Babylon Search

ARTICLE JOURNAL`S UNIVERSITY OF BABYLON FOR
ENGINEERING SCIENCES (JUBES)

‫مــــــجلــة جـــــــامعة بـــــــــابــل للعلــــــــوم الهندسية‬
Vol. 32, No. 1. \ 2024 ISSN: 2616 - 9916
Optimal Action of Soft Open Points in Distribution Systems

Based on Single and Multiple Objective Particle Swarm
Optimization: A Comparative Approach
Hawraa A. Al-Ta'I Riyadh Toman Thahab
Department of Electrical Engineering/ University of Babylon

[email protected] [email protected]
Received: 20/8/2023 Accepted: 13/11/2023 Published: 15/2/2024
Abstract
Distribution networks of electrical power are becoming increasingly active through
integration of small generating units. However, the topology of the distribution network may
hinder large scale integration of those generating units thereby decreasing the benefits of an
active distribution network. Soft open points (SOPs) are considered one of the solutions to
alleviate the hampering of integration of local sources by altering the topology of the distributed
network.
In this paper, the problem of finding the optimal operating points of SOPs are analyzed
and investigated. Two approaches are adopted; by using single objective particle swarm
optimization (PSO) and multiple objective (MOPSO). In the former, a set of objectives are
optimized on a one by one bases whereas in the second method, the set of objectives are
optimized together at the same time. The set of objectives are the typical, active power loss,
voltage profile and feeder load balancing. A comparative approach is adopted in this paper with
the intention of finding the best approach that yield optimum points. Optimum points which are
defined as a set of active/reactive powers for each converter of the SOP. These optimum points
are further employed by designers as reference settings for the control system which governs the
local operation of the SOP system.
Results from simulations carried out on the IEEE 33 bus system with and without multi
capacity local energy sources, shows that the three objectives are better optimized using PSO
than MOPSO keeping in mind that the PSO requires less computation effort compared to
MOPSO. Moreover, better PSO performance is observed as the amount of power injected by a
distributed source (DG) increases. The former is also true when the system is optimized by
MOPSO.
Keywords: Soft Open Points, distributed generators, single optimization method, multi-objective
optimization method, back-to-back power converters, power electronic devices.
54
Vol. 32, No. 1. \ 2024 ISSN: 2616 - 9916
1. Introduction
Extensive use of distributed power resources leads to multiple problems including:
increased power loss ,voltage deviation and load imbalance. The topology of the
distribution network may significantly contribute to these problems. In distribution networks,
there are usually open points between adjacent feeders. These open points (switches) are
usually closed, while other switches are open to achieve proper load distribution between the
feeders. Previous researches has been done on network reconfiguration in order to achieve
optimal network operation (reduce power loss, optimize voltage profiles and achieve load
balancing), but this method is very limited in use due to the high cost of remote controlled
switches and maintenance of devices [1-3]. Recently, a method to improve network operation
without resorting to the network reconfiguration is the use of power electronic converters
called soft open points (SOP) [4-8]. These devices have been adopted to improve network
operation due to its ability to flexibly control the flow of active power in addition to some of
the advantages it possesses compared to reconfiguration of the network [7]. SOPs can be
used to connect any collection of feeders, for example, supplied from dissimilar substations
with different levels of nominal voltages resulting in many benefits [8].
Work presented in [4], investigated the ability of SOP in mitigating overload conditions.
Only one objective was considered in the analysis. Authors employed Powell’s direct set
(PDS) as the optimization tool to find the optimal operating points. The gains of using SOPs
in the medium voltage distribution network for three objectives, feeder load balancing, power
losses minimization and voltage profile improvement tested in IEEE 33bus using Powell’s
direct set method (PDS) is presented in [5]. Authors in [6 ] presented a multi-objective PSO
integrated by the taxi cab method. In this work, the elements in a limited size archive stores
the non-dominated solutions that represent the optimal operating points of an SOP. These
points were further refined through an employment of a local search method. The results
were obtained for only one location of a tie line for the IEEE69 bus system.
To determine the KW and KVAR power operating settings for an SOP on an 11 kV
network, a non-linear programming optimization was devised in [7]. The SOP's operational
zone was determined within its voltage-limit boundaries using a Jacobean matrix-based
sensitivity analysis, and it was then graphically shown for various demand and generation
situations at the feeders for both terminals of the SOP. Three optimization objectives:
reducing energy loss, improving the voltage profile and balancing line utilization were used
to pinpoint the precise operating point.
In this work, two optimization approaches are investigated. The first is optimizing the
SOP operating point based on a single objective PSO, in which optimizing of an objective
function is carried out one at a time. The second is based on a multi-objective PSO which
optimizes all targeted objectives at the same time. The main emphasis of the present work is
to provide a simple platform for obtaining the operating points of the SOP at different
locations of the tested distribution network. The accuracy and simplicity of the optimization
method will play a crucial role in providing reference settings for the converter control of an
SOP system. Each approach is tested under a variant amount of DG capacity.
54
Vol. 32, No. 1. \ 2024 ISSN: 2616 - 9916
2. Model of Soft Open Point in Distribution Networks

