IoT based energy management in smart energy system: A hybrid SO2SA technique

Abstract

In this manuscript, an energy management system (EMS) is proposed to the distribution system (DS) using Internet of Things (IoT) framework with a hybrid system. The proposed hybrid method is the combination of the Seagull Optimization Algorithm (SOA) and Owl Search Algorithm (OSA), hence it is called SO2SA technique. The principle objective of the SO2SA technique is to optimize managing distribution system power and resources through continuous monitoring of the data from a communication framework based on IoT. In SO2SA technique, every home device is connected to the module of data acquisition, which indicates an IoT object along with a unique IP address as a result of huge mesh wireless network devices. The sending data are processed through SO2SA technique. Similarly, the IoT architecture of the distribution system enhances the flexibility of these networks and gives optimal utilization of obtainable resources. In addition, the SO2SA technique is responsible for meeting the overall power and supply requirements. The proposed method is implemented in MATLAB/Simulink site and the efficiency is likened to the other different methods. In 50 trail numbers, the RMSE, MAPE, and MBE range of SO2SA technique represents 5.63, 0.90, and 1.035. Thus, the proposed technique is highly competent over all the existing approaches.

Full text

RESEARCH ARTICLE
IoT based energy management in smart energy system:
A hybrid SO
2
SA technique
Sankaramoorthy Muthubalaji
1
| Sundararajan Srinivasan
1
|
Muthuramalingam Lakshmanan
2
1
Department of EEE, CMR College of
Engineering & Technology, Hyderabad,
India
2
Department of Electrical and Electronics
Engineering, CMR Institute of
Technology, Bengaluru, India
Correspondence
Sankaramoorthy Muthubalaji,
Department of EEE, CMR College of
Engineering & Technology, Kandlakoya,
Hyderabad, Telangana, India.
Email: muthusa15@gmail.com
Abstract
In this manuscript, an energy management system (EMS) is proposed to the
distribution system (DS) using Internet of Things (IoT) framework with a
hybrid system. The proposed hybrid method is the combination of the Seagull
Optimization Algorithm (SOA) and Owl Search Algorithm (OSA), hence it is
called SO
2
SA technique. The principle objective of the SO
2
SA technique is to
optimize managing distribution system power and resources through continu-
ous monitoring of the data from a communication framework based on IoT. In
SO
2
SA technique, every home device is connected to the module of data acqui-
sition, which indicates an IoT object along with a unique IP address as a result
of huge mesh wireless network devices. The sending data are processed
through SO
2
SA technique. Similarly, the IoT architecture of the distribution
system enhances the flexibility of these networks and gives optimal utilization
of obtainable resources. In addition, the SO
2
SA technique is responsible for
meeting the overall power and supply requirements. The proposed method is
implemented in MATLAB/Simulink site and the efficiency is likened to the
other different methods. In 50 trail numbers, the RMSE, MAPE, and MBE
range of SO
2
SA technique represents 5.63, 0.90, and 1.035. Thus, the proposed
technique is highly competent over all the existing approaches.
KEYWORDS
demand response, distribution system, fuel cell, main grid, Owl search algorithm,
photovoltaic, seagull optimization algorithm, wind turbine
List of Symbols and Abbreviations: p
i
, profit of each item; x
i
, binary number 1 and 0 represents ON/OFF state of every device; ω
i
, weight of each
item; γ
i
(t), ON/OFF state of the appliance; τ
sch
, start and end times of the appliance; τ
ot
, maximum allowable power to be consumed; P
pv
, output
power of PV panels; R
t
, solar radiation on a tilted plane module; η
pv
, performance of the PV panels; A
pv
, area of PV panel; WS, wind speed; WS
CI
,
wind cut-in speed; WS
R
, rated wind speed; WS
CO
, wind cut-off speed; P
r
, rated power; R
th
, thermal ratio for electrical energy; p
a
, auxiliary power
device; C
ng
, cost of natural gas; C
pump
, cost of hydrogen pumping; Pchp
Hi , power produced from CHP unit iat time H;Hchp
Hi , heat produced by the CHP
unit iat time H;A,B,Cand D, marginal points; E,F, marginal points of second type in CHP; x
1
H,x
2
H, operational state of CHP; A,B,C,D,E, and F,
CHP cost function coefficients; DR, randomly generated demand response; Y
ij
, random solutions; V, prey location that reached with fittest owl; O
i
,ith
owl; R
i
, each owl and prey distance information; p
vm
, probability of vole movement; α, uniformly distributed random number; β, linearly decreasing
constant; I
i
, intensity information of ith owl; UH
h
, charging and discharging of binary variables.
Received: 20 February 2020 Revised: 7 April 2021 Accepted: 13 April 2021
DOI: 10.1002/jnm.2893
Int J Numer Model. 2021;e2893. wileyonlinelibrary.com/journal/jnm © 2021 John Wiley & Sons Ltd. 1of26
https://doi.org/10.1002/jnm.2893

