3.87 KW (97.66%) for residential and for commercial load it is
reduced to 3.83 KW (97.57%) for both TDO and GTO algorithms.
Similarly, the results of OADG for QPL reduction at 0.85 pf (WT-DG) for different
load models are described in Table 5.20. The QPL for industrial load model is reduced
to 2.24 KVAR (97.17%), for residential reduced to 2.17 KVAR (97.16%) and for
commercial load model QPL reduced to 2.12 KVAR (97.11%) for both TDO and GTO
algorithms.
5.3.1.3. Voltage profiles for 69 bus RDS
Figure 5.6 illustrates the impact of the installation of various types of DGs on the
voltage profile for the 69 bus RDS under various types of loads i.e., constant,
residential, commercial, and industrial. In Figure 5.6, “const” stands for constant,
“Res” denotes the residential, “Comm” denotes the commercial and” Ind” stands for
Industrial loads. The outer values of the graphs represent the “bus number” and the
inner values represent “ bus voltage” corresponding to the particular bus number.

57
 Figure 5.6: Voltage profiles for IEEE 69 bus system for different loads at different
cases for scenario 1
5.3.1.4. Active power losses for 69 bus RDS
Figure 5.7 depicts the active power loss information for each bus of 69 bus RDS after
integration of three PV (unity pf) and three WT (0.85 pf) under various types of loads
i.e., constant, residential, commercial, and industrial under scenario 1.

Figure 5.7: Active power loss for 69 buses under various types of loads

58
5.3.1.5. Convergence characteristics for 33 bus RDS
Figures 5.8(a) and 5.8(b) display the convergence characteristics of GTO and TDO at
various power factors (unity and 0.85pf) for CP load model . It is apparent from the
below figure that GTO has higher efficiency than TDO.

(a)

(b)
Figure 5.8: Convergence characteristics of TDO and GTO for IEEE 69 bus system
for scenario 1 at different pf; a) Unity pf (PV type), (b) 0.85 pf (WT type)
5.3.2. Scenario 2: Evaluation of Multi-Objective Function (MOF)
In this scenario, TDO and GTO algorithms are employed to optimize the MOF based
on APL, VD, VSI and OCI to solve the problem of OADG in RDS considering CP
load model and VP load models for three PV (unity pf) and three WT (0.85 pf).
5.3.2.1. CP load Model
Table 5.21 presents the results of MOF in terms of APL reduction, VD minimization,
VSI maximization and minimization of total operating costs for optimum placement
and sizing of DGs for three PV (unity pf) and three WT (0.85 pf). The total power loss

59
of the system in the base case is (224.60KW + j101.99KVAR), the voltage deviation
is 0.09803 p.u, VSI is 0.6855 and operating cost is 17.274315 (M$).
It can be seen from Table 5.21 that at unity pf (PV-DG), the APL is reduced to 69.19
KW (69.19%) for TDO and 68.20 KW (69.18%) for GTO. The VD obtained by
employing TDO and GTO is 0.002093 p.u and 0.002074 p.u respectively while the
value of VSI computed for TDO and GTO is 0.9477 p.u and 0.9478 p.u respectively
and the operating cost index for TDO and GTO is 8280 and 8278 respectively.
Furthermore, the cost savings for GTO and TDO at unity pf (PV-type) after 20 years
are 17.20% and 17.22% respectively.
Table 5.21: Results of OADG for 69 bus RDN for MOF in CP load model

Unity pf (PV type) 0.85 pf (WT type)

Parameters Base case
TDO [P] GTO [P] TDO [P] GTO [P]
APL (KW) 224.60 69.19 69.20 7.21 7.21
APLR (%) - 69.19% 69.18% 96.79% 96.79%
Location - 61, 11, 18 11, 61, 17 17, 64, 61 17, 61, 64
DG Size - 2000, 682, 683, 2000, 613, 350, 617, 1691,
(KW) 441 444 1678 338
DG Size - - - 379, 217, 382, 1048,
(KVAR) 1040 209
VD (p.u) 0.09803 0.002093 0.002074 0.000339 0.000332
VSI (p.u) 0.68548 0.9477 0.9478 0.9773 0.9773
OCDG (M$) 17.274315 3.210458 3.193349 5.012274 4.990830
DGM (M$) - 1.913887 1.916338 1.618500 1.621564
DGO (M$) - 7.928960 7.939116 6.705214 6.717909
DGIC (M$) - 1.2492 1.2508 1.0564 1.0584
OCTDG (M$) 17.274315 14.302505 14.299603 14.392388 14.388703
OCI 1.0 0.8280 0.8278 0.8332 0.8330