In distribution system, the classical topologies are either radial or ring type. The common radial
structure has number of bus bars that are equal to one plus the number of nodes in the system.
Some radial feeders have nodes that are normally open point (NOP) [9]. The idea of SOP is to
connect any two NOPs together using two power electronics converters, operated as back-to-
back with a middle DC link port [10]. Depending on the power flow, one converter is operated as
an AC/DC converter whereas the other operates as DC/AC. Figure 1.a shows a typical two
feeder distribution system and Fig. 1b shows an SOP which is connected between bus bars x &
y.
F1 F2
Bx
Substation By NOP
F3 F4
(a)
VSC1 VSC2
Bx By
(AC/DC) (DC/AC)
(b)
Fig.1. Schematic of, (a) simple two feeder distribution system and (b) SOP.
To completely judge the potential merits during non-transient operation of the network, the SOP
device is usually modelled as an injection or supply of active/reactive power [11]. In power
analysis of distribution networks, usual procedure is to find the power flow injected into a bus
based on the flow from the previous bus. Referring to fig.2, the power equation at bus is
given by [12],
= - – (1)
= - - (2)
where and are the real and reactive powers flowing out of bus , while are
the real and reactive powers flowing into bus , and are the real and
reactive load powers at bus ,
54
Vol. 32, No. 1. \ 2024 ISSN: 2616 - 9916
and / are the real/reactive power losses of the line section linking buses
and . The losses in the connection of buses and 1 are calculated as [13],
(3)
(4)
Based on Eq. (3) & (4), (1) & (2) can be rewritten as,
(5)
(6)
The voltage profile at bus can be expressed interms of voltage at bus as [13],
+ (7)
, in eq.(7) are the voltages at bus & respectively. The ohmic resistance and
reactance of the line section between buses & are indicated by
respectively
As far as the SOP operation is concerned, the active power exchange between the converters
must sum up to zeros [7], whereas reactive power flow remains dependent on the exchange at
each AC side. Hence, assuming a lossless converter and neglecting the losses in the interfacing
filter, the active power exchange is denoted as,
= (8)
= (9)
Where the active power flows for each converter of the SOP are denoted by and . Hence
the sum of power exchange between buses p &q is given as,
(10)
From Eq.(10), the relationship between either ends of the SOP system is given as,
(11)
54
Vol. 32, No. 1. \ 2024 ISSN: 2616 - 9916
Vi Vi+1
Feeder 1 p
Feeder 2
Fig.2. Network with an SOP at terminals p & q buses [7]