1|INTRODUCTION
Increasing power around the world integrate along required energy safety distribution has led to the continuous effort
to transition from a grid of traditional power generation to a smart flexible and network of multiple power sources that
adds renewable energy sources (RES).
1
A more complex environment has emerged as both demand and supply change
over time, which can be accessed as a networking system that interconnects processes along with multiple variables
that need to handle real-time data gathered from device, system, and process.
2
Hence, a multiple-source energy net-
work's design and operation includes several decisions over numerous levels.
3-5
The network classification of power
along unique challenges and requirements is automated off-grid networks, wherever the important power source is
RES.
6-9
Using intermediate storage systems, the nature of RES intermittent may safely be managed, such as accumula-
tors (battery stacks).
10-12
Instead of the generation of hydrogen, by fuel cell, the storage and usage may require to serve
continuous power availability. Such kinds of networks may be used to the locations of off-grid where fuel transportation
is complex, including fewer communities wherever autonomy is significant.
13
Efficient use of energy in the distribution system saves money, improves sustainability, and decreases the footprint
of carbon greatly. As a result, the demand for the management of smart energy is increasing for the distribution system
and commonly for smart cities. Low cost, simple to array, less maintenance technology consists of a slightly limited
array of large-scale on these systems. The volume of data gathered in various cities in the country poses many chal-
lenges in data storage, organization, and analysis.
14
The interrelation in multiple technologies, diverse design controls,
need of an operation, and the result of communication may be realized by the new synergetic paradigm Internet of
Things (IoT).
15-17
IoT refers to three important factors such as (1) network, (2) things and semantics toward mid-
dleware, and (3) systems (from sensors to systems) and knowledge.
To meet these types of challenges, IoT technology is a natural candidate. To realize, monitor, and regulate the con-
sumption of energy of home appliances on a large scale the IoT technologies may offer a ubiquitous computing plat-
form. Utilizing several different types of wireless sensors established on residential units this data is gathered.
18-21
Different types of energy management systems (EMS), integrated energy systems, and customer equipment may pro-
duce several data formats, which do not generate interaction and interoperability at the system of industrial energy
management. The main purpose of EMS integration is to permit communication entities to communicate with every-
one utilizing a general information model. It makes it necessary for understanding and mapping the meaning and con-
text of information that lives in various domains.
22,23
At distribution systems, a general model of information is
required to support the capacity for aligning the information semantically to inter-domain communication for EMS.
24
1.1 |Recent research work: A brief review
Due to maximal power consumption at smart home utilizations is essential for installing a distribution system-centric
management of energy and ensures that the system was used correctly. Over the previous decade, academic efforts have
moved toward maximizing the efficiency of energy. Here, a part of the work is reviewed.
Faheem et al
25
have suggested a new method to deal with the power quality and reliability problems for the supplier
of service and consumers. The major objectives of the smart grid in smart cities were protecting the standard of living
few numbers of residents and support the full spectrum of its economic processes. The new energy-efficient, as well as
Reliable Data Gathering Routing Protocol to wireless sensor networks (WSNs)- based smart grid usage, was introduced.
Liu et al
26
have suggested IoT-based EMS design in terms of infrastructure of edge computing along with deep rein-
forcement learning. At first, IoT-based EMS in smart cities was demonstrated. Second, the model of architecture and
software of EMS based on IoT along edge computing were presented. Subsequently, an efficient energy planning pro-
gram along with deep reinforcement learning to present the framework was presented.
Yassine et al
27
have presented from smart homes a novel site, which activates the novelty examines on IoT captured
data. To permit the services of data-driven, the fog nodes with cloud system overcome the complications challenge
together with resource requirement for data processing of classification analysis, online, storage, and offline.
Li et al
28
have introduced a system of energy detection named SETS in terms of machine learning and statistic
models. The first three stages of decision-making were the predictive model that uses a multiple-model prediction
method. The model combined different machine learning models addicted to a particular predictive method to predict
the consumption of power. The primary decision model was the second level used by the simple moving average to
determine the abnormality. The secondary decision-making model was the third level it deals with the final stage of the
2of26 MUTHUBALAJI ET AL.

