ArticlePDF Available

Cooperative Energy Transactions in Micro and Utility Grids integrating Energy Storage Systems

Authors:

Abstract and Figures

Over the past few years, Micro Grids (MGs) have gained much popularity due to two-way communication in the power network with special emphasis on Distributed Energy Resources (DERs), which comprise of both Renewable Energy Sources (RESs) and Nonrenewable Energy Sources (NRESs). Currently, the main focus of researchers is to deal with the intermittent nature of RESs, which lead to the fluctuations in power production and dispatch. In this paper, direct energy trading among MGs is considered as an assuring solution for improving the grid stability, reducing power line losses and minimizing energy trading cost. The focus of this work is on energy transactions amongst multiple MGs within the same geographic region. In the proposed method, coalitions among MGs are made on the basis of the distance between them for energy transaction. Furthermore, an Energy Transaction Algorithm (ETA) is proposed for energy trading among MGs in the coalitions. Energy Storage System (ESS) is also integrated, which stores energy when an MG has surplus energy and utilizes the stored energy if it becomes energy deficient. Simulations are performed and the results demonstrate that energy transactions among MGs in the proposed method minimize the power line losses and energy trading cost up to 34.6% and 14%, respectively as compared to the existing method.
Content may be subject to copyright.
1
Cooperative Energy Transactions in Micro and
Utility Grids integrating Energy Storage Systems
Muhammad Usman Khalid1, Nadeem Javaid1,2,, Ahmad Almogren3,,
Abrar Ahmed4, Sardar Muhammad Gulfam4, Ayman Radwan5
1Deaprtment of Computer Science, COMSATS University Islamabad, Islamabad 44000, Pakistan,
2School of Computer Science, University of Technology Sydney, Ultimo, NSW, 2007, Australia,
3Department of Computer Science, College of CIS, King Saud University, Riyadh 11633, Saudi Arabia,
4Department of ECE, COMSATS University Islamabad, Islamabad 44000, Pakistan,
5Instituto de Telecomunicacoes and Universidade de Aveiro, Aveiro 3800, Portugal.
Correspondence: nadeemjavaidqau@gmail.com, ahalmogren@ksu.edu.sa
Abstract—Over the past few years, Micro Grids (MGs) have1
gained much popularity due to two-way communication in the2
power network with special emphasis on Distributed Energy3
Resources (DERs), which comprise of both Renewable Energy4
Sources (RESs) and Nonrenewable Energy Sources (NRESs).5
Currently, the main focus of researchers is to deal with the6
intermittent nature of RESs, which lead to the fluctuations in7
power production and dispatch. In this paper, direct energy8
trading among MGs is considered as an assuring solution for9
improving the grid stability, reducing power line losses and10
minimizing energy trading cost. The focus of this work is on11
energy transactions amongst multiple MGs within the same12
geographic region. In the proposed method, coalitions among13
MGs are made on the basis of the distance between them14
for energy transaction. Furthermore, an Energy Transaction15
Algorithm (ETA) is proposed for energy trading among MGs in16
the coalitions. Energy Storage System (ESS) is also integrated,17
which stores energy when an MG has surplus energy and18
utilizes the stored energy if it becomes energy deficient. Simu-19
lations are performed and the results demonstrate that energy20
transactions among MGs in the proposed method minimize the21
power line losses and energy trading cost up to 34.6% and22
14%, respectively as compared to the existing method.23
Index Terms—Microgrid, Distributed Energy Management,24
Power Line Losses, Energy Storage System, Energy Trading.25
HIGHLIGHTS26
A cooperative energy trading mechanism for MGs is27
proposed.28
Coalitions are formed amongst MGs on the basis of29
distance.30
Analyses of power losses and energy trading cost with31
and without ESS is performed.32
I. INTRODUCTION33
Electricity has become the fundamental part of human34
life. Its demand is increasing exponentially due to the35
massive increase in population and technological advance-36
ments. The exponential increase in power demand resulted37
in serious environmental hazards. The engineering industries38
are pondering on the ways to produce and utilize energy39
resourcefully [1], [2]. Looking at the power consumption40
patterns, it is observed that at certain hours, power demand41
is higher than other times. These high demand hours bear the42
higher electricity cost and are known as On-peak timings.43
In contrast, Off-peak timings are the ones where electricity44
demand and price are lower. Hence, a day can be divided45
into two kinds of intervals: Off-peak and On-peak [3]. Both46
intervals have seasonal and geographical variations. In the47
summer season, On-peak intervals are observed during noon48
or afternoon; however, in winter or autumn seasons, these49
intervals are regarded as Off-peak intervals. In Conventional50
Grid (CG), utility companies meet the increasing electricity51
demand through spinning reserves or Peaking Power Plants52
(PPP). Moreover, the high electricity demand is met by53
conventional energy sources such as coal, oil, gas, etc., which54
are not environment-friendly and produce harmful emissions.55
Therefore, there is a need for some intelligent framework56
that can meet the increasing and varying electricity demand.57
Moreover, such a system also tends to eliminate the need for58
spinning reserves and PPP.59
The concept of Smart Grid (SG) that revolutionized the60
power system came into view in the 20th century. SG in-61
tegrates Information and Communication Technology (ICT)62
between the Utility Grid (UG) and end users. Also, it enables63
the bi-directional flow of both information and electricity64
between the main grid and consumers. SG has the capability65
to fully exploit the energy resources in the power system66
and eliminate the need for spinning reserves and PPP. Micro67
Grid (MG), which is a small scale local energy generation68
system, is the fundamental part of a SG. It facilitates the69
electricity consumers [4]. Moreover, it helps the people living70
in remote areas to have uninterrupted electricity access. It has71
two modes of operation: islanded mode and grid-connected72
mode. In the former mode, MG operates autonomously73
isolated from the main grid and facilitates the users con-74
nected to it. In the latter mode, MG and main grid are75
linked with each other at the Point of Common Coupling76
2
(PCC). It is the point at which the voltage of MG is equal77
to the voltage of the main grid. The connection of MG78
with the main grid is established through a Macro Station79
(MS), which is also known as a substation. The conventional80
power system requires a centralized transmission medium81
to transfer electricity from the generation side to the end82
users. In contrast, MGs are composed of Distributed Energy83
Resources (DERs), which are decentralized in nature. DER84
reduces the generation and transmission costs, and results in85
fewer line losses because it is located close to the serving86
load [5], [6]. In DER, Distributed Generation (DG) sources87
use Renewable Energy Sources (RESs) such as biomass,88
biogas, geothermal power, Photovoltaic (PV) panels and89
wind turbine, which fulfill 14% of the world’s total energy90
demand [7]. Also, RESs are environment-friendly. However,91
they are intermittent in nature due to varying environmental92
conditions. In case of PV panels and wind turbines, energy93
generation depends upon solar irradiance and wind speed,94
respectively. Therefore, for the reliable operation of an MG,95
there is a need for connection with the main grid, which acts96
as a backup power source or some reliable power generation97
source for the MG operation [8].98
In a power system, satisfying the electricity demand in99
an efficient manner is the basic objective. Instead of using100
spinning reserves or PPP to meet the electricity demand,101
it can be fulfilled through supply and demand-side manage-102
ment. In supply-side management, power can be balanced by103
reducing power line losses, increasing electricity generation104
and efficiently utilizing the energy resources. In demand-105
side management, power can be balanced by modifying the106
users’ energy consumption patterns using Demand Response107
(DR) programs. In order to modify the energy consumption108
patterns, optimization techniques are utilized for appliance109
scheduling on the basis of the pricing signal provided by110
the utility company. The overall electricity bill is reduced111
due to scheduling. In addition, Peak to Average Ratio (PAR)112
is also minimized. The reduction in PAR directly benefits113
the utility companies by minimizing the need for spinning114
reserves or PPP. One of the main factors of power imbalance115
is power line losses, which are caused due to electricity116
transmission from power plants to the consumers through117
extensive networks.118
This research work is an extension of [9]. The main focus119
is the establishment a next-generation power system, which120
comprises multiple MGs connected to the main grid for121
efficient energy management. The objectives of this paper122
are to meet the users’ energy demand, alleviate power line123
losses and maximize the economic benefits. In order to meet124
the users’ energy demand, a cooperative model is proposed.125
In the Proposed Method (PM), MGs trade energy with other126
MGs and the UG through a coalition. The coalition is defined127
as the temporary alliance between two or more groups or128
parties for their mutual benefits. The coalitions are formed129
among MGs on the basis of two constraints [10]. First, the130
distance between any two MGs in a coalition should be131
less than or equal to the distance threshold. Second, each132
coalition must be comprised of at least two MGs for energy133
trading. After the coalition’s formation, MGs with surplus134
energy trade energy with the MGs having deficient energy.135
For energy trading among MGs, an Energy Transaction136
Algorithm (ETA) is proposed in which the power line losses137
and energy trading cost incurred during energy trading are138
calculated. Energy trading among MGs locally alleviates the139
power line losses. Besides, intra-coalition energy exchange140
brings economic benefits to the people. It is done by purchas-141
ing energy at a reduced price and selling it at an increased142
price as compared to the energy traded with the UG [11].143
Moreover, Energy Storage System (ESS) is integrated to144
further improve the economic efficiency. The following are145
the main contributions of the underlying work.146
A cooperative model, referred as PM, is proposed to147
minimize the power line losses.148
An ETA is proposed for trading energy with other MGs,149
and with UG and MG.150
Energy is traded between MGs present within the range151
of each other, and the UG and an MG.152
A comparative analysis of Existing Method (EM) [12]153
and PM is performed for the minimization of both154
power losses and energy trading cost.155
Effect on the socio-economic factor is analyzed due to156
ESS integration.157
The paper is organized as follows. The literature review158
is discussed in Section II. EM and PM are discussed in159
Section III. Also, Noncooperative Energy Sharing Algorithm160
(NCESA) and proposed ETA are discussed in Sections III-D161
and III-E, respectively. In Section IV, simulation results are162
demonstrated to evaluate the performance of the PM and163
ETA. Finally, the findings of the paper are concluded in164
Section V.165
A. Connection between Supercomputing Centers and Elec-166
tricity Service Providers167
The Supercomputing Centers (SCs) aim to perform high-168
performance computing, which require excessive resources169
and computational power. For performing such high com-170
puting, large amount of power is required. For that purpose,171
Electricity Service Providers (ESPs) have to support efficient172
electricity generation, transmission and distribution. Both173
UGs and MGs are examples of ESPs. The coalition in MGs174
play an important role in fulfilling the energy requirements175
of SCs. The MGs can supply surplus energy to the SCs at176
a reduced price as compared to the UG. Even if MGs do177
not have surplus energy, they can buy energy from the UG178