2.1 Constraints
The two VSCs' active power exchange and the voltage are restricted by [7],
(12)
(13)
(14)
(15)
Where and are the reactive power injections of each converter of the SOP . ,
are the nominal power of each VSC . and are the minimum and maximum
permitted voltages at buses to which the SOP is connected.
3.Optimization Formulation
Investigations were conducted into the advantages of SOPs for system ohmic losses, voltage
profile improvement and load balancing of a branch in the grid. Usual procedures involves
finding a solution to a set of combined nonlinear optimization problem subjected to some
enforced limits or constraints. Then the amounts of actual and reactive power injections of an
SOP can be calculated in order to quantify these advantages.
In this work, the improvements brought by SOP is initially calculated using the single objective
particle swarm optimization (PSO) technique using each of the three objective functions
separately. Then, to estimate the SOP improvement, we employ multi-objective particle swarm
optimization method (MOPSO) that considers all three objective functions simultaneously.
54
Vol. 32, No. 1. \ 2024 ISSN: 2616 - 9916
3.1 Objective Functions

The objectives of power loss reduction, improving the voltage profile and balancing the load, are
targeted to reach the best values of those functions. Apart from the load balancing factor, the
other two objectives are targeted for minimization. The following mathematical expressions
describe each objective:
Obj1: power loss reduction (P loss), which is formulated as,
(16)
Where is active power from bus , is the active power to bus , is the
total number of branches.
Obj2: voltage profile (VP), which is given by,
- )⎹ (17)
where is the voltage per bus-bar in the analyzed distribution network, 1 p.u. was taken as
for all bus-bars.
obj3: load balance index (LBI), which is given by,
(18)
where is the current flow in the branch connecting the bus to the next bus which is
, is the rated current of branch , In this study the feeders of the tested distribution
network are divided into two halves, the first half indicates that a current carrying capability, for
branches (1-9), to be 400 A. While the second half indicates that the current carrying capability,
for branches (10-32), is 200 A [ 14].
3.2 Optimization Algorithm Technique
3.2.1. Single objective Particle Swarm Optimization (PSO): PSO algorithm's performance in
providing an optimized objective has encouraged researchers to apply this biologically inspired
method to other domains.
The PSO algorithm was first put forth by James Kennedy and Russell C. Eberhart [16]. PSO is a
population-based search algorithm that mimics the social behavior of flocks of birds. Although
PSO was initially used to balance the weights in neural networks, it quickly gained popularity as
a global optimizer, especially for issues where the decision variables are real values . In order to
establish common terms for this algorithm, there are some definitions of many of the most
commonly used technical terms, which are detailed in [15]. Based on the details of the algorithm
as outlined in [15], particles are "flown" through a hyper dimensional search space in PSO.
Based on the social-psychological propensity of people to copy others' success, changes to the
position of the particles inside the search space are made. A population of randomly chosen
search sites known as particles is used to begin the search. A location vector (x) with M
45
Vol. 32, No. 1. \ 2024 ISSN: 2616 - 9916
dimensions of information is used to represent each particle; M is the number of choice