decision regarding energy theft. Tom et al. have introduced the use of consumer behavior based on historical data and
time-series analysis predicts the need for energy for individual consumers every hour for the next 72 h.
29
Furthermore,
the work statistically analyzes the use of devices in each household, thereby identifying the role that appliances play
during peak use. The work can help applications know how their customers utilize electricity and promote consumers
to move to the use of peak-time devices for peak hours. Furthermore, consumers could individual equipment grant con-
trol for applications to reduce the burden at peak times and reduce demand. Ahuja and Khosla
30
have examined the
framework of data analysis in smart energy meters using recent data processing methods/tools with a gamification
method to improve customer engagement. Ahuja and Khosla
30
have postulated tools of data analytic and combined
data analyzing methods utilized on smart energy meters (SEM). And also, when looking at different techniques and
structures of SEM data analysis, the authors provided a new structure for SEM by employing a gamification method to
improve consumer engagement in energy-conserving and improving performance.
Sureshkumar and Ponnusamy
31
have suggested an effective method for power flow management (PFM) in hybrid
renewable energy systems (HRES)-related SG system. Here, the suggested method combines modified elephant herding
optimization algorithm (MEHO) and Tabu search algorithm (TSA), hence it was named as MEHOTSA. Here, the
MEHO plays an evaluation procedure for establishing proper control signals of the system and generates a database of
the offline control signal in light of the power type among the source and load side. Roy et al
32
have described the
scheme of power flow management to the microgrid (MG)-connected system using a hybrid approach. The suggested
hybrid approach was the combined implementation of adaptive neuro fuzzy interference system (ANFIS) and advanced
salp swarm optimization algorithm (ASOA), hence it was named ANFASO technique. In the suggested technique,
ANFIS was used to continuously monitor the system of MG-connected load requirement. ASOA improves the exact
combination of MG based on predicted load requirement.
Durairasan and Balasubramanian
33
have introduced a hybrid method for PFM of MG system connected to HRES.
The introduced method was the consolidation of squirrel search algorithm (SSA) and whale optimization algorithm
(WOA), hence it was named SSAWO. SSA has improved the control signal of voltage source inverter (VSI) under the
power replaces difference among source and load side. WOA procedures guarantee the detection of online control sig-
nals using parallel implementation besides active and reactive power types.
1.1.1 |Background of research work
Energy is a vital part of households, industries, and agriculture. The energy-efficient is maintained and preserving it
intelligently for appliances is important. Energy usage is directly affected by coal, oil, and power generation. On
reviewing previous literature relating to the present subject of investigation, primarily the need for a revolution in clean
energy has been greatly implicated. It is also understood that power from photovoltaic (PV) and wind are disproportion-
ately produced resulting from heavy demand, line losses, and environmental factors. Renewable-based power produc-
tion is the most indispensable means to meet out the increasing load demand. In India, solar and wind energy are
available in plenty due to its location near to the tropic of cancer. But the percentage of power generated from these
sources is fluctuating due to variations in environmental influencing factors. Having understood the core perspectives
relating to renewable power production following are some of the major shortcomings and challenges in concise. Fore-
most, though solar energy is available in plenty, solar panels are less reliable due to weather conditions such as temper-
ature, relative humidity, irradiation, shading, soiling, panel materials, fill factor, elevation, and icing which causes
difficulty in forecasting its accessibility. Secondly, the land is a secret reserve in India and as per capital availability, it is
low. Large land area is required for installation which is sometimes not feasible. Another serious problem is the
increase in the cost of the project due to the storage of excess power. As a result, energy production from solar is quite
low compared to other forms of energy. With 28 regard to wind energy, due to limitations in grid infrastructure, there
is difficulty in transmitting energy produced from wind farms to consumers leading to wastage of energy. Finally,
there are fluctuations in solar and wind energy sources due to weather patterns and cycles of day and night. Besides
these challenges, in most of the previous research work, either wind energy or solar-based power production was
predicted. If both wind and solar were considered, then EMS was implemented by considering one or the minimum
number of the influencing environmental factors for power production. Further, many of the systems were not using
WSN for system integration and EMS. Also, the existing method is regulated by infra-red (IR) technology. But it has
few drawbacks, like IR acquire only 180radiations. Nevertheless, it transmits and receives signals but it does not sup-
port the network. Many analytical algorithms are present that may process large data volume, but most of these are not
MUTHUBALAJI ET AL.3of26