and then supply the energy to SCs, which is also feasible179
both for MGs and SCs. The authors in [13] presents a180
qualitative research on the service contracts between SCs and181
ESPs. The contracts involve various parameters like variable182
energy tariffs, demand charges, etc. An investigative study is183
performed in [14], which involves the evaluation of the rela-184
tionship and possible integration between SCs and respective185
ESPs. The authors in [15] carry out an extensive survey186
on integration of SCs and ESPs. The comparison between187
US based and European based grids is performed. For the188
comparison, demand management is taken into account. All189
3
works highlight the bright perspectives of integration of SCs190
and ESPs. The energy demand of SCs also191
II. LITERATURE REVIEW192
Nonrenewable Energy Sources (NRES) are found in the193
conventional energy systems. The fossil fuel based resources194
emit greenhouse gases, which adds to the air pollution.195
Furthermore, the expense of maintenance is also substantial.196
In contrast, MGs are composed of RESs, which are eco-197
friendly and have a low maintenance cost. Due to two-way198
communication and power flow, MGs can trade energy with199
the UG according to the needs. In [16], the authors evaluated200
smart homes with Plugin Hybrid Electric Vehicles (PHEV),201
ESS, and DG. The goal is to reduce the overall energy202
purchasing cost. To cut the power expenditure, the authors203
utilized a dynamic pricing method for appliance scheduling.204
Moreover, the concept of net metering is incorporated for205
energy trading. A coordination method is presented to control206
the load according to the transformer’s capacity limitations in207
order to minimize transformer overloading during low price208
intervals. Power line losses between MGs and between MG209
and UG, on the other hand, are not taken into account. In210
addition, the installation and maintenance expenses of the211
PV panel, ESS, and PHEV are not taken into account. The212
authors in [17] proposed a heuristic algorithm to tackle the213
electricity load requirements of IoT-enabled smart homes.214
Similarly, the authors in [18] proposed an optimized man-215
agement system to deal with the integration of RESs and ESS216
in the smart homes. The proposed work achieved significant217
reduction in electricity billing. In [19], the authors introduced218
a Multi-Agent System (MAS) for optimum coordination219
among DERs in order to improve power network efficiency220
and dependability. The MAS manages and dispatches power221
flow in a distribution network made up of numerous MGs222
that are linked to the main grid via substations. The goals are223
to minimize power line losses on the distribution network,224
reduce the load on the main grid, and maximize profit.225
For the development of coalitions among MGs, transferable226
utility cooperative Game Theory (GT) is employed. Surplus227
energy is shared with other MGs in the same coalition in228
order to balance power amongst MGs. According to the229
simulation results, MAS’s cooperative control outperforms230
noncooperative control. The authors, however, did not take231
into account the capacity of the substation voltage trans-232
former and the distance threshold for establishing coalition233
and consumers’ relation with related MG. The authors in [20]234
proposed a multi-agent based solution for energy manage-235
ment in integrated buildings and MGs. In the MGs, different236
RESs are used along with controllable loads. A coordinated237
system of DG and DR is used for the optimization of238
energy management system. Similarly, the authors in [21]239
proposed a stochastic predictive energy management system240
of multi MGs. The integration of MGs is done using a241
hierarchical control system. The proposed system ensures242
efficient energy management. However, it is confined to a set243
of predefined constraints and need to include more learning244
based methodologies so as to make it scalable.245
Distributed RESs fulfill energy demands well; neverthe-246
less, their variable nature impedes dependable operation.247
To meet the energy demand, one frequent approach is to248
deploy more generators and ESSs. Energy collaboration249
among MGs, on the other hand, is a compelling option.250
The authors in [22] investigated the effect of MGs’ energy251
cooperation and ESS on total energy costs. They took into252
account two MGs with DG, ESS, and aggregated load. The253
offline algorithm demonstrated that energy cooperation and254
ESS integration alleviated the varying nature of RESs while255
lowering overall energy cost. The authors also proposed256
two online algorithms for real-time energy management of257
cooperative MGs to reduce the computational complexity.258
The importance of cooperation for energy management in259
MGs can also be visualized from [23], [24]. On the same260
grounds, the energy management in residential MGs is261
ensured in [25], [26]. The authors in [27] discussed the262
coalition formation among MGs using GT. The establishment263
of a coalition has the goal of maximizing economic profit.264
MGs build alliances based on mutual benefits. Due to the265
reduced distance between MGs, coalition formations enhance266
average payout and minimize power line losses. The primary267
benefit of MG coalition creation is local power transfer,268
which reduces power flow to and from the main grid and269
results in minimum line losses. According to simulation270
results, MG coalitions enhance the average payout of each271
MG by up to 31% and considerably minimize power line272
losses. In [28], the authors proposed an optimal coalition273
formation strategy for MGs known as Hierarchical Priority274
based Coalition (HRCoalition). The primary goal is to lower275
the grid network’s total power line losses, reduce energy276
stress on the main grid, and maximize intra-coalition energy277
exchange. HRCoalition is used to build optimum coalitions278
based on distance threshold and physical connections be-279
tween MGs. A greedy energy exchange method is utilized to280
trade energy within the established coalitions. Shapley Value281
(SV) is used to share the advantages received through the282
establishment of a coalition of MGs. Traditional coalition283
creation techniques are computationally less efficient than the284
suggested coalition strategy. Consumers’ interactions with285
the corresponding MG, on the other hand, are not taken into286
account.287
The authors of [12] suggested a Game Theoretic Coali-288
tion Formulation Strategy (GT-CFS), in which MGs created289
coalitions based on differences in power line losses per290
power unit. Power line losses on the distribution network are291
reduced as a result of coalitions established among MGs.292
MGs use SV to divide the average reward generated by293
coalition formation fairly. However, client interaction with294
their assigned MG is not taken into account. Moreover, the295
integration of ESS for storing surplus energy is neglected.296
Contrarily, the authors [29], [30] dealt with the integration297
of ESS in energy management platforms considering uncer-298
tainty and bilateral trading, respectively. Zhenyu et al. [31]299
used the firework algorithm to optimize the economic and en-300
vironmental performance of MGs. The goals are to eliminate301
harmful emissions in the environment and to optimize MGs302
to reduce operational costs. In order to achieve the Pareto303
4
optimum point, the authors employed Gravitational Search304
Algorithm Operator (GSAO) to increase the quality of the305
obtained solution. The findings indicate that there is a trade-306
off between cost minimization and pollution reduction. The307
authors in [32] presented a scheduling method for optimizing308
the load associated with the UG. The goals of this article are309
to reduce power costs while balancing energy supply and310
demand.311
In paper [33], the authors discussed the utilization of RESs312
in a residential area. The integration of DERs improves the313
reliability of the power system and decreases the harmful314
emissions. To tackle the intermittent nature of RESs, the315
authors formed an energy trading strategy as a repeated316
game. For performing simulations, 12 homes are considered317
having RESs, which are interconnected with each other as318
well as with the MG. Simulation results show that energy319
generation using RESs saved 20% of the total energy cost320
while reducing the carbon emissions. The authors in [34]321
proposed different operational modes for forming clusters322
of interconnected MGs. To tackle the uncertainties in cre-323
ating a balance between collective and individual interests,324
chance constrained programming is used. The simulations325
proved that the proposed transactive energy approach is quite326
efficient. However, equal distribution of profits and cost327
saving is not ensured. The authors in [35] proposed a hier-328
archical stochastic energy management system to efficiently329
manage the operations of interconnected grids. Moreover,330
the uncertainties and unscheduled power exchange with331
the main grid are also minimized. However, the proposed332
work lacked in taking into account the communication and333
power line failures. The isolated mode of operation is also334
not dealt with. Daneshwar et al. [36] proposed a novel335
operational model for the interconnected MGs with 100%336
RESs in the transactive energy market. The novelty lies in the337
development of a free energy trading environment. However,338
only some specific parameters were used for modeling the339
system and more parameters need to be considered for better340
modeling. Mohammad et al. [37] discussed the MG’s load341
demand management in grid-connected mode. The objec-342
tive of this paper is to minimize the operational cost. A343
cooperative power sharing algorithm is used to trade energy344
within the grid network. Simulation results demonstrated that345
operational cost is minimized in the grid-connected mode346
as compared to the islanded mode in which no energy is347
traded within the grid. However, power line losses are not348
considered.349
In [38], the authors aimed to reduce power line losses350
as well as energy trading costs. They presented a coalition351
game strategy based method for reducing needless power352
transfers between an MG and a UG. MGs create alliances to353
swap power in exchange for transmission fees. The exchange354
of power across MGs increases the overall network’s cost355
efficiency and optimizes the average reward, which is shared356
among MGs via SV. The authors in [39] studied direct energy357
trading between Small Scale Electricity Suppliers (SSESs)358
and end users. The cooperation between SSESs and end359
users is formulated as a game. This game is solved through360
Coalitional Game Theory (CGT). The contribution of each361
SSES in a coalition is calculated through SV and the average362
payoff is distributed among SSESs according to their SV.363
Chao et al. focused on reducing line losses of the entire364
grid [40]. A greedy CGT algorithm is used to make coalitions365
among MGs on the basis of power line losses. ESS is366
also considered for storing the surplus energy. Simulation367
results show that power line losses are reduced significantly.368
However, energy trading cost is not calculated. The authors369
in [41] investigated the optimum scheduling of several MGs370
coupled in a hierarchical design. The goal is the energy371
peak reduction of each feeder to lower total operating cost,372
increase overall energy system profit and minimize the power373
flow from the distribution grid. An Energy Management374
System (EMS) is utilized at the lowest level of the proposed375
hierarchical design to compute the local optimum scheduling376
of MGs. At the highest level, the MG of MGs Centre (MMC)377
is employed, which generates a global ideal schedule for378
each MG in order to realize the minimum cost. If an MG379
has excess energy, it is shared with other MGs to enhance the380
total profit. When compared to the standard grid algorithm,381
the coordination technique reduces the overall system costs382
by 1.34 percent. Moreover, computational time is reduced383
significantly for obtaining the global schedule. However,384
power line losses are not calculated.385
The authors in [42], [43] studied the operations of ESS386
in MGs. Due to the intermittent nature of RESs, ESS is387
