2016 International Conference on Advanced Communication Control and Computing Technologies (ICACCCT)

SOS algorithm for DG placement for loss

minimization considering reverse power flow in
the distribution systems
M. Padma Lalitha P.Suresh babu B.Adivesh
Professor Assistant professor P.G.Scholar,
Department of EEE, Department of EEE, Department of EEE,
AITS Rajampet,kadapa, AITS Rajampet,kadapa, AITS Rajampet,kadapa
A.P.India A.P.India A.P.India

Abstract—Distribution systems play a vital role in Type 1: DG’s that produce only active power (e.g.
delivering power to the consumers. But, with the PV, micro turbines, fuel cells)
demand for increased power, it is considered as the Type 2: DG’s that produce only reactive power (e.g.
weakest section in the entire power system as most of capacitors, synchronous compensators)
the power loss occurs in this section only. The main goal
Type 3: DG’s that produce both active and reactive
of the utility is to deliver the power of acceptable
quality. Placement of Distributed Generators (DG) powers (e.g. synchronous generators)
finds a viable solution for reducing the power loss and Type 4: DG’s that produce active power by
voltage profile improvement and hence improved consuming reactive power (e.g. induction generators)
reliability. Non optimal locations and sizes may lead to Different methods have been developed by
increased losses and bad voltage profiles. Also the researchers to determine the optimum location and
optimal placement of DG should not involve in reverse size of the DG’s. N.Acharya et al., used analytical
flow into the slack bus. In this paper a symbiotic approach for the allocation of DG’s [5]. M.P.Lalitha
relation based algorithm called Symbiotic Organisms et al., used ABC algorithm for loss reduction by DG
Search (SOS) algorithm is used for the optimal
placement which is done in two stages[6]. M.Kefayat
placement and sizing of DG’s simultaneously. The
performance of this proposed method is tested on IEEE et al., used combination of genetic algorithm and
33 and 69 bus test systems and the results are presented simulated annealing for DG placement [7]. Moradi
and compared with existing methods. MH et al., used GA and PSO for DG placement [10].
In this paper both optimal location and optimal sizing
Index Terms – Symbiotic Organisms Search (SOS); of DG’s are determined by using SOS algorithm
optimal size; optimal placement; Distributed Generation simultaneously.
(DG); Power losses reduction.
II. PROBLEM FORMULATION

I. INTRODUCTION The main objective function of this paper is to

minimize the real power loss in the distribution
Distributed generation (DG) also called
system. It can be mathematically formulated as
decentralized generation, dispersed generation
below
generally refers to small-scale (typically 1 kW – 50
MW) electric power generators that produce electric ୲୭୲ୟ୪୪୭ୱୱ ൌ ‹‹‹œ‡σ୬୧ୀଵ ୧ଶ  ୧ 
power at a site near to the load. DG’s share has been Where n=number of branches
increased in the power system from the last few I is the magnitude of the branch current and
i
years. Distributed generation include fuel cells, solar
R is the resistance of the ith branch.
photovoltaic, and wind turbines. Jiayi H.,Chuanwen J i
et al., had given a review of different DG The branch currents are obtained from the load
technologies available [1]. flows. The branch current has two components,
By optimal placement of DG’s, distribution active component (I ) and reactive component (I ).
A r
network performance can be improved in terms of The loss associated with the active and
improved voltage profile, reduction in flows and reactive components of branch currents are given as:
power losses, improved power quality, reliability and ଶ
୐ୟ ൌ σ୬୧ୀଵ ୟ୧  ୧ 
satisfying the increased demand. P.Chiradeja,et al ୬ ଶ
୐୰ ൌ σ୧ୀଵ ୰୧  ୧ 
.,mentioned the benefits of DG placement [2-3].
Inappropriate DG placement may increase system In addition voltage constraints should be
losses and network capital and operating costs. In satisfied, given as
addition, DGs plays a key role in dealing with |Vmin| ≤ |Vi |≤ |Vmax|
environmental concern due to their ability to produce By placing DG’s, the active and reactive
green power and reduce the dependency on fossil currents flowing in the branches can be reduced,
fuels. The following are the different types of DG’s which results in loss minimization.