variables.
Based on [16], in subsequent repetitions, the location vector (x) is continuously modified using a
particle velocity. The velocity (v) of a particle is changed at every iteration using the two best
values. The first is the position that each particle has individually or personally earned ( ).
The other is the global best position ( ) attained by any population particle, which serves
as a roadmap directing the population to its finest location. The particle's velocity and
position update equations are [17] ,
) (19)
(20)
Where is the inertia weight that regulates how a particle's previous velocity affects its
current velocity, or how likely a particle is to keep moving in the same direction. The cognitive
learning factor symbolizes a particle's desire to its finest work [15]. The social learning factor
, represents a particle's attraction to the best of its neighbors here positive constants are
typically used to define and [15]
Two random values [0,1] called and are utilized to avoid being confined by a local
optimum and to allow for a diversity of particles in the search space [18].
3.2.2 . Multi-Objective Particle Swarm Optimization
To reflect the nature of multi_objective issues, where there are multiple objective functions to be
optimized instead of just one, the single objective formulation is expanded [15]. As a result, there
are multiple solutions rather than just one. Using Pareto Optimality Theory [19], this set of
solutions can discovered. Given that PSO is population-based, it is preferable to generate
multiple (different) non-dominated solutions in a single run. Therefore, while extending PSO to
Multi-objective optimization, the three essential difficulties to be taken into account, as with any
other evolutionary algorithm, [19],
1. What particles should be chosen to act as leaders in order to favor non-dominated solutions
over dominated ones?
2. How many non-dominated solutions discovered during the search process are retained such
that they can be reported as non-dominated solutions across all previous populations, not just the
present one? It is also ideal if these solutions are evenly distributed throughout the Pareto front.
3. How can the swarm's diversity be preserved to prevent convergence to a single solution?
In Multi-objective optimization issues, every particle may have a variety of leaders, but
only one can be chosen.
Such a group of leaders is typically kept apart from the swarm in what is called an external
archive [15]. The non-dominated solutions that have been discovered so far are kept in this
repository (A) [15]. When it is time to update the positions of the swarm's particles, the solutions
found in the exterior repository are used as leaders. Additionally, the external archive's contents
45
Vol. 32, No. 1. \ 2024 ISSN: 2616 - 9916
are frequently given as the algorithm's termination output. According to [15], the initialization of
the locations and velocities of the particles is random, and A is empty at the beginning of the
optimization. The beginning location is initialized as each particle's personal bests, then for each
iteration updates to each particle's velocities v and locations x are implemented in accordance
with eq. (19) and (20). The particle positions after updating might not be within the range of
workable solutions.
In this situation, it must be confined to the viable region. Solutions might also be
considered based on a given set of constraints. In order to make sure that repository (A) is a non-
dominating set, the set out objectives can be assessed with the vector of particle positions in the
feasible region . Solutions that are not weakly dominated by any member of the archive are then
added to A , and any components of repository (A) that are dominated by x are then removed
from A.
4. Simulation and Results of Distribution System
The IEEE 33 bus distribution network is used as a case study to illustrate effectiveness of
SOP to reduce power losses, improve voltage profile, load balancing using PSO and MOPSO
algorithms. Both algorithms are implemented in MATLAB . Figures (3 & 4) show the flow
chart of the PSO and MOPSO studied in this work. All load flow analysis is carried out using
MATPOWER 7.0 software [20]. This distribution network has a substation with 37 lines, where
32 of these lines are normally closed switches while five other are normally open switches with a
nominal voltage of 12.66 kV [21]. The total P and Q power consumption is 3715 KW and 2300
KVAR respectively. Five normally-open switches (the switches between buses 25 - 29, 33 - 18, 8
- 21, 12 – 22 and 9 - 15) are chosen as nominees venue for SOP connection in this study. Figure
(5) shows the IEEE 33 system with locations of SOP connections. In this study both PSO and
MOPSO are used to optimize the distribution network via different locations of the SOP system.
Table (1) shows parameter used in the simulations carried out in this study.
Table (1). Parameters of SOP system used in simulation
VSC apparent power, & 3 MVA
Nominal voltage Table (1). Parameters of 12.66 KV used in
SOP system
Number of populations simulation 40
Number of particles 4
Size of repository 20 (for MOPSO)
Number of iterations 60
C1, C2 (1, 2) [17]
r1,r2 (0-1) , (0-1) [18]
45
Vol. 32, No. 1. \ 2024 ISSN: 2616 - 9916
DG1
SOP1
SOP2
DG2 SOP3
SOP4
SOP5
Fig. 5. IEEE 33bus distribution network with SOP and