integrated to maintain the stability of an MG. A distributive388
strategy for cooperative control is used to maintain the supply389
and demand of the MG [42]. Two types of optimizations390
are performed in [43]. First, MGs are optimized individually391
through the noncooperative game. Afterwards, optimization392
of all MGs is conducted through the cooperative game.393
Simulation results show that the daily cost of MG is reduced.394
The authors in [44], [45] also worked on energy management395
in multi MGs. Xingzheng et al. [46] proposed an adaptive396
distributed energy scheduling scheme for optimal manage-397
ment of RESs in cooperative MGs. In order to cope up398
with RESs’ generation and demand, Lyapunov Optimization399
(LO) is used to make an online algorithm. The proposed400
algorithm achieved a near to optimal solution. Simulation401
results showed that the proposed algorithm reduced the op-402
erational cost and improved the RESs’ utilization. However,403
there is a trade-off between operational cost and battery size.404
To deal with efficient power scheduling using real time and405
critical peak pricing, authors proposed a hybrid optimization406
technique in [47]. The load is shifted from On-peak intervals407
to Off-peak intervals keeping in view the electricity cost. The408
proposed work achieved significant electricity bill reduction.409
For both cost and comfort based optimization of residential410
load in a SG, the authors implemented a dynamic program-411
ming technique along with two heuristic techniques in [48].412
The results showed a reasonable reduction in peak power413
consumption and energy cost with minimum user discomfort.414
Traditionally, MGs’ energy management is formulated415
as an offline optimization problem, which forecasts RESs’416
generation and market demand, which is difficult to achieve417
practically. On the other hand, online algorithms focused418
on balancing the power between supply and demand by419
5
simplifying the MG model while ignoring the underlying420
distribution network and power flow. Wenbo et al. [49]421
proposed an online algorithm for real-time MG operation422
while considering the underlying distribution network and423
associated power flow. Simulation results proved that the424
proposed online algorithm outperformed the offline greedy425
algorithm. In addition, the effect of the underlying distri-426
bution network on energy management is also studied. The427
authors in [50] proposed a mathematical model for energy428
resource scheduling and load management in isolated MGs.429
The focus is to mitigate the greenhouse gas emissions.430
Simulation results showed that smart charging of PHEV431
and the presence of DR effectively addressed the energy432
management problem and reduced the power line losses.433
Moreover, Battery ESS (BESS) and DR provide more ro-434
bustness and flexibility to the power system when energy435
demand is high. The authors in [51] proposed a distributed436
control strategy with the aim to deal with optimal dispatch437
of isolated MGs. In [52], the authors suggested that instead438
of sharing only surplus and deficient energy information439
with the community energy management system, adjustable440
power information must also be shared. The main objective441
is to reduce the operational cost. The authors proposed a442
hierarchical optimization algorithm for multiple MG system443
to trade energy with each other. DR is also incorporated for444
increasing supply reliability and minimizing operational cost.445
Simulation results showed that community ESS is a better446
option in order to reduce the operational cost.447
No doubt, a large amount of work has been done in the448
literature to ensure cooperative energy transaction in MGs.449
Still, there is a room to simultaneously discuss EM [12]450
and PM from the perspective of power losses’ minimization451
and energy trading cost reduction. The works discussed in452
literature review either focused on integration of ESS and453
cost reduction or the coalition of MGs with energy losses’454
minimization. The proposed study takes into account both the455
integration of ESS and MGs coalition. Moreover, the readers456
are provided with a deep insight of how and why coalitions457
should be established between MGs, and how both power458
losses and energy trading cost can be minimized.459
III. SYS TE M MOD EL460
In the proposed system, a distribution network having N461
number of MGs is considered. The MGs are also linked with462
the UG through an MS. In the proposed work, MGs are the463
main shareholders as their roles are decided depending on464
the amount of energy they have. They either sell energy to465
users or buy energy from other MGs or UG. On the other466
hand, the users are considered as the stakeholders because467
they are part of the entire system and are affected by the468
operations of MGs. Moreover, the technical aspects of the469
MGs are handled by MS or UG. For example, the clustering470
of MGs and the setting of energy trading price when MGs471
are in coalition is done by MS. On the other hand, energy472
trading price is decided by the UG when MG and UG are473
in coalition. Users are connected to the UG through an MG474
and an MS. Electricity is supplied to the end users through475
MGs. When an MG will be energy deficit, it will get energy476
from the UG and provide it to the users. From Fig. 1 it is477
seen that each user is connected with only one MG. User478
1 is connected with MG1. Similarly, User 2 is linked with479
MG2. It is also assumed that an MG is capable of meeting480
the demand of users connected to it. The accumulated users’481
demand is equal to the energy demand needed by the MG at482
any particular time t. In smart homes, a Smart Meter (SM)483
is commissioned to keep track of energy consumption. It has484
the ability to communicate with a specific MG. Moreover,485
the radial topology is used to establish a link between all486
MGs and the MS. Energy trading is conducted amongst487
MGs and UG via power distribution lines. In Fig. 1, two-488
way communication and distribution lines are found, which489
are represented by the dotted and solid lines, respectively.490
The energy management horizon considered in this work is491
24 hourly slots for a day. In any particular time slot t, the492
energy generation of ith MG is Giand the energy demand is493
Di. It is the users’ accumulated demand connected with the494
MG. Energy requirement of an MG in any particular time495
slot tcan be calculated using Equation 1.496
Ereqi=GiDi.(1)
If Ereqi>0, it means that the MG has the capability to497
generate and store surplus energy. It also sells surplus energy498
to other MGs or the UG. On the other hand, Ereqi<0 means499
that the MG is energy deficit and buys energy from other500
MGs or the UG. Moreover, Ereqi=0 represents that energy501
generation is equal to the energy demand. If Ereqi>0,502
MGs’ generation and demand are taken randomly. Under503
the described scenario, two distinct cases are studied: EM504
and PM. EM is the base model, which does not include any505
coordination amongst MGs. In it, energy trading takes place506
between MGs and UG. An NCESA is proposed for energy507
transaction between an MG and the UG. In the PM, MGs508
trade energy locally through coalitions. Moreover, energy509
is also traded with the UG. For trading energy, an ETA is510
proposed. It is assumed that the MG has enough money to511
buy energy from other MGs and the UG.512
A. Noncooperative Model513
In the EM, energy trading between an MG and the UG514
is conducted through the MS, which is connected to both515
grids. There is no energy trading among MGs. Both UG and516
MG are connected through a medium voltage line having517
a voltage rating of V0. The conventional power system is518
always accompanied with the power line losses due to I2R519
effect and the presence of a voltage transformer for voltage520
conversion. I2Rpower losses are severe if both MG and521
UG are located far away from each other. In addition, there522
are other factors for power losses, such as leakage current,523
corona effect, heat dissipation, dielectric losses and magnetic524
losses in transformers. When MGiwants to sell or buy energy525
to or from the UG, then the power line losses (P
i0) are526
calculated by Equation 2 [12].527
P
i0=I2
0Rio +αPtrn
i,(2)
6
Macro Sta on
Transmission Line
MG1
MGn
MG1
MG4MG1
MG3MG2
MGn
MGn
MG2
Coali on 1
N x N Km2Area
Coali on 2
Communica on Line
Energy storage system
Two-way power ow
. . . Coali
. . .
on n
Distribu on Line
Genera on Plant
Wind Farm Generator
User 1
Solar Panel
Two-way communica on
U lity Grid
Wind Farm Generator
User 2
Solar Panel Wind Farm Generator
User n
Solar Panel
Fig. 1. Illustration of the PM
where I0=Ptrn
i
V0is the current flowing over the distribution528
line when energy trading takes place between UG and MG.529
Rio is the distribution line’s resistance and Ptrn
iis the amount530
of energy being transferred between MGiand UG at initial531
voltage level, denoted as V0.αis the constant of power losses532
faced due to other factors like transmission loss, natural533
disasters, etc. After substituting I0in Equation 2, Equation534
3 is obtained.535
P
i0=(Ptrn
i)2Rio
V2
0
+αPtrn
i.(3)
Ptrn
ican be found out using Equation 4.536
Ptrn
i=
Ereqi,Gi>Di
E
i,Gi<Di
0,otherwise,
(4)
where E
iis the amount of energy produced or required by537
the UG to fulfill the demand of MGi, when it is energy538
deficient. E
iis more than Ereqibecause when energy is539
traded between the UG and the MG, then some power losses540
are incurred. If there are no power losses during energy541
trading, then E
ibecomes equal to Ereqi.542
E
it = (P
i0×t) + |Ereqi|
=(Ptrn
i)2Rio
V2
0
+αPtrn
i×t+Ereqi(5)
In Equation 5, E
it is the total energy produced or required543
by the UG, which is equal to the energy required by the MG544
plus the power losses. Equation 6 is used to calculate the545
energy trading cost among MGs or between UG and MG.546
ECshr =Ptrn
i.ϕ,(6)
where ϕis the unit power price. It is different when energy547
is traded to and from the UG. If an MG sells energy to the548
UG, the value of ϕis low and when an MG buys energy,549
then the value of ϕis high. In the EM, the objective is the550
MGs’ average payoff (υi) maximization, which is defined as551
follows.552
υi=ϕP
i0.(7)
B. Proposed Cooperative Model553
In PM, besides energy trading between MG and UG,554
energy exchange among multiple MGs is made possible.555
It happens through coalitions among MGs. The coalitions556
are made on the bases of two conditions: 1)there should557
be minimum of two MGs and 2)the distance between the558
MGs should be less than the threshold distance. Energy559
trading in coalitions alleviates the power line losses over560
distribution network and maximizes the average payoff as561
given in Equation 7. For the power line losses’ minimization,562
cooperative groups of MGs are made known as coalitions563
and denoted by S. In each coalition, there are two sets of564
MGs. The clustering is performed on the bases of energy565
7
surplus and deficiency. The entities with surplus energy are566
grouped into one cluster, termed as energy sellers Ss. On the567
other hand, the entities with energy deficiency are grouped568
into another cluster, termed as energy buyers Sb. Initially,569
energy is traded among MGs. If there is surplus energy,570
it is traded amongst MGs and UG. Ereqi>0 means that571
MGiis capable of selling surplus energy and it belongs to572
Ss. Similarly, Ereqi<0 means that MG jis energy deficient573
and it belongs to Sb. For coalition formation, there must574
exist one Ssand one Sb. Otherwise, there will be no energy575
trading among MGs. When energy is traded between MGi
576
and MG j, which act as seller and buyer MGs, respectively,577
then the power line losses are calculated by Equation 8.578
P
i j =(Pt rn
i j )2Ri j
V2
1
,(8)
where Ptrn
i j and Ri j are the power transfer and resistance579
between two MGs, respectively. V1is the voltage level at580
which energy is traded among MGs. When energy is traded581
among MGs within formed coalitions, then only I2Rlosses582
are witnessed and the value of αbecomes 0. The cases of583
energy trading among MGs within formed coalitions are as584
follows.585
Ptrn
i j =(Ptrn
i,Ptrn
iPtrn
j
Ptrn
j,otherwise,
where Ptrn
iand Ptrn
jare the energy transfer values of seller586
and buyer MGs, respectively. If MGicannot meet the demand587