ISBN No.978-1-4673-9545-8 443

 2016 International Conference on Advanced Communication Control and Computing Technologies (ICACCCT)

This paper presents the method that is used organism of the ecosystem is considered as a
to minimize the active power losses only associated candidate solution to the corresponding problem and
due to active component of the currents, by finding will have certain fitness .The new solutions are
optimal locations and sizes of DG’s simultaneously. generated by simulating the symbiotic interactions
between two organisms in the ecosystem which
III. CONCEPT OF REVERSE POWER FLOW includes the mutualism, commensalism and
Under normal operating conditions of the parasitism phases. Each organism in the ecosystem
distribution system, reverse power flow in the randomly interacts with the other through all these
internal branches occurs and allowed, but should not three phases and this process of interaction is
be allowed to flow back into the grid. Reverse power repeated until the termination criterion is fulfilled.
flow in the system may occur during light loading The details of operation of these three phases of
conditions or if the total power generated by the symbiotic interaction are as below.
DG’s exceeds the total load. Reverse power flow
C. mutualism phase
may disturb the inherent integrity of the protection
and regulating devices already placed in the system This phase of SOS algorithm mimics the
considering one direction of power flow. It also mutualistic interaction between two organisms where
causes increased branch currents and potentially both the organisms are benefited. One example of
increased power losses. Thus, optimal placement and mutualism is the relation between oxpecker and
sizing of DG’s should consider the loading zebra. Oxpeckers eat ticks and parasites from zebra’s
conditions of the system. For this purpose daily load skin. In this way, oxpeckers get food and zebra gets
curves are integrated with optimal sizing of the DG’s pest control. In this phase, Yi is considered as ith
which should not cause reverse power flow in the organism in the ecosystem and another organism Yj is
system. This reverse power can be simply examined selected randomly to interact with Yi . Both the
by the voltage profiles of first (slack) bus and second organisms exhibit a mutualistic relationship to
bus. If the difference is negative it leads to the increase their mutual survival advantage in the
reverse power flow. ecosystem and the new solutions for Yi and Yj are
given as below,
IV. SYMBIOTIC ORGANISMS SEARCH Yinew Yi  rand ( 0,1) * (Ybest  Mutual _ Vector * BF1)
(SOS) ALGORITHM
Y jnew Y j  rand ( 0,1) * (Ybest  Mutual _ Vector * BF 2)
A. Introduction

SOS algorithm is a recently developed In above equations, Mutual _Vector is determined as

population based algorithm developed by Cheng and ୧ ൅ ୨
Prayogo in 2014. This algorithm is inspired from the —–—ƒŽ̴˜‡…–‘” ൌ 
ʹ
symbiotic interactions that exist between two
organisms in the ecosystem. Organisms in the real BF1 and BF2 are benefit factors and their values are
world rarely live in isolation due to dependence on either 1 or 2. These factors represent the level of
other species for their survival. benefit to each organism, as the organisms may get
partially or fully benefited from the interaction.
B. Concept of the algorithm Mutual _Vector represents the relationship between
Organisms in the environment develop Yi and Yj . Ybest represents the highest degree of
symbiotic relationships with others to adapt to adaptation. The new solutions are only accepted if
changes in their environment. Three types of they give better fitness value than the previous
symbiotic relationships are found in nature. They are solutions.
mutualism, commensalism and parasitism. By
performing these three phases, SOS tries to move D. Commensalism phase
population, called an ecosystem of possible The relationship between spider and trees or
solutions, to promising areas of the search space herbs is the example of commensalism. The spider
during the search for the optimal solution. makes net on the trees or herbs to trap insects. In this
Mutualism refers to the relationship way, the spider gets food but the trees or herbs
between two different species of organisms where remain unaffected. In the SOS algorithm, to simulate
both individuals get benefited. Commensalism this commensalism phase, an organism Yj is selected
describes the symbiotic relationship between two randomly from the ecosystem which is made to
organisms in which one benefits but the other is not interact with the organism Yi . Now, the organism Yi
affected by the benefited. Parasitism is a kind of tries to get benefited from the interaction while it
symbiotic relationship where one organism is does not benefit or harm the organism Yj . The new
benefited and the other is, effectively, harmed. candidate solution of Yi, generated by the commensal
Similar to other population based interaction, is given by
algorithms, SOS algorithm also employs a population
of candidate solutions for searching the optimal Yinew Yi  rand (1,1) * (Ybest  Y )
j
global solution. SOS algorithm starts with an initial
population of organisms called the ecosystem. Each