DG locations
45
Vol. 32, No. 1. \ 2024 ISSN: 2616 - 9916
4.1. Simulation Results for Base Case
In this case the SOP at any location is deactivated and the load flow is run only to find
ohmic power losses, voltage profile and load balance factor. Results show that the active loss is
is 208.459 kw, voltage profile index (VPI) is 1.6765 p.u with a smallest voltage of 0.91075 P.u,
recorded at bus 18 and the load balance index (LBI) is 4.5646. All the simulation cases presented
in this work will be compared against values of the base case. In this simulation case, lines
connecting buses 25 - 29, 33 - 18, 8 - 21, 12 – 22 and 9 – 15 are left open circuited.
4.2 Optimization of Soft Open Points System using Single Particle Swarm Optimization
The effects of various SOP installation on the three objective functions (power loss
minimization ,voltage profile improvement and feeder load balancing) are examined. In this
case, the VSC internal losses (switching devices and interfacing filter resistance) are neglected.
Applying single objective particle swarm optimization algorithm (PSO) that considers each
objective function separately, Table (2) below illustrate results of objective function for each
location of SOP in the network.
Table (2) Results of three objective using PSO/optimized separately
Objective(1): minimization of
Objective(2):Voltage profile Objective(3):Load balance
active power losses
improvement indexing
SOP
Location
P loss VPI LBI P loss VPI LBI P loss VPI LBI
8-21 123.355 0.950 3.846 187.421 0.526 6.091 138.884 1.211 3.620
9- 15 165.833 1.241 3.852 278.408 0.995 6.908 169.945 1.326 3.707
12-22 128.478 0.907 4.464 147.673 0.638 4.901 137.173 1.140 3.745
25-29 131.591 1.200 3.169 226.455 0.854 6.929 134.503 1.321 2.966
3.363
18-33 149.341 1.319 3.665 229.371 1.069 7.056 158.936 1.414
45
Vol. 32, No. 1. \ 2024 ISSN: 2616 - 9916
4.3 Optimization of Soft Open Points System using Multiple Objective Particle Swarm
Optimization
In this section of the simulation the SOP active/reactive powers are found through
optimization of the three objectives that are of vital importance in the distribution system
operation. Here a combination of Pareto optimal solutions for the SOP operational points without
DG penetration were obtained, which represent the Pareto frontiers [22]. MOPSO algorithm [23]
,which is augmented with MATPOWER functions and load data is used to find the Pareto
frontier extreme points on each axis. This shows the best values that can be found for each
objective function for all candidate places of SOP which is shown in Fig.6 below. The MOPSO
method's results for these extreme points are reported in Table (3) below. Here, it is seen that
solutions obtained vary according to the location of the SOP. Some locations such as, (8-21)
revels good active power loss decrease and significant voltage profile enhancement but no
significant impact of the load balancing factor. The second best location that impacts the power
loss and voltage profile is seen at (12-22) but again no that significant impact on the feeder
balance profile.
Table(3) results of three objective function using MOPSO
(P loss kw, VPI, LBI)
SOP Location
(124.815, 0.527, 3.622)
8-21
(166.97, 0.999, 3.731)
9- 15
(128.763, 0.640, 3.757)
12-22
(132.814, 0.858, 2.97)
25-29
(150.412, 1.073, 3.439)
18-33
From the results illustrated in Tables (2) and (3) it is noted that the objective functions values of
the SOP using single objective particle swarm optimization method (PSO) is better and more
accurate than those obtained by using multi-objective particle swarm optimization (MOPSO).
44
Vol. 32, No. 1. \ 2024 ISSN: 2616 - 9916
44
Vol. 32, No. 1. \ 2024 ISSN: 2616 - 9916
4.4 Optimization of Soft Open Points System using Single and Multiple Objective Particle
Swarm Optimization with Variable Capacity Distributed Generation Penetration
In this case two DGs are penetrated into the distribution system with a variable capacity.
The objective is to first study the performance of the distribution system with SOP
interconnection with these DGs. The second is to compare the performance of single PSO and
MOPSO in this case. The capacity of the DG is measured as a percentage of the total active
power load measured without any overload on the feeders [24]. The DG is considered as a type
1, injecting active power into the bus to which it’s connected [25]. In this work, the total power
delivered by these sources is divided between the two DGs equally. For this case, a bundle of
non-dominated solutions were obtained for each penetration using MOPSO method after
modifying the electrical grid data to reflect the injected DG power. Table 4 shows the results of
the MOPSO for a DG power ranging from 10%-50%. Figure 7 shows a graphical interpretation
of the obtained results. The voltage profile index, load balance index, and network power loss all
show a very significant decline. To show the usefulness of MOPSO, the results were compared
with PSO in which the algorithm was also modified to reflect DG interconnection. Here, each
objective function was optimized separately. Results for this case are shown in Table 5 and Fig.8
Clearly both methods show effectiveness in improving the system performance and producing a
well optimized distribution system. Moreover, as the percentage power of DG increase, both
methods produce nearly same optimization effectiveness.
Table(4) set of non-dominated solutions by using MOPSO method with DG