of MG j, the seller MG only sells Ptrn
i, which is less than or588
equal to (Ptrn
j) of buyer MG. The power line losses in any589
coalition Scomprises of three parts.590
1) Power line losses caused by energy trading among MGs.591
2) Power line losses due to energy transfer from MG to592
the UG.593
3) Power line losses due to energy transfer from UG to the594
MG.595
The total power line losses in any coalition Scan be596
calculated using Equation 9.597
P
Total =P
i0+Pj0+P
i j,(9)
where P
i0and Pj0are the energy selling and buying power598
line losses, respectively. The total payoff function of any599
coalition Scan be determined using Equation 10.600
µ(S,Π) = ϕ1
iεSs
P
i0+ϕ2
jεSb
Pj0+ϕ3
iεSs,jεSb
P
i j,(10)
where ϕ1is the unit power energy selling price to UG by601
MG, ϕ2is the unit power energy buying price from UG602
by MG and ϕ3is the unit power price for energy trading603
among MGs. The price of unit power is fixed for all cases.604
The objective is to maximize the total payoff by minimizing605
the power losses.606
max
SNµ(S,Π)(11)
C. Integration of Energy Storage System607
When RESs expand, the demand for new and efficient608
methods to deal with fluctuations and uncertainties also arise.609
ESS is one of the options. It plays a vital role in energy610
infrastructure and brings cost savings to the utility companies611
as well as to the consumers. The major task of ESS is to store612
the surplus energy. The charging and discharging to and from613
the ESS of MG in any time interval tare represented as Ci,t614
0 and Di,t0, respectively. The charging and discharging615
efficiencies of the ESS show energy losses caused during616
both processes. The charging and discharging parameters are617
denoted by 0 <ψc
i<1 and 0 <ψd
i<1, respectively. In order618
to increase the life of ESS, energy is stored up to 90% and619
discharged down to 10% of the rated power.620
D. Noncooperative Energy Sharing Algorithm621
In NCESA, MGs are not involved in coalition formation.622
MGs only trade energy with the UG through MS. Each623
MG first fulfills the demand of users connected to it. After624
fulfilling the demand of users, MG shares surplus energy625
with the UG in order to maximize the economic benefits.626
When an MG is not viable to meet the users’ demand,627
then it gets the required energy from the UG and has to628
pay for it. In NCESA, power line losses are high due to629
energy transfer between an MG and the UG. It is because the630
power line resistance is high due to a large distance between631
both entities. Also, the economic benefit is less because UG632
purchases energy at low price and sells it at high price. Due633
to this, UG gets more benefit or profit as compared to the634
MG. In the case of ESS, MG with surplus energy first stores635
energy in ESS and then shares the remaining energy with636
the UG. Moreover, an energy deficit MG first obtains energy637
from ESS and then buys the remaining energy from the UG.638
NCESA pseudocode is shown in Algorithm 1.639
E. Proposed Energy Transaction Algorithm640
Proposed ETA is a game in which players make different641
choices to achieve the best possible outcome. In this scenario,642
MGs act as players and their choices to join the coalition643
decide the amount of benefits that can be achieved. The644
choices also decide the amount of power losses that can be645
reduced in the distribution network. Coalitions among MGs646
are made depending upon the distance threshold. Each MG647
can only join one coalition in a particular time interval t. The648
main objectives of the proposed algorithm are minimization649
of total power line losses and maximization of average650
payoff.651
When coalitions are formed, then energy trading takes652
place. Information regarding intra-coalitional trading in-653
cludes position, and amount of energy generation and de-654
mand of the MGs. In a coalition, MGs are sorted into two655
groups. Seller MGs (MGs) are sorted in one group in the656
descending order of their energy requirement. Buyer MGs657
(MGb) are sorted in another group in the descending order of658
their energy deficiency. Afterwards, the proposed algorithm659
selects the most energy deficit MG from MGband the MG660
8
Algorithm 1: NCESA Pseudocode
1begin
2Initialize total number of MGs N, total number of time slots T, initial state of ESS of each MGi
3Define ϕ1and ϕ3
4for t=1:T do
5for i=1:N do
6MGicalculates its E reqiusing Equation 1 and send Ereqito the UG
7if Ereqi>0then
8MGihas surplus energy and acts as a seller MG
9if ESSi<ESSmax
ithen
10 ESSstate =E SSi
11 ESSrequirement
i=ESSmax
iESSstate
12 if Ereqi>E SSrequirement
ithen
13 ESSi=ESSst ate +ESSrequirement
i
14 Ereqi=E reqiESSrequirement
i
15 else
16 ESSi=ESSst ate +Ereqi
17 Ereqi=0
18 end
19 end
20 Ptrn =E reqi
21 Power line losses and energy trading cost are calculated using Equation 3 and 6, respectively
22 end
23 if Ereqi<0then
24 MGihas deficient energy and acts as a buyer MG
25 if ESSi>ESSmin
ithen
26 ESSstate =E SSi
27 ESSrequirement
i=ESSstate E SSmin
i
28 if ESSrequirement
i>Ereqithen
29 ESSi=ESSst ate Ereqi
30 Ereqi=0
31 else
32 Ereqi=E reqiESSrequirement
i
33 ESSi=ESSmin
i
34 end
35 end
36 Ptrn =E reqi
37 Power line losses and energy trading cost are calculated using Equation 3 and 6, respectively
38 end
39 end
40 end
41 end
with maximum surplus energy from MGs. If the demand661
of energy deficit MG is fulfilled by MGs, the next energy662
deficit MG is selected. Otherwise, the next MG with surplus663
energy is selected to fulfill the need of the buyer MG. The664
energy trading cost of MGs is calculated by Equation 6.665
Power line losses due to energy trading among MGs are666
calculated by Equation 8. After energy trading is performed667
within formed coalitions, the remaining energy required by668
the deficit MG is provided by the UG and MGstrades the669
excessive energy with the UG. For calculating power line670
losses and energy trading cost, Equation 3 and Equation 6671
are used, respectively. If ESS is integrated, an MG first stores672
or obtains energy to or from the ESS, respectively. After that,673
energy trading takes place within formed coalitions and with674
the UG.675
IV. SIMULATION RESULTS AN D DISCUSSION676
In this section, the capabilities of the PM and ETA677
are validated through a series of experiments performed in678
MATLAB. For simulations’ purpose, a distribution network679
of 10 x 10 km2and having 15 MGs is taken into account.680
Fig. 2 shows random deployment of MGs in a certain area681
with MS being located at the center. MS is considered as the682
origin of a 2-D space with coordinates (0, 0). It is assumed683
9
Algorithm 2: Proposed ETA Pseudocode
1begin
2Initialize total number of MGs N, total number of time slots T, initial state of ESS of each MGi
3Define ϕ1,ϕ2and ϕ3
4for t=1:T do
5for i=1:N do
6MGicalculates its E reqiusing Equation 1
7if Ereqi>0then
8From line 9 to 19 of Algorithm 1
9else
10 From line 25 to 35 of Algorithm 1
11 end
12 end
13 for c=1:C do
14 Sort buyer MGs MGbin descending order of energy demand
15 Sort seller MGs MGsin descending order of surplus energy
16 for y=1:B do
17 P
req =MGb
y
18 for z=1:S do
19 P
sell =MGs
z
20 if P
sell >0 && P
req >0then
21 P
rem =P
req P
sell
22 if P
rem >0then
23 P
sell =0, MGs
z=0 and MGb
y=P
rem
24 Energy trading cost and power line losses are calculated using Equation 6 and 8, respectively
25 else
26 P
req =0, MGb
y=0 and MGs
z=P
sell P
req
27 Energy trading cost and power line losses are calculated using Equation 6 and 8, respectively
28 end
29 end
30 end
31 end
32 end
33 for i=1:N do
34 if Ereqi>0then
35 Power line losses and energy trading cost are calculated using Equation 3 and 6, respectively
36 else
37 Power line losses and energy trading cost are calculated using Equation 3 and 6, respectively
38 end
39 end
40 end
41 end
that MG’s energy demand is equal to the users’ demand. The684
demand and generation of an MG are regarded as Gaussian685
random variables that lie in the range of 1-3 MW. Moreover,686
the amount of energy required by the MG is computed using687
the energy generation and demand difference as given in688
Equation 1. MGs’ energy requirements vary from -2 to 2689
MW in any particular time slot t. If Ereq >0, MG acts as a690
seller and if Ereq <0, MG acts as a buyer. The values of line691
voltage and resistance are taken to be 11 kV and 0.4 /km692
in the case of an MG and an MS. While, in the case of two693
MGs, these values are taken to be 11 kV and 0.2 /km,694
respectively. When energy is transferred between MGs and695
between an MG and a UG, the price is taken to be fixed.696
MG pays 40 cents per kW for electricity purchased from UG.697
While MG charges 10 cents per kW when selling electricity698
to UG. The intra-coalition trading price for energy is fixed699
at 20 cents per kW. The power loss constant throughout700
energy trading between UG and MG, on the other hand, is701
assumed as α=0.02. Moreover, there exist different types702
of batteries that are designed for various purposes. In the703
proposed work, Lithium-Ion (Li-Ion) batteries are used. It704
is because they are mostly used in the MGs’ infrastructure.705
The reasons for using Li-Ion batteries are that they have high706
voltage capacity, increased energy density and require zero707
maintenance.708
Fig. 3 depicts the total energy generation and demand of709
10
multiple MGs for one day. Blue and red bars show energy710
generation and demand , respectively. An MG with a surplus711
amount of energy sells it to the other MGs or UG to get712
profit. Likewise, energy deficit MG buys energy from other713
MGs or UG to fulfill the user’s demand. It can be seen from714
Fig. 3 that MG number 2, 4, 6, 7, 9, 12, 13 and 15 act as715
seller MGs while remaining MGs act as buyer MGs. Table716
I shows the parameters of MGs considered for performing717
simulations. The positive value shows MG has surplus energy718
to sell while negative value shows that MG needs to purchase719
energy.720
Table II shows storage capacities of ESSs, which vary from
-5 -4 -3 -2 -1 0 1 2 3 4 5
X (km)
-5
Y (km)
5
4
3
2
1
0
MG 1
MG 2
MG 3
MG 4
MG 5
6
MG 7
MG 8 MG 9
MG 10
MG 11
MG 12
MG 13
MG 14
MG 15
MS
MG 6
MGs
MS
-4
-3
-2
-1
Fig. 2. Deployment of MGs and MS (N=15)
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
MG
-20
-10
0
10
20
30
Power (MW)
Energy Generation
Energy Demand
Fig. 3. MGs surplus and deficient energy
721
0 to 1 MW. ESS charging and discharging limits are set at722
90% and 10%, respectively.723
A. Noncooperative Model with and without Energy Storage724
System725
While considering EM, two cases are discussed: EM with726
ESS and EM without ESS. Fig. 4 shows power line losses727
of MGs for one day in the EM without considering ESS.728
MGs initially satisfy the users’ demand and then trade the729
excessive energy with the UG through an MS. If extra energy730
is needed to assist the consumers by an MG, it is bought from731
the UG. The power line losses faced during energy trading732
between UG and MGs are high because UG is located far733
from the MGs. In the figure, blue bars represent the total734
power losses incurred when selling surplus energy to the UG.735
While red bars show the total power losses incurred when736
receiving the deficient amount of energy from the UG to meet737
the users’ demand. In Fig. 5, the energy trading cost in the738
EM without considering ESS is demonstrated. When MGs739
sell surplus energy to the UG, they receive energy selling cost740
and when MGs purchase deficient energy from the UG, they741
pay the respective amount. Energy selling and purchasing742
costs are represented by blue and red bars, respectively.743
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
MG
0
0.5
1
1.5
2
Power Losses (MW)
Energy Selling Losses
Energy Purchasing Losses
Fig. 4. Power losses (EM without ESS)
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
MG
0
2000
4000
6000
8000
10000
Cost ($)
Energy Selling Cost
Energy Purchasing Cost
Fig. 5. Energy trading cost (EM without ESS)
The total power losses of the MGs for one day in the EM744
with ESS are shown in Fig. 6. The addition in EM is the745
inclusion of ESS. Hence, the system checks the status of the746
corresponding ESS that whether it is fully charged or not747
before trading energy with the UG. If MG has some extra748
energy and ESS is not fully charged, it charges the ESS up749
to 90% of the rated power and after that remaining energy is750