444
 2016 International Conference on Advanced Communication Control and Computing Technologies (ICACCCT)

where (Ybest − Yj ) interprets the benefit provided by parasite enters into human body through anopheles
Yj to help Yi to increase its degree of adaptation so mosquitoes and it reproduces inside the host human
that it can survive in the ecosystem. body. As a result, the human host suffers from
malaria and may also die.
E. Parasitism phase
In parasitism phase of SOS algorithm, an organism
Yi is chosen, which is similar to the anopheles
Initialization maximum iterations
mosquito, and it creates an artificial parasite named
DG sizes, DG locations., Parasite _Vector. This Parasite _Vector is created by
duplicating Yi and then its randomly selected
Iter. i = 1 dimensions are modified using a random number.
Now, an organism Yj is selected randomly from the
ecosystem which is treated as a host to the parasite. If
Select first DG size DGi, where DGi ≠ DGj
the fitness value of Parasite _Vector is better than
that of Yj , then it will kill the organism Yj and take
Modify organisms DGi and DGj based on their mutual relationship over its position in the ecosystem. On the other hand,
M
U if the fitness value of Yj is better, then it builds
T immunity against the Parasite _Vector and the
U
A parasite will no longer exist in the ecosystem.
L
I
The flow chart for SOS algorithm is given below:
Calculate the fitness of the modified DGs
S The flow chart is explained for DG sizes
M
only and the similar follows for DG locations
NO Are the modified DGs fitter than previous? YES
simultaneously.
keep the previous DGs Accept the new DGs
V. PARAMETERS TAKEN FOR THE
i = i+1 PROPOSED METHOD
C
O
Randomly select one DG DGj , where DGj ≠ DGi DG sizes =50
M
M DG locations =50
E Beneficial factors =1 or 2
N Modify DG DGi with the help of DGj and calculate
S the fitness value of modified DG
A
L
VI. TEST SYSTEMS AND RESULTS AND
I DISCUSSIONS
S
M NO Is the modified DG fitter than previous? YES Programmes are written in MATLAB for
finding the total losses in both 33 and 69 bus
keep the previous DG Accept the new DG systems. And then SOS algorithm code is written for
both optimal placement and sizing of DG’s
simultaneously.
P Randomly select one DG DGj , where DGj ≠ DGi A. IEEE 33 bus system data and results
A
R
A
S
Create parasite vector using DG DGi
I 23 24 25
T
I
S
NO
M NO Is the parasite vector fitter than DGj? YES 26 27 28 29 30 31 32 33

Keep DGj and remove parasite

Replace DGj with parasite vector
vector

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

Increment DG by one

19 20 21 22
Is termination criteria full filled?
Fig. 2. Single line diagram of the IEEE 33-bus
YES distribution system.

Optimal DG size This 33 bus test system has the following data:
Buses : 33
Fig. 1. Flow chart of the SOS algorithm Branches : 32
Base voltage : 12.66 KV
The relationship between plasmodium parasite and
Base MVA : 100
the human being is an example of this phase. The
Total Load connected : 3.715 MW and 2.3 MVar

445
 2016 International Conference on Advanced Communication Control and Computing Technologies (ICACCCT)

TABLE –I. Load flow results System data:

Buses : 69
Loss due to Loss due to Branches : 68
active reactive Total Base voltage : 12.66 KV
Bus
component of component of the power loss
system
the current current PLt (KW)
Base MVA : 100
PLa (KW) PLr (KW) Total Load connected : 3.80 MW and 2.694 MVar