Similarly, at 0.85 pf (WT-DG), the APL is reduced to 7.21 KW (96.76%) for both
TDO and GTO. The VD obtained by employing GTO and TDO is 0.000339 p.u and
0.000332 p.u respectively while the value of VSI computed for both TDO and GTO is
0.9773 p.u and the operating cost index for TDO and GTO is 8332 and 8330
respectively. Moreover, the cost savings for TDO and GTO at 0.85 pf (WT-DG) after
20 years are 16.68% and 16.78% respectively. It can be concluded that as compared
to 0.85 pf (WT-DG), the results of the cost saving for unity pf (PV-DG) is high.
5.3.2.2. VP load Model
The results of MOF in terms of APL reduction, VD minimization, VSI maximization
and minimization of total operating cost for optimum placement and sizing of DGs for
VP load models for three PV (unity pf) and three WT (0.85 pf) are shown in Table
5.22 and Table 5.23.

60
 Table 5.22: Results of OADG for 69 bus RDN for MOF for PV-DG in VP load
model
Parameters Industrial Residential Commercial
APL (KW) 171.72 165.31 157.39
VD (p.u) 0.079855 0.076462 0.072316
Base case
VSI (p.u) 0.7151 0.7218 0.7299
OC (M$) 16.902644 16.300692 15.831996
Parameters TDO [P] GTO [P] TDO [P] GTO [P] TDO [P] GTO [P]
APL (KW) 29.44 29.44 38.23 38.21 41.70 41.73
APLR (%) 82.86% 83.86% 76.87% 76.88% 73.50% 73.49%
Location 61, 18, 11 18, 61, 11 18, 61, 11 18, 11, 61 18, 61, 11 18, 61, 11
DG Size 1996, 429, 431, 1996, 419, 1887, 418, 601, 411, 581, 409, 1798,
(KW) 616 614 593 1882 1796 590
DG Size
- - - -
(KVAR)
With DG VD (p.u) 0.000882 0.000876 0.00117 0.00117 0.00129 0.00127
VSI (p.u) 0.9714 0.9714 0.9665 0.9660 0.9643 0.9647
OCDG (M$) 3.246820 3.246814 3.319258 3.310582 3.375599 3.337137
DGM (M$) 1.863634 1.863634 1.776612 1.777837 1.708587 1.714102
DGO (M$) 7.720771 7.720771 7.360248 7.365326 7.078431 7.101281
DGIC (M$) 1.2164 1.2164 1.1596 1.1604 1.1152 1.1188
OCTDG (M$) 14.047625 14.047619 13.615718 13.614146 13.277817 13.271321
OCI 0.8311 0.8311 0.8353 0.8352 0.8387 0.8383

It can be seen from the Table 5.22 that at unity pf (PV-DG), by employing TDO, the
APL is reduced to 29.44 KW (82.86%) for industrial load model, for residential APL
reduced to 38.23 KW (76.87%) and for commercial APL is reduced to 41.70 KW
(73.50%) while by employing GTO, the APL is reduced to 29.44 KW (82.86%) for
industrial, 38.21 KW (76.88%) for residential and for commercial APL is reduced to
41.73 KW (73.49%). The obtained VD from TDO algorithm for industrial load model
is 0.000882 p.u, for residential is 0.00117 p.u and for commercial is 0.00129 p.u while
VD obtained from GTO is 0.000876 p.u for industrial load model, 0.00117 p.u for
residential and for commercial load model is 0.00127 p.u. The VSI computed by
employing TDO for industrial load model is 0.9714 p.u , for residential is 0.9665 p.u
and for commercial load models is 0.9643 p.u while for GTO is 0.9714 p.u, 0.9660
p.u and 0.9647 p.u for industrial, residential and commercial load models respectively.
The operating cost index obtained from TDO and GTO for industrial load is 0.8311,
for residential 0.8353 and 0.8352, and for commercial 0.8387 and 0.8383 respectively.
Besides, the cost saving after 20 years for industrial load will be 16.89% for both TDO