traded with the UG. Otherwise, if MG does not have enough751
energy to fully charge the ESS, it is charged to the maximum752
available energy and no energy is traded with the UG. If MG753
has deficient energy, it first checks the status of ESS. If ESS754
is charged, MG initially utilizes the stored energy and then755
buys the remaining energy from the UG. Red bars show the756
power losses caused when MGs purchase energy from the757
UG while blue bars show the power losses incurred when758
MGs sell energy to the UG.759
Fig. 7 shows energy trading cost regarding EM with ESS.760
While considering ESS, the energy traded with UG is less761
as compared to EM without ESS. Energy selling cost is762
represented by blue bars while red bars show the energy763
purchasing cost. The integration of ESS within the EM764
11
TABLE I MGs’ Parameters In Simulation Experiments
MG ID Deployment Position Total Energy
Generation (MW)
Total Energy
Demand (MW)
Energy
Requirement (MW)
1 (0.22315, -4.6766) 28.034 32.057 -4.0232
2 (-1.0264, 0.57067) 40.275 35.612 4.663
3 (-0.2089, 2.198) 37.731 37.909 -0.17791
4 (4.939, -3.8959) 33.027 32.927 0.10075
5 (1.0448, -2.8335) 29.446 34.842 -5.3959
6 (4.4491, 3.1102) 42.268 32.208 10.06
7 (-0.095578, -3.6134) 37.449 35.825 1.6238
8 (-0.62053, 3.819) 35.473 42.026 -6.5536
9 (2.7266, 4.2356) 37.757 37.474 0.28269
10 (2.4407, -4.8724) 31.405 35.808 -4.4033
11 (-0.57096, -1.2284) 38.578 43.277 -4.699
12 (-4.47, -3.3219) 35.830 33.703 2.1271
13 (-4.1218, 0.40223) 42.471 30.146 12.325
14 (2.9799, -3.9834) 30.585 36.255 -5.6697
15 (1.5558, -4.6073) 38.494 35.392 3.1015
TABLE II Capacities of ESSs in MGs
MG ID Total Capacity of ESS (MW)
1 0.96661
2 0.9858
3 0.68046
4 0.8221
5 0.53397
6 0.60396
7 0.5198
8 0.73468
9 0.57505
10 0.99565
11 0.71353
12 0.97769
13 0.86212
14 0.79045
15 0.77013
reduces losses and maximizes the economic benefits of the765
MGs. The simulation results of the EM with and without766
ESS are shown in Table III. The results show that both the767
power losses and energy trading cost are less when ESS is768
used as compared to when ESS is not used.769
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
MG
0
0.2
0.4
0.6
0.8
1
1.2
1.4
Power Losses (MW)
Energy Selling Losses
Energy Purchasing Losses
Fig. 6. Power losses (EM with ESS)
A comparative analysis is conducted for power line losses’770
minimization amongst EM without ESS and EM with ESS.771
Fig. 8 shows that MGs wih ESSs have minimum line losses772
as compared to MGs without ESS. The reason behind this773
is that ESS stores energy when MG has a surplus amount of774
energy and provides it when MG becomes energy deficient.775
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
MG
0
1000
2000
3000
4000
5000
6000
7000
8000
Cost ($)
Energy Selling Cost
Energy Purchasing Cost
Fig. 7. Energy trading cost (EM with ESS)
Hence, less amount of energy is traded with the UG. When776
the energy that flows to and from the UG decreases, the777
energy trading cost also decreases. A comparative analysis778
with respect to cost of energy trading in the EM with and779
without ESS is shown in Fig. 9.780
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
MG
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
Power Losses (MW)
Existing_Method without ESS
Existing_Method with ESS
Fig. 8. Comparative analysis of power losses
in EM
The following challenges are most likely to be faced when781
the MGs act independently.782
In the absence of coalition between MGs, there would783
be excessive energy losses. The MG with surplus energy784
would either have to store the energy in ESS or sell it785
back to the UG.786
12
TABLE III Simulation Results of EM with and without ESS
MG ID Power Losses
without ESS (MW)
Energy Trading
Cost without ESS ($)
Power Losses
with ESS (MW)
Energy Traing
Cost with ESS ($)
1 0.52497 4533.3 0.28289 2708.1
2 0.62637 5370.6 0.54799 4757.3
3 0.929 7223 0.64482 5248
4 0.73016 3966 0.42917 2493.9
5 1.2008 8852.3 1.0284 7852.9
6 1.7233 6840.3 1.2982 5277.6
7 0.65655 4545 0.41508 3004
8 0.88675 6648.4 0.71252 5594.6
9 1.0551 6171.4 0.75217 4711.3
10 1.029 6259 0.63062 4397.6
11 0.57073 6067.8 0.42217 4728.8
12 1.5127 7152.6 1.0752 5079.5
13 1.1464 5208.7 0.84893 3688.9
14 1.3359 8111.7 0.89966 6357.9
15 1.2622 6625.4 0.85766 4908.7
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
MG
2000
3000
4000
5000
6000
7000
8000
9000
Cost ($)
Existing_Method without ESS
Existing_Method with ESS
Fig. 9. Comparative analysis of energy trading
cost in EM
When MGs act independently, then they have to rely787
solely on the batteries to store the surplus energy as788
they are unable to sell energy to the nearby MGs. As a789
result, the batteries’ lifetime is reduced and the batteries790
are changed frequently, which lead to the increase in791
the maintenance cost. This increased cost is then paid792
by the consumers, when MGs sell energy to the users,793
which ultimately leads to a bad reputation of the MG.794
When there would not be any coalition between MGs,795
the energy deficient MG has to buy energy from the796
UG, which increases the energy cost.797
If MGs act independently, an increased burden on the798
UG will be observed as all energy requests are to be799
directly fulfilled by it.800
B. Proposed Cooperative Model with and without Energy801
Storage System802
As in EM, PM also has two scenarios: PM with and803
without ESS. Fig. 10 shows losses experienced in the PM804
by the MGs. In the PM without ESS, MGs make coalitions805
among themselves based upon the distance threshold and806
energy requirement. The distance threshold considered for807
coalition formation is 2.5 km and each coalition must have808
one seller and one buyer MG. MGs, after fulfilling the users’809
demand, trade excessive energy. If MGs still have extra810
energy, they sell it to the UG. In case, an MG’s energy811
demand is not met by the coalition, the remaining energy812
is bought from the UG. Power losses due to energy trading813
among MGs within coalitions are less as compared to the814
energy trading with the UG. The reason is that MGs are815
found in the close vicinity of each other as compared to the816
distance between MGs and UG. Fig. 10 shows the power817
losses incurred in the PM. Fig. 11 shows the cost of trading818
energy in the PM without considering ESS. Energy trading819
cost is high for seller MGs and low for buyer MGs. It is due820
to the reason that the prices for selling and buying energy to821
and from various MGs are high and low, respectively. MGs822
with surplus energy get more profit by selling energy to the823
MGs instead of UG. On the other hand, energy deficient824
MGs have to pay less energy trading cost if they buy energy825
from the MGs instead of UG. Energy trading cost of MGs826
in the PM is less as compared to the EM.827
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
MG
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
Power Losses (MW)
Energy Selling Losses
Energy Purchasing Losses
Fig. 10. Power losses (PM without ESS)
In the PM with ESS, a positive effect is observed on the828
power system. In Fig. 12, power losses experienced in the829
PM with ESS are illustrated. When ESS is integrated within830
PM, MGs first store energy in ESSs and then trade energy831
with MGs within the coalition and with the UG. If MGs832
are energy deficient, they first obtain energy from ESSs and833
then from other MGs within the coalition. If the demand is834
still not fulfilled, MGs buy the remaining energy from the835
UG. In Fig. 13, energy selling and buying cost is shown.836
Simulation results of the PM with and without ESS are837
shown in Table IV. The table shows that energy trading cost838
13
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
MG
0
1000
2000
3000
4000
5000
6000
7000
8000
Cost ($)
Energy Selling Cost
Energy Purchasing Cost
Fig. 11. Energy trading cost (PM without ESS)
is reduced as compared to the PM without ESS. It is because839
energy trading with other MGs and UG is decreased. Also,840
the cost of trading energy in the PM with ESS is less than841
the EM and PM without ESS.842
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
MG
0
0.2
0.4
0.6
0.8
1
1.2
Power Losses (MW)
Energy Selling Losses
Energy Purchasing Losses
Fig. 12. Power losses (PM with ESS)
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
MG
0
1000
2000
3000
4000
5000
6000
Cost ($)
Energy Selling Cost
Energy Purchasing Cost
Fig. 13. Energy trading cost (PM with ESS)
A comparative analysis is conducted for power line losses’843
minimization in the PM with and without ESS. Fig. 14844
shows that the power losses are minimized in the PM when845
ESS is integrated as compared to the PM without ESS. A846
comparative analysis of energy trading cost in the PM with847
and without ESS is shown in Fig. 15.848
C. Comparative Analysis of EM and PM849
In this subsection, the comparative analyses between the850
EM and the PM, from the perspective of power line losses851
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
MG
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
Power Losses (MW)
Proposed_Method without ESS
Proposed_Method with ESS
Fig. 14. Comparative analysis of power losses
in PM
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
MG
2000
3000
4000
5000
6000
7000
8000
Cost ($)
Proposed_Method without ESS
Proposed_Method with ESS
Fig. 15. Comparative analysis of energy trad-
ing cost in PM
and energy trading cost, are presented. The analyses prove852
the advantage of using the PM over the EM.853
1) Comparative Analysis of Power Line Losses’ Profiles:854
In Fig. 16, power losses of the EM and the PM, with855
and without considering ESS, are illustrated. The simulation856
results clearly show that the PM minimizes the power losses857
better than the EM. In the case of MGs with ESS, energy858
is stored when MG has surplus energy and is utilized in the859
hour of need. By utilizing the stored energy, power losses are860
decreased. The reduction is more in the PM as compared861
to the EM. Simulation results show that power losses are862
reduced by 34.6% in the PM as compared to the EM. Power863
losses are reduced by 28.6% and 26% due to the integration864
of ESS in the PM and the EM, respectively when compared865
with energy losses of both models without ESS. It means866
that the PM surpasses the EM by 2.6% in terms of power867
losses’ reduction when using ESS. The overall maximum868
power losses’ reduction is achieved by the PM with ESS,869
which is 51.6%.870
2) Comparative Analysis of Energy Trading Cost Profiles:871
Fig. 17 shows a comparative analysis of the EM and the PM872
with and without ESS. Energy trading cost of the PM with873
and without ESS is less as compared to the EM with and874
without ESS, respectively. It implies that the energy trading875
cost of MGs having deficient energy is reduced significantly876
in the PM because MGs provide energy at a lower rate than877
UG. Energy trading cost of MGs is lower because most of878
14
TABLE IV Simulation Results of PM with and without ESS
MG ID Power Losses
without ESS (MW)
Energy Trading
Cost without ESS ($)
Power Losses
with ESS (MW)
Energy Trading
Cost with ESS ($)
1 0.46736 4285.5 0.25297 2455.5
2 0.55085 5291.7 0.4805 4718.2
3 0.51079 5726.3 0.3657 3913.3
4 0.67022 3691.6 0.40086 2300
5 0.50231 6135.6 0.47744 5917.6
6 1.2239 5957.2 1.0201 4777.6
7 0.41086 4197.8 0.29224 2645.1
8 0.45321 5870.8 0.44836 5237.2
9 0.70272 5496.3 0.54381 4268.4
10 0.42877 3962.3 0.26554 2799.2
11 0.39816 5247.7 0.32764 4386.2
12 1.5127 7152.6 1.0752 5079.5
13 1.1464 5208.7 0.84893 3688.9
14 0.40105 5740 0.26499 3713.2
15 0.54081 6508.7 0.27641 5083
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
MG
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
Power Losses (MW)
Existing_Method without ESS
Existing_Method with ESS
Proposed_Method without ESS
Proposed_Method with ESS
Fig. 16. Comparative analysis of power
lossses: EM and PM with and without ESS
the energy is traded locally and energy deficit MGs purchase879
energy at a lower price from other MGs as compared to the880
UG. Moreover, because of the less distance between energy881
trading MGs, power losses are also low, which further reduce882
the cost of electricity. Also, maximum energy is stored by the883
ESS. Moreover, some MGs’ energy trading cost is equal to884
the EM because these MGs are located far away from other885
MGs and are not involved in any coalition. When an MG is886