TABLE-II. Load flow results

33 135.942 66.724 202.666
Bus Loss due to Loss due to Total power
system active reactive loss
component of component of PLt (KW)
1. DG placement results: Results of type1 DG the current the current
PLa (KW) PLr (KW)
DG’s that produce only active power (e.g. 69 152.826 71.715 224.541
PV, micro turbines, fuel cells)
By placing these types of DG’s, the active 1. DG placement results: Results of type1 DG
currents flowing in the line are compensated, as a
DG’s that produce only active power
result the loss associated with active component of
By placing these types of DG’s, active
branch currents is reduced drastically. As number of
component losses are reduced drastically reactive
DG units placed increases, the % reduction in loss
component losses are almost remains same and the
increases. Type I DG placement yields better results
theory as explained before. The results of DG
for the systems where the loss associated with active
placement for type 1 DG’s for 69 bus system are as
component of branch current (PLa) is more compared
shown in table-IV
to the loss associated with reactive component of the
The results show that the active component
branch current (PLr). The modified active power at
power loss is reduced from 152.82KW to 2.41 KW,
buses by placing DG is given as and the total loss reduction can be obtained as high
as 69.33%. Also this placement does not involve the
୬ୣ୵ ሺ‹ሻ ൌ ୭୪ୢ ሺ‹ሻ െ ୈୋ ሺ‹ሻ (4)
reverse power flow.
The results of DG placement for type-1
VII. COMPARISION WITH EXISTING
DG’s for 33 bus system are as shown in table-III.
METHODS
The results show that the active component
power loss is reduced from 135.9421 KW to 66.0281
M. Kefayat et.al.,[11] used hybrid of Ant Colony
KW due to reduction in active currents flowing by
Optimization (ACO) and Artificial Bee colony
placing the DG’s, but the reduction in reactive
(ABC) for DG placement. The results of this paper
component loss very less. The total loss reduction
are compared with the proposed method. The
can be obtained as high as 67.42%. Also this
comparison with 33 bus system and 69 bus system
placement does not involve the reverse power flow.
DG placement results are shown in TABLE-V and
B. IEEE 69 bus system data and results TABLE-VI.
The comparison shows the effectiveness of the
proposed method for DG placement. The authors in
36 37 38 39 40 41 42 43 44 45 46
[11] used hybrid of Ant Colony Optimization (ACO)
and Artificial Bee colony (ABC) for DG placement.
But the proposed method used only single algorithm
for the DG placement and the results show the
47 48 49 50
effectiveness of the algorithm.
51 52 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

66 67

53 54 55 56 57 58 59 60 61 62 63 64 65

28 29 30 31 32 33 34 35

Fig. 3. Single line diagram of the IEEE 33-bus

distribution system.

446
 2016 International Conference on Advanced Communication Control and Computing Technologies (ICACCCT)

TABLE –III. DG placement results of 33 bus test system

Loss with DG (kw)

Number Optimal Total Active Reactive Total Loss without
Optimal % reduction
of DG’s Sizes Size DG’s
Locations PLa PLr PLt in total loss
placed (Mw) (Mw) PLt (Kw)
1 61 1.8724 1.8724 15.8222 66.9810 82.8032 63.1234
17 0.5253
2 61 1.7826 2.3079 4.9054 66.6425 71.548 68.136
11 0.564
18 0.3842 2.8916
3 2.6821 66.3740 69.2659 69.1523
61 1.7339
10 0.4024
224.5419
23 0.452
49 0.4401
4 2.9688 2.4124 66.4294 68.8418 69.331
61 1.6743

TABLE –IV. DG placement results of 69 bus test system TABLE –V. Comparison of 33 bus system DG placement
results
Methods Num Optimal Optimal Total Loss with 2 17 0.5253
ber locations sizes size DG’s
of (Mw) (Mw) (KW) SOS 61 1.7826 2.3079 71.54
DG’s
11 0.5597
ACO- 6 2.5775 2.5753 103.9
ACO-
ABC(11 21 0.3468 2.6224 69.42
1 ABC(11)
)
SOS 6 2.5737 2.5737 103.9649 61 1.7159
3 11 0.564
ACO- 13 0.8464
ABC(11 SOS 18 0.3842 2.6821 69.26
30 1.1588 2.0052 85.9
)
2 13 0.8463 61 1.7339