penetration(10%-50%)
Sop (8-21) SOP (9-15) Sop (12-22) Sop (25-29) Sop (18-33)
%DG Ploss VPI LBI Ploss VPI LBI Ploss VPI LBI Ploss VPI LBI Ploss VPI LBI
10% 107.58 0.482 3.071 139.09 0.89 3.1 108.83 0.58 3.17 111.63 0.82 2.51 121.18 0.94 2.88
20% 90.38 0.442 2.6 115.31 0.79 2.56 92.84 0.53 2.71 94.92 0.76 2.04 96.97 0.82 2.26
30% 82.94 0.422 2.3 97.46 0.71 2.18 80.23 0.49 2.37 80.01 0.72 1.69 77.05 0.71 1.84
40% 72.49 0.406 2.013 83.98 0.607 1.92 71.82 0.45 2.16 70.87 0.67 1.43 61.44 0.58 1.54
50% 67.36 0.406 1.88 75.15 0.529 1.73 67.36 0.44 1.99 58.94 0.63 1.2 50.24 0.47 1.31
44
Vol. 32, No. 1. \ 2024 ISSN: 2616 - 9916
Table 5. Objectives obtained using PSO method with DG penetration (10%-50%)
Results for one of the SOP location, line (8-21), is summarized in Table 6 below. Along with the
results obtained, the percentage improvement in each objective function is presented. The results
are shown for both approaches adopted in this study, that is SOP and MOPSO for each
44
Vol. 32, No. 1. \ 2024 ISSN: 2616 - 9916
percentage of DG capacity. It is clear that the percentage of improvement for the three objective
functions as a result of using MOPSO, starts to increase starting from 10% penetration rate and
up to 50%. This is also true for PSO, but here the percentage of improvement for each objective
function is separately determined. From this comparison the results appear better in the case of
using PSO than using MOPSO
Table(6) comparison of three objective function using PSO,MOPSO for one site of SOP (8-21)
between without DG penetration and with DG penetration for ratio percentage (10%-50%)
6.Conclusion
Investigations are implemented on the performance of a distribution system with soft open
points. A simple model is used that represents the SOP as power injectors/suppliers at/from a
specified location. Five different sits for SOP were investigated in the selected distribution
network, and it is concluded that the SOP site plays a significant part on the amount of
improvement in the operation and optimization of the network. Results obtained by applying
PSO, which optimizes one cost function at a time, revealed that the percentage improvements in
(Ploss, VPI ,LBI), for the case of using one SOP, considering site (8-21) as an example with a
DG penetration capacity of 50%, are (45.95%, 23.07 %, 48.61% ) compared to those values for
base case with no SOP and DG penetration. This indicates the superiority of SOP to impact
positively the three objective functions. A comparison between employing single objective
particle swarm optimization method (PSO) on the network, using injected power for each VSCs
as a decision variable, with a Multiple objective Particle Swarm Optimization method (MOPSO)
confirms better results in favor of PSO. Results showed that the three objective functions were
better using PSO compared to MOPSO. It is also concluded that as DG percentage penetration
increased, percentage improvement for each objective function also increased . This conclusion
44
Vol. 32, No. 1. \ 2024 ISSN: 2616 - 9916
coincides for both single PSO and MOPSO. The present study provides a platform procedure for
obtaining reference power settings for VSC control system designs.
References
[1] M. A. Kashem, V. Ganapathy, and G. B. Jasmon, “Network reconfiguration for load balancing in