not a part of the coalition, it solely exchanges energy with the887
UG, resulting in no cost savings during energy trading. The888
cost of energy trading in the PM is 14% less as compared889
to the EM. When ESS is integrated with the PM and the890
EM, then the overall energy trading cost is reduced by 24.3%891
and 24.2%, respectively. It means that the PM reduces energy892
trading cost more than the EM. Moreover, energy trading cost893
of both the EM and the PM with ESS is less as compared to894
the models in which ESS is not considered. This shows that895
ESS integration maximizes economic benefits by reducing896
energy purchasing cost of buyer MGs and increasing energy897
selling cost of the seller MGs. Moreover, maximum energy898
trading cost reduction is achieved by the PM with ESS,899
which is 34.8%.900
D. SoC Analysis of PM and EM901
From SoC perspective, both PM and EM with ESS would902
be considered. It is because SoC is linked with battery, which903
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
MG
2000
3000
4000
5000
6000
7000
8000
9000
Cost ($)
Existing_Method without ESS
Existing_Method with ESS
Proposed_Method without ESS
Proposed_Method with ESS
Fig. 17. Comparative analysis of energy trad-
ing cost: EM and PM with and without ESS
is used only in the mentioned cases. In the first case, a904
gradual change in SoC will be observed as compared to the905
third case. It is because the amount of energy being wasted906
is less in the former case than the latter case, which means907
the batteries would not drain quickly. Fig. 18 shows the908
accumulative average SoC value of 15 MGs for 24 hours909
during charging. The demand of all 15 MGs is added for910
obtaining the total electricity generated by all MGs. It is911
assumed that 10% of the total load is stored in ESS. For912
SoC, the upper and lower bounds are set as 90% and 10%,913
respectively. From the figure a gradual increase in the SoC is914
observed over the span of 24 hours during charging. Fig. 19915
shows the accumulative average SoC value of 15 MGs for916
24 hours during discharging. A gradual decrease is observed917
in the figure with increasing time.918
E. Test Analysis of the PM919
The proposed system is tested for a varying number of920
MGs, ranging from 15 to 75. The results provided in Table921
V show that the power losses increase with the increase922
in the number of MGs, both with and without ESS. The923
increase is gradual in the presence of ESS. Whereas, it is924
rapid when ESS is not employed in the system. Similarly,925
the energy trading cost is also increased with the increase in926
the number of MGs. The cost increases drastically with the927
increase in the number of MGs when ESS is not employed.928
15
Fig. 18. SoC behavior with respect to time
during charging
Fig. 19. SoC behavior with respect to time
during discharging
Contrarily, the increase in energy trading cost is steady in929
the case of ESS. Table V provides the simulation results for930
power losses and energy trading cost with and without ESS931
for different number of MGs. Both the power losses and932
the energy trading price are increased with the increase in933
the number of MGs because more MGs means more energy934
transactions and more use of resources. Moreover, if energy935
transactions are made in near real-time, the computational936
complexity will also increase. However, it would be within937
a feasible range. Moreover, it is assumed that an MS has938
enough computational resources to form coalitions between939
MGs and respond in near real-time. This ensures that the940
PM is robust, scalable and does not suffer when the size of941
the test system is increased.942
V. CONCLUSION AND FUTURE WO RK943
In this paper, a novel ETA is proposed for MGs. The944
proposed ETA allows MGs to decide whether to charge or945
discharge the ESS and trade energy with other MGs and UG946
or not. It is done so that power line losses can be minimized947
and financial gains can be maximized. The coalitions are948
formed in which energy trading is performed. Also, intra-949
TABLE V Simulation Results of PM for Increasing
Number of MGs
Cases Parameters MG=15 MG=30 MG=45 MG=60 MG=75
Without
ESS
Power
Losses
(MW)
3.99912 7.99824 11.99736 15.99648 19.99956
Energy
Trading
Cost ($)
29857.7 59715.4 89573.1 119430.8 149288.5
With
ESS
Power
Losses
(MW)
2.79317 5.58634 8.37951 11.17263 13.96585
Energy
Trading
Cost ($)
21950.8 43901.6 65852.4 87803.2 109754
coalition energy trading is employed, which is the backbone950
of the proposed model. Coalitions among MGs are made951
on the basis of the distance between them. The distance952
threshold value must be more than the distance between two953
MGs for forming coalitions and trading energy efficiently.954
By making coalitions among MGs, energy is traded locally955
due to which power losses are minimized and the load956
is alleviated from the UG that enhances the grid stability.957
Moreover, it is shown that the energy trading cost of deficient958
MGs is less because they purchase energy from other MGs959
at a lower price than purchasing energy from the UG. In960
contrast, MGs with surplus energy get more benefits by961
selling energy at a higher price to other MGs instead of962
selling at a lower price to the UG.963
Simulation results show that the PM and ETA outperform964
the EM in terms of power line losses’ reduction and eco-965
nomic benefit maximization. Moreover, comparison results966
show that losses of the system in the PM are reduced by967
34.6% and economic benefits of MGs are maximized up to968
14% as compared to the EM. ESS further reduces the losses969
by storing surplus energy and utilizing it in the hour of need.970
Results show that maximum losses’ reduction and economic971
benefit maximization are achieved by the PM with ESS. It972
reduces the power losses up to 51.6% and maximizes the973
economic benefits by 34.8%.974
In future work, the interaction of the customers with the975
MG and energy trading among neighbourhood houses at976
customer level will be studied.977
ACK NOW LE DG EM EN T978
This work was supported by King Saud Univer-979
sity, Riyadh, Saudi Arabia, through Researchers Sup-980
porting Project number RSP-2021/184. The work of au-981
thor Ayman Radwan was supported by FCT / MEC982
through Programa Operacional Regional do Centro and by983
the European Union through the European Social Fund984
(ESF) under Investigador FCT Grant (5G-AHEAD IF/FCT-985
IF/01393/2015/CP1310/CT0002)986
REFERENCES987
[1] The International Energy Agency (IEA) (2014) World energy outlook.988
http://www.iea.org/. Accessed Jun 2018989
[2] A. Ipakchi and F. Albuyeh, "Grid of the Future," in IEEE Power and990
Energy Magazine, vol. 7, no. 2, pp. 52-62, March-April 2009.991
16
[3] Rheinisch-Westfalische Elektrizitatswerke (RWE), Typical Daily Con-992
sumption of Electrical Power in Germany, 2005.993
[4] N. R. Darghouth, G. Barbose, and R. Wiser, "The Impact of Rate994
Design and Net Metering on the Bill Savings from Distributed PV995
for Residential Customers in California," Energy Policy, vol. 39, no. 9,996
pp. 5243-5253, Sep. 2011.997
[5] H. Farhangi, "The Path of the Smart Grid," in IEEE Power and Energy998
Magazine, vol. 8, no. 1, pp. 18-28, January-February 2010.999
[6] S. Massoud Amin and B. F. Wollenberg, "Toward a Smart Grid: Power1000
Delivery for the 21st Century," in IEEE Power and Energy Magazine,1001
vol. 3, no. 5, pp. 34-41, Sept.-Oct. 2005.1002
[7] F. Manzano-Agugliaro, A. Alcayde, F. G. Montoya, A. Zapata-Sierra1003
and C. Gil, "Scientific production of renewable energies worldwide:1004
an overview," in Renewable and Sustainable Energy Reviews, vol. 18,1005
pp.134-143, 2013.1006
[8] N. Hatziargyriou, H. Asano, R. Iravani and C. Marnay, "Microgrids," in1007
IEEE Power and Energy Magazine, vol. 5, no. 4, pp. 78-94, July-Aug.1008
2007.1009
[9] M. U. Khalid, N. Javaid, M.N. Iqbal, A. A. Rehman, M. U. Khalid1010
and M. A. Sarwar, "Cooperative Energy Management Using Coalitional1011
Game Theory for Reducing Power Losses in Microgrids," in Conference1012
on Complex, Intelligent, and Software Intensive Systems, vol. 772, pp.1013
317-328, July, 2018.1014
[10] K. R. Apt and A. Witzel, "A Generic Approach to Coalition Forma-1015
tion," in International Game Theory Review, vol. 11, no. 3, pp. 347-367,1016
Sept. 2009.1017
[11] I. S. Bayram, M. Z. Shakir, M. Abdallah and K. Qaraqe, "A Survey1018
on Energy Trading in Smart Grid," 2014 IEEE Global Conference on1019
Signal and Information Processing (GlobalSIP), Atlanta, GA, pp. 258-1020
262, 2014.1021
[12] C. Wei, Z. M. Fadlullah, N. Kato and A. Takeuchi, "GT-CFS: A1022
Game Theoretic Coalition Formulation Strategy for Reducing Power1023
Loss in Micro Grids," in IEEE Transactions on Parallel and Distributed1024
Systems, vol. 25, no. 9, pp. 2307-2317, Sept. 2014.1025
[13] A. Clausen, G. Koenig, S. Klingert, G. Ghatikar, P. M. Schwartz and1026
N. Bates, "An analysis of contracts and relationships between super-1027
computing centers and electricity service providers." In Proceedings of1028
the 48th International Conference on Parallel Processing: Workshops,1029
pp. 1-8. 2019.1030
[14] N. Bates, G. Ghatikar, G. Abdulla, G. A. Koenig, S. Bhalachandra,1031
M. Sheikhalishahi, T. Patki, B. Rountree and S. Poole, "Electrical1032
grid and supercomputing centers: An investigative analysis of emerging1033
opportunities and challenges." Informatik-Spektrum 38, no. 2, 111-127,1034
2015.1035
[15] J. J. Palop, L. Mucke and E. D. Roberson, "Quantifying biomarkers of1036
cognitive dysfunction and neuronal network hyperexcitability in mouse1037
models of Alzheimer’s disease: depletion of calcium-dependent proteins1038
and inhibitory hippocampal remodeling." In Alzheimer’s Disease and1039
Frontotemporal Dementia, pp. 245-262. Humana Press, Totowa, NJ,1040
2010.1041
[16] N. G. Paterakis, O. Erdinc, I. N. Pappi, A. G. Bakirtzis and J. P. S.1042
Catalao, "Coordinated Operation of a Neighborhood of Smart House-1043
holds Comprising Electric Vehicles, Energy Storage and Distributed1044
Generation," in IEEE Transactions on Smart Grid, vol. 7, no. 6, pp.1045
2736-2747, Nov. 2016.1046
[17] B. Hussain, Q. U. Hasan, N. Javaid, M. Guizani, A. Almogren and A.1047
Alamri, "An innovative heuristic algorithm for IoT-enabled smart homes1048
for developing countries," IEEE Access, 6, pp.15550-15575, 2018.1049
[18] A. Ahmad, A. Khan, N. Javaid., H. M. Hussain, W. Abdul, A.1050
Almogren, A. Alamri and I. A. Niaz, "An optimized home energy1051
management system with integrated renewable energy and storage1052
resources," Energies 10, no. 4, 549, 2017.1053
[19] F. Mangiatordi, E. Pallotti, D. Panzieri and L. Capodiferro, "Multi1054
Agent System for Cooperative Energy Management in Microgrids,"1055
2016 IEEE 16th International Conference on Environment and Elec-1056
trical Engineering (EEEIC), Florence, 2016, pp. 1-5.1057
[20] A. Anvari-Moghaddam, A. Rahimi-Kian, M. S. Mirian and J. M. Guer-1058
rero, "A multi-agent based energy management solution for integrated1059
buildings and microgrid system." Applied energy 203, pp. 41-56, 2017.1060
[21] N. Bazmohammadi, A. Anvari-Moghaddam, A. Tahsiri, A. Madary,1061
J. C. Vasquez and J. M. Guerrero, "Stochastic predictive energy1062
management of multi-microgrid systems." Applied Sciences 10, no. 14,1063
4833, 2020.1064
[22] K. Rahbar, C. C. Chai and R. Zhang, "Energy Cooperation Opti-1065
mization in Microgrids With Renewable Energy Integration," in IEEE1066
Transactions on Smart Grid, vol. 9, no. 2, pp. 1482-1493, March 2018.1067
[23] X. Xing, L. Xie and H. Meng, "Cooperative energy management1068
optimization based on distributed MPC in grid-connected microgrids1069
community." International Journal of Electrical Power & Energy Sys-1070
tems 107, pp. 186-199, 2019.1071
[24] A. C. Luna, N. L. Diaz, M. Graells, J. C. Vasquez and J. M.1072
Guerrero, "Cooperative energy management for a cluster of households1073
prosumers." IEEE transactions on consumer electronics 62, no. 3, pp.1074
235-242, 2016.1075
[25] D. Arcos-Aviles, G. Garcia-Gutierrez, F. Guinjoan, P. Ayala, A. Ibarra,1076
E. Motoasca, J. Llanos and J. Pascual, "Fuzzy-based power exchange1077
management between grid-tied interconnected residential microgrids."1078
In 2020 IEEE ANDESCON, pp. 1-7. IEEE, 2020.1079
[26] D. Arcos-Aviles, F. Guinjoan, J. Pascual, L. Marroyo, P. Sanchis,1080
R. Gordillo, P. Ayala and M. P. Marietta, "A review of fuzzy-based1081
residential grid-connected microgrid energy management strategies for1082