SOS 30 1.1585 2.0048 85.8983

V.CONCLUSION
14 0.7547
ACO- This paper presented a new algorithm called
24 1.0999 2.9260 71.4
ABC(11 Symbiotic Organism Search (SOS) for optimal
) 30 1.0714
placement of DG’s for loss minimization considering
3 14 0.7544 reverse power flow. Placement of different types of
SOS 24 1.1021 2.9287 71.4483 DG’s will results in loss reduction of both active and
30 1.0722 reactive components of currents. If the objective is to
reduce only real power loss, we can place type 1
7 0.9114
DG’s. The advantage of this algorithm compared to
ACO- 14 0.5858 other is that, it does not require any algorithm
ABC(11
) 24 0.9877 3.2000 65.9 controlling parameters and it is fast converging. The
31 0.7151 results obtained show the effectiveness of the
4 proposed algorithm for real power loss minimization
7 0.8536
in distribution systems.
14 0.5891
SOS 24 0.9843 3.115 66.0281 REFERENCES
31 0.6887 [1] Jiayi H., Chuanwen J. and Rong X., “A review on distributed
energy resources and microgrid,” Renewable and
Sustainable Energy Reviews vol 12, pp 2472–2483, 2008.
TABLE –VI. Comparison of 69 bus system DG placement [2] P. Chiradeja, and R. Ramakumar, “An approach to quantify
results technical benefits of distributed generation,” IEEE Trans.
Methods Num Optimal Optimal Total Loss with Energy, vol. 19, no. 4, pp. 764-773, Dec. 2004.
ber location sizes size DG’s [3] Waseem, M. Pipattanasomporn, and S. Rahman, “Reliability
of s (Mw) (Mw) (KW) benefits of distributed generation as back up source,” IEEE
DG’s
Power and Energy Society General Meeting, Calgary, 2009,
ACO- 61 1.8726 2.5753 83.32 pp. 1 - 8.
ABC(11) [4] H. M. Ayres, W. Freitas, M. C. De Almeida, L. C. P. Da
1 Silva, “Method for determining the maximum allowable
SOS 61 1.8724 1.8724 82.80
penetration level of distributed generation without steady-
ACO- 18 0.5309 state voltage violations’, IET Gen. Trans. Dist., vol. 4, no. 4,
ABC(11) pp. 495–508, 2010.
61 1.7818 2.3127 71.65

447
 2016 International Conference on Advanced Communication Control and Computing Technologies (ICACCCT)

[5] N. Acharya, P. Mahat, and N. Mithulananthan, “An

analytical approach for DG allocation in primary distribution
network,” Int. J. Electr. Power Energy Syst., vol. 28, no. 10,
pp. 669-678, 2006.
[6] M. P. Lalitha, N.S. Reddy, and V. V. Reddy, “Optimal DG
placement for maximum loss reduction in radial distribution
system using ABC Algorithm,” Int J Rev. Comp. pp. 44-52,
2009.
[7] M. Kefayat , A. Lashkar Ara , S.A. Nabavi Niaki. Energy
Conversion and Management 92 (2015) 149–161
2014Gandomkar, M., Vakilian, M., Ehsan, M.: “A
combination of genetic algorithm and simulated annealing
for optimal distributed DG allocation in distributed
networks”. Proc. IEEE Electrical and Computer Engineering
Canadian Conf., 2005, pp. 645–648.
[8] J. A. M. García and A. J. G. Mena “Optimal distributed
generation location and size using a modified teaching–
learning based optimization algorithm” Electrical Power and
Energy Systems, vol. 50, pp. 65-75, 2013.
[9] F. S. Abu-mouti and M. E. El-Hawary, “Optimal distributed
generation allocation and sizing in distribution systems via
Artificial Bee Colony algorithm,” IEEE Transcations on
Power Delivery, vol. 26, no. 4, pp. 2090–2101, 2011.
[10] Moradi MH, Abedini M. A combination of genetic algorithm
and particle swarm optimization for optimal DG location
and sizing in distribution systems. Int J Elect Power Energy
Syst 2012;34:66–74.
[11] M. Kefayat , A. Lashkar Ara , S.A. Nabavi Niaki , “A hybrid
of ant colony optimization and artificial bee colony
algorithm for probabilistic optimal placement and sizing of
distributed energy resources”, Energy Conversion and
Management 92 (2015) 149–161, 2014

448