61
and GTO, for residential load 16.47% and 16.48% for TDO and GTO, and for
commercial 16.13% and 16.17% for TDO and GTO respectively.
Table 5.23: Results of OADG for 69 bus RDN for MOF for WT-DG in VP load
model
Parameters Industrial Residential Commercial
APL (KW) 171.72 165.31 157.39
VD (p.u) 0.079855 0.076462 0.072316
Base case
VSI (p.u) 0.7151 0.7218 0.7299
OC (M$) 16.902644 16.300692 15.831996
Parameters TDO [P] GTO [P] TDO [P] GTO [P] TDO [P] GTO [P]
APL (KW) 3.95 3.96 3.93 3.93 3.88 3.88
APLR (%) 97.70% 97.69% 97.62% 97.62% 97.54% 96.54%
Location 18, 11, 61 11, 61, 18 18, 61, 11 61, 18, 11 61, 11, 18 61, 18, 11
DG Size 422, 612, 622, 1843, 406, 1732, 1727, 409, 1640, 595, 1641, 399,
(KW) 1844 418 600 602 395 590
DG Size 220, 327, 323, 778, 241, 927, 928, 236, 980, 354, 979, 242,
(KVAR) 775 220 335 352 241 365
With DG VD (p.u) 0.0000890 0.0000890 0.0000888 0.0000888 0.0000887 0.0000882
VSI (p.u) 0.9778 0.9778 0.9777 0.9777 0.9777 0.9777
OCDG (M$) 3.8367397 3.8153204 3.862792 3.862786 3.891135 3.891138
DGM (M$) 1.763742 1.766806 1.677945 1.677945 1.611759 1.611759
DGO (M$) 7.306932 7.319626 6.951487 6.951487 6.677286 6.677286
DGIC (M$) 1.1512 1.1532 1.0952 1.0952 1.052 1.052
OCTDG (M$) 14.058613 14.054953 13.587424 13.587417 13.232180 13.232183
OCI 0.8317 0.8315 0.8336 0.8336 0.8358 0.8358

Moreover, the results of OADG for MOF at 0.85 pf (WT-DG) are shown in Table
5.23. By employing TDO, the APL is reduced to 3.95 KW (97.70%) for industrial load
model, 3.93 KW (97.62%) for residential and for commercial load model is 3.88 KW
(97.54%) while by employing GTO, the APL is reduced to 3.96 KW (97.69%), 3.93
KW (97.62%) and 3.88 KW (96.54%) for industrial, residential and commercial load
models respectively. The obtained VD from TDO and GTO algorithm for industrial
load model is 0.0000890 p.u and for residential load model is 0000888 p.u while VD
obtained from TDO and GTO for commercial load model is 0.0000887 p.u and
0.0000882 p.u respectively. The VSI computed for industrial, residential and
commercial load models by employing both TDO and GTO is 0.9778 p.u, 0.9777 p.u
and 0.9777 p.u respectively. The operating cost index obtained from TDO and GTO
after 20 years for industrial load is 0.8317 and 0.8315, for residential 0.8336, and for
commercial is 0.8358 respectively. Besides, the cost saving after 20 years for industrial

62
load will be 16.83% and 16.85% for TDO and GTO, for residential load 16.65% for
both TDO and GTO, and for commercial 16.42% for both TDO and GTO respectively.
It is concluded that as compared to 0.85 pf (WT-DG), the results of MOF in terms of
APL reduction, minimization of VD and OCI, and maximization of VSI are improved
at 0.85 pf (WT-DG). Furthermore, maximum saving of cost is achieved at 0.85 pf
(WT-DG) for residential and commercial load models while for industrial load model,
maximum cost saving occurs when DG operates at unity pf (PV-DG).

63
Summary
In this chapter, problem of OADG is addressed by using the TDO and GTO algorithms
to optimize the single objective function and multi objective function. In SOF, active
and reactive power loss is optimized while to minimize the active power loss, voltage
deviation, operating cost index along with the maximizing the voltage stability index,
MOF is used. For multi objectivity, the weighted sum method is used by assigning the
proper weight to each of the objectives. The proposed techniques are examined on the
IEEE 33 bus and IEEE 69 bus RDN for CP and VP load models. The obtained results
are compared with the other existing optimization techniques to validate the results