grid power profile smoothing." Energy Sustainability in Built and Urban1083
Environments, pp. 165-199, 2019.1084
[27] Y. Wang, W. Saad, Z. Han, H. V. Poor and T. Basar, "A Game-1085
Theoretic Approach to Energy Trading in the Smart Grid," in IEEE1086
Transactions on Smart Grid, vol. 5, no. 3, pp. 1439-1450, May 2014.1087
[28] S. Chakraborty, S. Nakamura and T. Okabe, "Real-time Energy Ex-1088
change Strategy of Optimally Cooperative Microgrids for Scale-flexible1089
Distribution System," in Expert Systems with Applications vol. 42, no.1090
10, pp. 4643-4652, 2015.1091
[29] N. Bazmohammadi, A. Tahsiri, A. Anvari-Moghaddam and J. M.1092
Guerrero, "A hierarchical energy management strategy for intercon-1093
nected microgrids considering uncertainty." International Journal of1094
Electrical Power & Energy Systems, 109, pp. 597-608. 2019.1095
[30] G. van Leeuwen, T. AlSkaif, M. Gibescu and W. van Sark, "An1096
integrated blockchain-based energy management platform with bilateral1097
trading for microgrid communities." Applied Energy, 263, 114613,1098
2020.1099
[31] Z. Wang, Q. Zhu, M. Huang and B. Yang, "Optimization of Eco-1100
nomic/Environmental Operation Management for Micro-grids by using1101
Hybrid Fireworks Algorithm," in International Transactions on Electri-1102
cal Energy Systems, vol. 27, no. 12, 2017.1103
[32] G. Hafeez, N. Javaid, S. Iqbal and F. A. Khan, "Optimal Residential1104
Load Scheduling Under Utility and Rooftop Photovoltaic Units," in1105
Energies, vol. 11, no. 3, pp. 611-637, 2018.1106
[33] T. AlSkaif, M.G. Zapata, B. Bellalta and A. Nilsson, "A Distributed1107
Power Sharing Framework among Households in Microgrids: A Re-1108
peated Game Approach," in Computing, vol. 99, no. 1, pp. 23-37, Jan.1109
2017.1110
[34] M. Daneshvar, B. Mohammadi-Ivatloo, S. Asadi, A. Anvari-1111
Moghaddam, M. Rasouli, M. Abapour and G. B. Gharehpetian,1112
"Chance-constrained models for transactive energy management of1113
interconnected microgrid clusters." Journal of Cleaner Production, 271,1114
122177, 2020.1115
[35] N. Bazmohammadi, A. Tahsiri, A. Anvari-Moghaddam and J. M.1116
Guerrero, "A hierarchical energy management strategy for intercon-1117
nected microgrids considering uncertainty." International Journal of1118
Electrical Power & Energy Systems, 109, 597-608, 2019.1119
[36] M. Daneshvar, B. Mohammadi-Ivatloo, K. Zare, S. Asadi and1120
A. Anvari-Moghaddam, "A novel operational model for intercon-1121
nected microgrids participation in transactive energy market: A hybrid1122
IGDT/stochastic approach." IEEE Transactions on Industrial Informat-1123
ics 17, no. 6, pp. 4025-4035, 2020.1124
[37] M. Fathi and H. Bevrani, "Statistical Cooperative Power Dispatching1125
in Interconnected Microgrids," in IEEE Transactions on Sustainable1126
Energy, vol. 4, no. 3, pp. 586-593, July 2013.1127
[38] J. Ni and Q. Ai, "Economic Power Transaction using Coalitional1128
Game Strategy in Micro-grids," in IET Generation, Transmission and1129
Distribution, vol. 10, no. 1, pp. 10-18, July. 2016.1130
[39] W. Lee, L. Xiang, R. Schober and V. W. S. Wong, "Direct Electricity1131
Trading in Smart Grid: A Coalitional Game Analysis," in IEEE Journal1132
on Selected Areas in Communications, vol. 32, no. 7, pp. 1398-1411,1133
July 2014.1134
[40] C. Wei, Z. M. Fadlullah, N. Kato and I. Stojmenovic, "On Optimally1135
Reducing Power Loss in Micro-grids With Power Storage Devices," in1136
IEEE Journal on Selected Areas in Communications, vol. 32, no. 7, pp.1137
1361-1370, July 2014.1138
[41] W. P. Lee, J. Y. Choi and D. J. Won, "Coordination Strategy for1139
Optimal Scheduling of Multiple Microgrids Based on Hierarchical1140
System," in Energies, vol. 10, no. 9 pp. 1336-1353, 2017.1141
[42] T. Zhao and Z. Ding, "Cooperative Optimal Control of Battery Energy1142
Storage System Under Wind Uncertainties in a Microgrid," in IEEE1143
17
Transactions on Power Systems, vol. 33, no. 2, pp. 2292-2300, March1144
2018.1145
[43] X. Liu, B. Gao, Z. Zhu and Y. Tang, "Non-cooperative and Coop-1146
erative Optimisation of Battery Energy Storage System for Energy1147
Management in Multi-microgrid," in IET Generation, Transmission &1148
Distribution, vol. 12, no. 10, pp. 2369-2377, 2018.1149
[44] W. Zhang and Y. Xu, "Distributed optimal control for multiple1150
microgrids in a distribution network." IEEE Transactions on Smart Grid,1151
10, no. 4, pp. 3765-3779, 2018.1152
[45] F. Khavari, A. Badri and A. Zangeneh, "Energy management in1153
multi-microgrids considering point of common coupling constraint."1154
International Journal of Electrical Power & Energy Systems, 115,1155
105465, 2020.1156
[46] X. Zhu and K. C. Leung, "An Adaptive Distributed Power Scheduling1157
Algorithm for Renewable Microgrid Cooperation," 2017 IEEE Interna-1158
tional Conference on Communications (ICC), Paris, 2017, pp. 1-6.1159
[47] M. Naz, Z. Iqbal, N. Javaid, Z. A. Khan, W. Abdul, A. Almogren and1160
A. Alamri, "Efficient power scheduling in smart homes using hybrid1161
grey wolf differential evolution optimization technique with real time1162
and critical peak pricing schemes," Energies 11, no. 2, 384, 2018.1163
[48] N. Javaid, F. Ahmed, I. Ullah, S. Abid, W. Abdul, A. Alamri and A.1164
Almogren, "Towards cost and comfort based hybrid optimization for1165
residential load scheduling in a smart grid," Energies 10, no. 10, 1546,1166
2017.1167
[49] W. Shi, N. Li, C. C. Chu and R. Gadh, "Real-Time Energy Manage-1168
ment in Microgrids," in IEEE Transactions on Smart Grid, vol. 8, no.1169
1, pp. 228-238, Jan. 2017.1170
[50] T. Alharbi and K. Bhattacharya, "Optimal Scheduling of Energy1171
Resources and Management of Loads in Isolated/Islanded Microgrids,"1172
in Canadian Journal of Electrical and Computer Engineering, vol. 40,1173
no. 4, pp. 284-294, Fall 2017.1174
[51] J. Llanos, D. E. Olivares, J. W. Simpson-Porco, M. Kazerani and1175
D. Sáez, "A novel distributed control strategy for optimal dispatch1176
of isolated microgrids considering congestion." IEEE Transactions on1177
Smart Grid, 10, no. 6, pp. 6595-6606, 2019.1178
[52] V. H. Bui, A. Hussain and H. M. Kim, "A Multiagent-Based Hierar-1179
chical Energy Management Strategy for Multi-Microgrids Considering1180
Adjustable Power and Demand Response," in IEEE Transactions on1181
Smart Grid, vol. 9, no. 2, pp. 1323-1333, March 2018.1182
Muhammad Usman Khalid completed an MS1183
in electrical engineering with the communication1184
over Sensors (ComSens) Research Laboratory, De-1185
partment of Computer Science, COMSATS Uni-1186
versity Islamabad, Islamabad, Pakistan, under su-1187
pervision of Dr. Nadeem Javaid. He has 6 research1188
publications in well reputed international journals1189
and conferences. His research includes energy1190
management in smart grids, game theory, etc.1191
1192
NADEEM JAVAID (S’8, M’11, SM’16) received1193
the bachelor degree in computer science from1194
Gomal University, Dera Ismail Khan, Pakistan,1195
in 1995, the master degree in electronics from1196
Quaid-i-Azam University, Islamabad, Pakistan, in1197
1999, and the Ph.D. degree from the University1198
of Paris-Est, France, in 2010. He is currently an1199
Associate Professor and the Founding Director1200
of the Communications Over Sensors (ComSens)1201
Research Laboratory, Department of Computer1202
Science, COMSATS University Islamabad, Islam-1203
abad. He is also working as visiting professor at the School of Computer1204
Science, University of Technology Sydney, Sydney, Australia. He has1205
supervised 137 master and 24 Ph.D. theses. He has authored over 9001206
articles in technical journals and international conferences. His research1207
interests include energy optimization in smart grids and in wireless sensor1208
networks using data analytics and blockchain. He was recipient of the1209
Best University Teacher Award from the Higher Education Commission of1210
Pakistan, in 2016, and the Research Productivity Award from the Pakistan1211
Council for Science and Technology, in 2017. He is also Associate Editor1212
of IEEE Access and Editor of Sustainable Cities and Society journals.1213
AHMAD ALMOGREN (SM) received the1214
Ph.D. degree in computer science from Southern1215
Methodist University, Dallas, TX, USA, in 2002.1216
He is currently a Professor with the Computer1217
Science Department, College of Computer and1218
Information Sciences (CCIS), King Saud Univer-1219
sity (KSU), Riyadh, Saudi Arabia, where he is1220
currently the Director of the Cyber Security Chair,1221
CCIS. Previously, he worked as the Vice Dean of1222
the Development and Quality at CCIS. He also1223
served as the Dean for the College of Computer1224
and Information Sciences and the Head of the Academic Accreditation1225
Council, Al Yamamah University. He served as the General Chair for1226
the IEEE Smart World Symposium and a Technical Program Committee1227
member of numerous international conferences/workshops, such as IEEE1228
CCNC, ACM BodyNets and IEEE HPCC. His research interests include1229
mobile-pervasive computing and cyber security.1230
ABRAR AHMED received the B.S. degree in1231
computer engineering from COMSATS University1232
Islamabad, (Formerly, COMSATS Institute of In-1233
formation Technology) Pakistan, in 2006, the M.S.1234
degree from Lancaster University, U.K., in 20081235
and the Ph.D. degree in electrical engineering from1236
COMSATS University Islamabad, in 2017. Since1237
2006, he has been associated with COMSATS1238
University Islamabad, where he currently holds the1239
position of an Assistant Professor. His research1240
interests include wireless channel modelling and1241
characterizing, smart antenna systems, nonorthogonal multiple access tech-1242
niques and adaptive signal processing.1243
SARDAR MUHAMMAD GULFAM received1244
his master’s degree from Tampere University of1245
Technology Finland in 2010. He received his1246
PhD degree in 2017 from COMSATS University1247
Islamabad. Currently, he is working as an As-1248
sistant professor in the Department of Electric1249
and Computer Engineering, COMSATS University1250
Islamabad, Islamabad Pakistan.1251
1252
Ayman Radwan received his Ph.D. degree from1253
Queen’s University, Kingston, ON, Canada, in1254
2009. He is a Senior Research Engineer and an1255
Assistant Professor (Investigador Auxiliar) with1256
the Instituto de Telecomunicações and University1257
of Aveiro, Aveiro, Portugal. He is mainly spe-1258
cialized in coordination and management of EU1259
funded projects. He participated in the coordi-1260
nation of multiple EU projects. He is currently1261
the Project Coordinator of the CELTIC+ Project1262
“SAFE-HOME”, as well as participating in multi-1263
ple other EU projects. He has also been the Technical Manager of the FP7-1264
C2POWER Project and the Coordinator of the CELTIC projects “GreenT”1265
and “MUSCLES”. His current research interests include the Internet of1266
Things, 5G, 6G and radio resource management.1267
Article
Full-text available
Electric power industry is continually adopting new techniques to improve the reliability and efficiency of the energy system and to cope with the increasing energy demand and the associated technical challenges. In recent years, the maturation of Artificial Intelligence (AI) led researchers to solve various problems in the power system by using AI techniques. Voltage Source Converter is the result of advancements in the field of power electronics and semiconductors technology, which holds a promising future for the realization of smart grid, renewable energy integration, and HVDC transmission system. Usually hit and trial method or the design engineer's experience is used for the manual tuning of the PI controllers, which cannot yield superior performance. The process becomes more complicated when multiple grids are involved, such as in VSC-based MTDC grids. This research article use a deep learning optimization technique for the tuning of the VSC controllers, which resulted in quick settling time, better slew rate, less undershoot and low overshoot. The deep learning neural network is trained through the Particle Swarm Optimization (PSO) algorithm to produce the best possible tuned or optimally tuned parameters for the controllers. The optimal tuning of the controller will result in an overall better performance of the converter and the grid. A four-layered deep learning neural network and a three-terminal MTDC grid were designed and simulated in MATLAB/SIMULINK environment.