able to finish tasks enough. Hence, to alleviate the mentioned issues, the IoT-based smart control method is developed.
The above-mentioned drawbacks are motivated to do this research work.
1.2 |Objectives and contribution of proposed work
1. An EMS is proposed for distribution system (DS) with IoT framework using a hybrid approach. The proposed hybrid
method is the combination of the Seagull Optimization Algorithm (SOA) and Owl Search Algorithm (OSA), hence it
is called SO
2
SA technique. The main purpose of SO
2
SA method is optimal managing the DS power and resources
through continuous monitoring data as a communication framework based on IoT. The performance of SO
2
SA
method is implemented at MATLAB/Simulink working platform and implementation is evaluated to existing
methods.
2. The innovation of the proposed system is the SOA uses the searching behavior of OSA. Therefore the proposed sys-
tem assumes the optimal better phase, while disregarding the best individual phase.
3. In the proposed technique, every home device is connected to the module of data acquisition, which is an object of
IoT along with a single IP address that goes out in a large network of mesh wireless devices. Here, IoT-based com-
munication framework is used to ease the demand response (DR) development EMS to DS. This framework collects
DR as every device in every smart home and data sends to the centralized server. The sending data are processed by
SO
2
SA technique. Similarly, the IoT architecture of DS enhances the flexibility of these networks and presents opti-
mal utilization of obtainable resources. In addition, the SO
2
SA technique is responsible for meeting the total distri-
bution and energy requirement.
4. The proposed technique function is performed by the SO
2
SA technique. Here, the generated output demand
response, power generated with source, cost function, and limits of associated generation are input and randomly
generate demand response. The main goal is achieved by maximizing the profit with reduced cost. In this way, after
completion of the process, the proposed system is ready to give optimal energy management in DS with IoT
framework.
This manuscript is deliberated as: Section 1 describes the introduction, the literature assessment and its background.
Section 2 describes the overview of IoT-based energy management in smart cities. Section 3 explains the mathematical
modeling of sources. Section 4 illustrates the execution of the proposed technique. Section 5 demonstrates the simula-
tion results and discussion. Finally, Section 6 concludes the manuscript.
2|OVERVIEW OF IOT-BASED ENERGY MANAGEMENT IN SMART CITIES
An IoT is the main component and generates the combination of various sources of energy in smart cities.
34
To moni-
tor, control and save energy, an important number of associations for locating shrewd strategies for controlling the
growing cost and energy demand. The EMS contains a lower cost of resources when meeting the required energy
requirement. At large private, enterprise, modern sections, optimum management of EC is realized by increasing
advances of the IoT with a large amount of information. This manuscript proposes an EMS for DS with IoT configura-
tion through a hybrid system to smart houses. In the proposed method, the data acquisition module interfaces with
every home appliance in an IoT configuration along with a single IP address, providing a huge wireless network of
mesh devices. For further dealing and examining, the EC data collected through the data acquisition module of every
smart house device is sent to the centralized server. From each local location, this information is gathered in huge data
on the application server. To handle the energy consumption in superior way, the proposed EMS portrays that FORDF's
hybrid approach for reducing energy cost through maximizing profit. Figure 1A depicts the IoT-based EMS. Figure 1B
depicts the overall proposed method process.
The residential area has many household devices like charged for water and electricity. Here, an EMS is proposed to
dispatch home smart devices to minimize the electricity bills and maximize user comfort (profit).
35
The smart home
includes smart devices, the proposed controller, and Cloud IoT (CIoT) resources, storage, and data acquisition modules,
where CIoT performs a bridge amid the utilities and enterprise consumer. The device sends its energy consumption
mode to the proposed controller, which dispatches them based on price signals sent by utilities. Cloud IoT prices are ris-
ing send a signal from the utility and send it to the proposed controller. At the same time, it needs energy to consume
4of26 MUTHUBALAJI ET AL.