64
References
[1] F. F. W. Mesut E. Baran, “Network reconfiguration in distribution systems for
loss reduction and load balancing,” IEEE Trans. Power Deliv., vol. 4, no. 2, pp.
1401–1407, 1989, doi: 10.1109/61.25627
[2] O. D. Montoya, W. Gil-González, and L. F. Grisales-Noreña, “An exact MINLP
model for optimal location and sizing of DGs in distribution networks: A
general algebraic modeling system approach,” Ain Shams Eng. J., vol. 11, no.
2, pp. 409–418, 2020, doi: 10.1016/j.asej.2019.08.011.
[3] B. K. Malika, V. Pattanaik, B. K. Sahu, and P. K. Rout, Quasi-oppositional
Forensic-Based Investigation for Optimal DG Selection for Power Loss
Minimization, no. 0123456789. Springer Nature Singapore, 2022. doi:
10.1007/s41660-022-00277-9.
[4] V. Tiwari, H. M. Dubey, and M. Pandit, “Assessment of Optimal Size and
Location of DG/CB in Distribution Systems using Coulomb–Franklin’s
Algorithm,” J. Inst. Eng. Ser. B, vol. 103, no. 6, pp. 1885–1908, 2022, doi:
10.1007/s40031-022-00811-w.
[5] A. Selim, S. Kamel, A. S. Alghamdi, and F. Jurado, “Optimal Placement of
DGs in Distribution System Using an Improved Harris Hawks Optimizer Based
on Single- And Multi-Objective Approaches,” IEEE Access, vol. 8, pp. 52815–
52829, 2020, doi: 10.1109/ACCESS.2020.2980245.
[6] T. Yuvaraj et al., “Optimal integration of capacitor and distributed generation
in distribution system considering load variation using bat optimization
algorithm,” Energies, vol. 14, no. 12, 2021, doi: 10.3390/en14123548.
[7] T. P. Nguyen, T. A. Nguyen, T. V. H. Phan, and D. N. Vo, “A comprehensive
analysis for multi-objective distributed generations and capacitor banks
placement in radial distribution networks using hybrid neural network
algorithm,” Knowledge-Based Syst., vol. 231, p. 107387, 2021, doi:
10.1016/j.knosys.2021.107387.
[8] A. Ali, M. U. Keerio, and J. A. Laghari, “Optimal Site and Size of Distributed
Generation Allocation in Radial Distribution Network Using Multi-objective
Optimization,” J. Mod. Power Syst. Clean Energy, vol. 9, no. 2, pp. 404–415,
2021, doi: 10.35833/MPCE.2019.000055.
[9] F. F. W. Mesut E. Baran, “Optimal sizing of capacitors placed on a radial

65
distribution system,” IEEE Trans. Power Deliv., vol. 4, no. 1, pp. 735–743,
1989, doi: 10.1109/61.19266

66
 Chapter 6: Conclusion and Future
Work
6.1. Conclusion
In this study, two metaheuristic optimization algorithms named as TDO, and GTO are
employed for optimal sitting and sizing of renewable based DGS (PV,WT) at unity
and 0.85 pf. The proposed algorithms are utilized to optimize the single and multi-
objective functions. The reduction of active and reactive power loss is achieved
through optimizing single objective function while multi-objective function is
optimized to reduce the active power loss, minimize the voltage deviation, maximize
the voltage stability index and minimize the total operating cost index. These
objectives are achieved by using a multi-objectivity approach i.e., weighted sum
method. The effectiveness and the validation of the proposed algorithms are examined
on IEEE 33 bus and IEEE 69 bus RDN. As SOF, the active power loss for both TDO
and GTO algorithms at 0.85 pf is reduced more than other existing optimization
algorithms, that is, 93.15% and 96.87% of initial value for 33 bus and for 69 bus RDN
respectively. The purposed algorithms are also examined for PV-DG and WT-DG on
IEEE 33 and IEEE 69 bus RDN considering voltage dependent load models or variable
power load model like industrial, residential and commercial load models and they
also show the better results as compared to the existing optimization techniques in
terms of power loss reduction. As SOF, the purposed GTO algorithm outperform the
other purposed TDO algorithm in terms of efficiency to converge the solution while
for MOF, the results of APL minimization for GTO algorithm are comparable to the
other purposed TDO algorithm but in terms of other objectives like VD minimization,
OCI minimization and VSI maximization, GTO has better results than TDO.
Therefore, both GTO and TDO can be used for optimal placement and sizing of DGs
in DN.
6.2. Future Work
In future, following work can be extended in further directions such as:
• The hybrid version of the proposed techniques can be employed for the
optimization of the other higher distribution network systems.

67
• The effect of the DG penetration on the environment in terms of reduction of GHG
emission can be analyzed.
• The uncertainty associated with the renewable energy resources like PV and wind
could be considered along with seasonal variations.

68
 Appendix
Publication

69