Article
Full-text available
Distributed power generation technologies have been, over the years, gaining more and more attention from the industry as well as from the academia, due to the global interest in the exploitation of renewable energy sources even in urban areas. From this perspective, small size wind turbines, often with vertical axis layout and a nominal power between 1-3 kW, are being installed in urban areas on rooftops of residential, commercial, industrial buildings and even on metallic carpentry towers. As wind turbines tend to produce noise and vibrations, especially in presence of vortexes, wakes and unsteadiness, efficient devices for mitigating the vibrations transmitted to the supporting structure have to be added in order to limit the generation of stresses. Actually, the social acceptance of the installation of small wind turbines in built environment is tightly connected to the acoustic comfort for the inhabitants. The present study has been developed thanks to a joint collaboration between the Department of Engineering of the University of Perugia (Italy) and FH Technikum Wien (Austria) within the participation to the IEA Wind-Task 41: “Distributed Wind”. The work aims at mechanically characterizing the vibration damping provided by a mechanical device installed on the top of the turbine tower support. The decoupler, specifically designed for this purpose, has been tested in different conditions (in field and laboratory) and with different approaches, in order to have an overall evaluation of its performance. At first, vibrations are measured in open field in a real test case scenario with two installation arrangements: on the rooftop of a small building and on a tubular tower. Subsequently, the decoupler has been removed from its original location and tested in laboratory with external excitation from shaker.
Article
Full-text available
Next-generation power systems will require innovative control strategies to exploit existing and potential capabilities of developing renewable-based microgrids. Cooperation of interconnected microgrids has been introduced recently as a promising solution to improve the operational and economic performance of distribution networks. In this paper, a hierarchical control structure is proposed for the integrated operation management of a multi-microgrid system. A central energy management entity at the highest control level is responsible for designing a reference trajectory for exchanging power between the multi-microgrid system and the main grid. At the second level, the local energy management system of individual microgrids adopts a two-stage stochastic model predictive control strategy to manage the local operation by following the scheduled power trajectories. An optimal solution strategy is then applied to the local controllers as operating set-points to be implemented in the system. To distribute the penalty costs resulted from any real-time power deviation systematically and fairly, a novel methodology based on the line flow sensitivity factors is proposed. Simulation and experimental analyses are carried out to evaluate the effectiveness of the proposed approach. According to the simulation results, by adopting the proposed operation management strategy, a reduction of about 47% in the average unplanned daily power exchange of the multi-microgrid system with the main grid can be achieved.
Article
Full-text available
In this paper, an integrated blockchain-based energy management platform is proposed that optimizes energy flows in a microgrid whilst implementing a bilateral trading mechanism. Physical constraints in the microgrid are respected by formulating an Optimal Power Flow (OPF) problem, which is combined with a bilateral trading mechanism in a single optimization problem. The Alternating Direction Method of Multipliers (ADMM) is used to decompose the problem to enable distributed optimization and a smart contract is used as a virtual aggregator. This eliminates the need for a third-party coordinating entity. The smart contract fulfills several functions, including distribution of data to all participants and executing part of the ADMM algorithm. The model is run using actual data from a prosumer community in Amsterdam and several scenarios of the model are tested to evaluate the impact of combining physical constraints and trading on social welfare of the community and scheduling of energy flows. The scenario variants are trade-only, where only a trading mechanism is implemented , grid-only where only OPF optimization is implemented and a combined scenario where both are implemented. Results are compared with a baseline scenario. Simulation results show that import costs of the whole community are reduced by 34.9% as compared to a baseline scenario, and total energy import quantities are reduced by 15%. Total social welfare is found to be highest without a trading mechanism, however this platform is only viable when all costs are equally shared between all households. Furthermore, peak imports are reduced by over 50% in scenarios including grid constraints.
Article
Full-text available
This work presents a novel distributed control strategy for frequency control, congestion management, and optimal dispatch (OD) in isolated microgrids. The proposed strategy drives the distributed generators (DGs) within the microgrid to a dispatch that complies with the Karush-Kuhn-Tucker (KKT) conditions of a linear optimal power flow (OPF) formulation. The controller relies on local power and frequency measurements, information from neighbouring DGs, and line-flow measurements transmitted through a communications network. Extensive simulations show a good performance of the controller against sudden changes in the load, congested lines and availability of DGs in the microgrid, being able to successfully drive the system to an optimal economic operation.
Conference Paper
This paper presents a fuzzy-based power exchange management between two neighboring residential grid-connected microgrids comprising both photovoltaic generation and battery energy storage system (BESS). The proposed power exchange management accounts for the magnitude of the energy rate-of-change of each microgrid and the charge difference between the BESSs of both microgrids to charge the ESS that has an energy deficit. As such, the proposed power exchange management can reduce the amount of power absorbed from the mains of each microgrid by operating jointly with each other rather than separately, and it also synchronizes the ESS of both microgrids, improving the behavior of ESSs. A comparison of the simulated results for a scenario with and without power exchange is presented in order to demonstrate the adequate behavior of the proposed power exchange management.
Article
Recently, the decarbonization of the electric power system has led to substantial efforts for designing a pathway toward 100% renewable energy resources (RERs). This paper proposes a novel operational model for the effective participation of the interconnected microgrids with 100% RERs in the transactive energy market. The novelty of the proposed model is mostly related to the use of transactive energy technology for developing the free energy trading environment for the microgrids with 100% RERs as the local energy trading market to establish a dynamic energy balance in the system. To capture the intermittencies in the system, a hybrid IGDT/stochastic method is proposed in the deregulated environment. The proposed model is validated by selecting the modified IEEE 14-bus test system. The results indicate the effectiveness of the proposed model in providing the same percentage of cost-saving for microgrids when they simultaneously participate in the transaction energy market.
Article
Transactive energy as an emerging approach and sustainable technology can provide an exceptional opportunity for microgrids to exchange energy with each other for greater benefits in the cluster mode. In this mode of operation, some collective and individual interests can be realized for the microgrids based on transactive energy management. This paper proposes mathematical models for microgrid clusters using a transactive energy structure to manage energy exchange in the smart grid. In order to make an informed decision for the operation of microgrid clusters, chance-constrained programming is employed to consider the uncertainties in balancing collective and individual interests under the transactive energy management. In this research, sixteen commercial microgrids are considered in the process of evaluating the efficiency of the proposed models using the chance-constrained programming method. Simulation results prove the effectiveness of the transactive energy approach accompanying the implementation of chance-constrained programming in energy management of the microgrid clusters.
Article
There are different models for Energy Management of Multi-Microgrids (MMGs). Generally, the owners of microgrids are not identical. In this case, due to privacy concerns and overcomes drawbacks of conventional decentralized systems, hybrid energy management system is proposed. Unlike other energy management models, in hybrid model, multi-microgrids are connected to the grid through the common line entitled Point of Common Coupling (PCC). Energy management in hybrid multi-microgrids considering optimal utilization of PCC capacity is a critical issue that has been less taken into account. In this paper, a new bi-level method is proposed for optimal energy management in hybrid MMG systems taking into account the PCC line capacity. In the first level, each microgrid implements its day-ahead scheduling based on different quantities of PCC line capacity and extracts its profit-quantity curve. This novel curve shows the variations of microgrid profits versus different PCC limits. Subsequently, a two-stage optimization problem is presented in the second level, in which in the first stage an introduced microgrid aggregator (MGA) maximizes microgrids aggregated profit and determines optimal quota of each microgrid from PCC line based on corresponding profit-quantity curves while in the second stage, this profit is fairly divided among microgrids via Shapely value. Numerical results demonstrate the efficiency and reliability of the proposed method.
Conference Paper
Increases in peak electricity demands and the growing use of renewable energy --- with associated intermittency and variable output --- present new challenges to electricity service providers (ESPs). ESPs employ demand charges, variable tariffs and demand response (DR) programs to influence the consumption behavior of consumers to partially mitigate these challenges. Due to their high load and potential flexibility, supercomputing centers (SCs) are increasingly gaining importance in the grid. This paper presents a qualitative study of service contracts between ESPs and SCs in the United States and Europe. From this we extract a contract typology used to understand how, and to what extent, variable tariffs, DR programs, and demand charges are imposed on SCs. Further, we highlight the actions taken by SCs in response to these contractual elements. Finally, we present perspectives on grid integration of SCs to enhance their collaboration with their ESPs to benefit supply stability and resilience.
Article
With the fast development of microgrids (MGs), the microgrid community (MGC) integrating adjacent MGs has raised more and more attentions. In an MGC, each MG shares available power reciprocally to minimize the operation costs and maintain power balance. As the amount of MGs grows, the MGC becomes complicated and difficult to control. Different from centralized strategy, this paper proposes a cooperative energy management scheme based on distributed model predictive control (DMPC) for grid-connected MGC. The normal MGC is virtualized as two-level structure to simplify the internal interactions among subsystems. All the subsystems iterate in parallel adopting the proposed cooperative logarithmic-barrier method, which follows a communication protocol to derive control signals. The real-time hardware-in-the-loop (HIL) study verifies that, the optimization results of proposed DMPC approximate to Pareto solutions of centralized MPC (CMPC), whereas the computing speed is much faster than CMPC.