distribution networks,” Generation, Transmission and Distribution, IEE Proceedings, vol. 146, no. 6, pp. 563-
567, 1999.
[2] Baran ME, Wu FF, “Network reconfiguration in distribution systems for loss reduction and load
balancing”, IEEE Power Engineering Review ., vol. 9, no. 4, pp.101–102, 1989.
[3] A. Botea, J. Rintanen, and D. Banerjee, “Optimal Reconfiguration for Supply Restoration With Informed
A* Search,” Smart Grid, IEEE Transactions on, vol. 3, no. 2, pp. 583-593, 2012.
[4] W. Cao, J. Wu and N. Jenkins, "Feeder load balancing in MV distribution networks using soft normally-
open points," IEEE PES Innovative Smart Grid Technologies, pp. 1-6, 2014,doi:
10.1109/ISGTEurope.2014.7028874.
[5] Wanyu Cao a , Jianzhong Wu a , Nick Jenkins a , Chengshan Wang , Timothy Green ,” Benefits analysis of
Soft Open Points for electrical distribution network operation”, ELSEVIER. Applied Energy., vol. 165, no.C,
pp. 36-47, March 2016.
[6] Qi Qi, Jianzhong Wu , Chao Long, “Multi-objective operation optimization of an electrical distribution
network with soft open point “ ELSEVIER. Applied Energy., vol. 208, pp. 734-744, 15 December 2017.
[7] Chao Long, Jianzhong Wu , Lee Thomas, Nick Jenkins, “Optimal operation of soft open points in medium
voltage electrical distribution networks with distributed generation” , ELSEVIER. Applied Energy, vol. 184,
pp. 427-437, 15 December 2016.
[8] Romero-Ramos E, Gómez-Expósito A, Marano-Marcolini A, Maza-Ortega JM, Martinez-Ramos JL,
“Assessing the loadability of active distribution networks in the presence of DC controllable links”, IET Gener
Transm Distrib ,vol. 5, pp. 1105 – 1113, November 2011.
[9] Peng Li, H. Ji, Chengshan Wang, Jinli Zhao, G. Song, F. Ding, Jianzhong Wu, “Optimal Operation of Soft
Open Points in Active Distribution Networks Under Three-Phase Unbalanced Conditions”, IEEE
TRANSACTIONS ON SMART GRID, DOI:10.1109/TSG.2017.2739999,2019.
[10] Bloemink JM, Green TC, “Benefits of distribution-level power electronics for supporting distributed
generation Growth”. IEEE Trans. Power Deliv., vol.28, no.3, pp.911–919, 2013.
[11] Chengshan Wang , Guanyu Song , Peng Li , Haoran Ji, Jinli Zhao , Jianzhong Wu ., “Optimal siting and
sizing of soft open points in active electrical distribution networks”, Apply Energy, vol. 189, pp. 301–309,
2017.
[12] Baran, Mesut, and Felix F. Wu. "Optimal sizing of capacitors placed on a radial distribution
system." IEEE Transactions on power Delivery, vol. 4, no.1, pp. 735-743, 1989.
[13] R. Srinivasas Rao a, S.V.L. Narasimham , M. Ramalingaraju , “Optimal capacitor placement in a radial
distribution system using Plant Growth Simulation Algorithm”, ELSEVIER. Electrical Power and Energy
Systems , vol. 33 , no. 5, pp. 1133–1139, 2011.
[14] Arif Wazir, Naeem Arbab, “ Analysis and Optimization of IEEE 33 Bus Radial Distributed System Using
Optimization Algorithm” , Journal of Emerging Trends in Applied Engineering, vol.1, no. 2 , 2016.
[15] Reyes-Sierra M, Coello CAC, “ Multi-objective particle swarm optimizers: a survey of the state-of-the-
art” , International Journal of Computational Intelligence Research, vol. 2, no.3, pp. 287–308 ,2006.
[16] James Kennedy and Russell C. Eberhart, “Particle swarm optimization.”, In Proceedings of the 1995
IEEE International Conference on Neural Networks , vol. 4, pp. 1942–1948,1995.
[17] Stoppato A, Cavazzini G, Ardizzon G, Rossetti A., “A PSO (particle swarm optimization)-based model
for the optimal management of a small PV(Photovoltaic)-pump hydro energy storage in a rural dry area”,
ELSEVIER. Energy, vol. 76, pp. 168–174, 2014.
[18] Parsopoulos KE, Vrahatis MN., “Particle swarm optimization method in multiobjective problems”, ACM
Symposium on Applied Computing, pp. 603-607, 2002
45
Vol. 32, No. 1. \ 2024 ISSN: 2616 - 9916
[19] Carlos A. Coello Coello, David A,Van Veldhuizen, and Gary B. Lamont, “Evolutionary Algorithms for
Solving Multi-Objective Problems”, Kluwer Academic Publishers, May 2002.
[20] R. D. Zimmerman, C. E. Murillo-Sánchez and R. J. Thomas, "MATPOWER: Steady-State Operations,
Planning, and Analysis Tools for Power Systems Research and Education," in IEEE Transactions on Power
Systems, vol. 26, no. 1, pp. 12-19, Feb. 2011, doi: 10.1109/TPWRS.2010.2051168.
[21] Diego Feroldi, Pablo Rullo, “Optimal Operation for the IEEE 33 Bus Benchmark Test System With
Energy Storage”, IEEE URUCON , DOI:10.1109/URUCON53396.2021.9647175, 2021.
[22] Gunter Rudolph, “On a multi-objective evolutionary algorithm and its convergence to the pareto set” , In
Proceedings of the 5th IEEE Conference on Evolutionary Computation, DOI: 10.1109/ICEC.1998.700081,
1998.
[23] Víctor Martínez-Cagigal (2023). Multi-Objective Particle Swarm Optimization (MOPSO)
(https://www.mathworks.com/matlabcentral/fileexchange/62074-multi-objective-particle-swarm-optimization
mopso), MATLAB Central File Exchange.
[24] V.H. Méndez. , J. Rivier , J.I. de la Fuente , T. Gómez , J. Arceluz , J. Marín , A. Madurga , “Impact of
distributed generation on distribution investment deferral”, ELSEVIER. Electrical Power and Energy Systems,
vol. 28, no. 4, pp. 244-252, May 2006.
[25] Rajendra Prasad Payasi1, Asheesh K. Singh , Devender Singh, “Planning of different types of distributed
generation with seasonal mixed load models” ,MultiCraft. International Journal of Engineering, Science and
Technology, vol. 4, no. 1 , pp. 112-124, 2012.
45
Vol. 32, No. 1. \ 2024 ISSN: 2616 - 9916
/
[email protected] [email protected]
(SOPs)
.
(SOPs).
(PSO)
(MOPSO)
SOP.
IEEE 33
PSO MOPSO PSO
PSO MOPSO
MOPSO DG
45

Babylon Search

Uploaded by

Copyright:

Available Formats

Babylon Search

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Babylon Search

Uploaded by

Copyright:

Available Formats

ARTICLE JOURNAL`S UNIVERSITY OF BABYLON FOR

ENGINEERING SCIENCES (JUBES)

Optimal Action of Soft Open Points in Distribution Systems

Department of Electrical Engineering/ University of Babylon

Received: 20/8/2023 Accepted: 13/11/2023 Published: 15/2/2024

2. Model of Soft Open Point in Distribution Networks

Fig.2. Network with an SOP at terminals p & q buses [7]

3.1 Objective Functions

obj3: load balance index (LBI), which is given by,

dimensions of information is used to represent each particle; M is the number of choice

3.2.2 . Multi-Objective Particle Swarm Optimization

Fig. 5. IEEE 33bus distribution network with SOP and

4.1. Simulation Results for Base Case

Table (2) Results of three objective using PSO/optimized separately

9- 15 165.833 1.241 3.852 278.408 0.995 6.908 169.945 1.326 3.707

Table(4) set of non-dominated solutions by using MOPSO method with DG

Table 5. Objectives obtained using PSO method with DG penetration (10%-50%)

You might also like