the data from the proposed controller then send it to the utility. Utility and CIoT communicate with each other via
wireless networks, like Z-wave, ZigBee, or wired protocol. Zigbee is defined as a wireless protocol that works in a mesh
network, it utilizes a device to relay a signal to other devices, reinforcement, and optimizing the network. Zigbee is built
in dimmers, door locks, thermostats, and so on. Like Zigbee, Z-Wave is an open-source mesh network protocol. The
major difference is the data throughput—Z-wave has been approximately six times slower than Zigbee. Nevertheless, it
needs lesser energy for enveloping a similar range as Zigbee. The Smart Things and Lowes Iris are utilized Z-Wave.
Appliances, CIoT including proposed controller exchange information via the home area network. In this manuscript,
smart homes are considered, each of them is fitted with several electrical appliances. In the proposed method, the deci-
sions are carried out within 24 h, 1 week, and 1 year. Dividing a day into 24-time intervals can allocate 1 h to the
device, which only takes few minutes to operate, and since the scheduler is not turn on devices other than the selected
device during that specific time interval, the rest of the time is wasted. Therefore, the electricity bills are reduced by
selected time interval, also creates a more robust system. To perform scheduling, smart devices are divided into various
categories according to the nature of their power consumption modes.
2.1 |Smart buildings
It consists of knowledge and humanity. The EMS in the smart building is characterized by three levels:
1. Level of device
2. Level of system
3. Inter level system
FIGURE 1 (A) IoT-based energy management system. (B) Overall proposed system process
MUTHUBALAJI ET AL.5of26

2.1.1 |Level of device
This is used to evaluate the input data and it is related to the operating conditions for the awareness of the situation in
smart buildings.
36
In the EMS, the processing unit of IoT is used for processing the energy data in a predictive way and
output is sent to appropriate commands including control units. Thus, the needed actions are carried out by the final
actuators.
2.1.2 |Level of system
At the level of system, behavior on different energy-related components is synchronized and coordinated in the smart
building. To manage and control the energy units with the help of short term IoT system the smart building can auto-
matically re-configure the internal basis on buildings. Most of the intelligent IoT system also gives its support to the
energy units.
37
FIGURE 1 (Continued)
6of26 MUTHUBALAJI ET AL.

2.1.3 |Inter level system
The self-contained subsystems of several autonomous energy management filled to smart buildings are required at inter
level system. Like electricity, gas, heat, and so on are the incorporated system. An inter-system is called a smart build-
ing. For the satisfaction of EMS, this smartness specifies the ability to manage, interconnect, and adapt different assets
and functionalities.
2.2 |Smart grid system
Smart grid system acts as a key component toward renewable energy. In a smart power grid, the IoT does not generate
the combination of RES and transport electrification, it generates the value-added service related to new energy.
Based on the deployment and design of the smart grid, the management of energy transformation is obtained.
38
Smart grid systems can able to enlarge its capacity with its intelligence and data flows. At smart grid system, energy
management with the help of IoT can achieve each city corner and can simulate infrastructure of knowledge grid and
interaction of user friendly. The power management system contains a new development of smart energy plus Internet
with the Internet deep fusion and production of energy transmission storage, consumption, and market. The important
properties are smart appliances, synergistic multi-energy sorts, distributed supply and requirement, flat the system, and
open the transactions.
39
The examples are the generation of smart energy and consumption facilities, such as smart coal
mines, smart PV stations, and so on and the cloud site of Internet-based smart operation, so that remote control optimi-
zation can be realized and develop operational performance and advantage.
2.3 |Multi energy networks
Multiple energy networks are used to develop the total performance and benefits of the energy system at various places
such as big buildings, parks, islands, towns, and so on in IoT. Depending on information technologies and communica-
tion exploitation at IoT networks, multiple energy networks may combine the smart electrical grid, provide heat and
gas grid, to smart cities network traffic to unified energy management.
40
Moreover, utilizing IoT advantages at multiple
energy systems is given as follows,
2.3.1 |Promote deep energy fusion and information infrastructure
Several energy systems depending on the technology of IoT may coordinate construction on grids of energy such as gas,
cool air, heat, electric appliances, and so on. Based on the information the infrastructure includes architecture informa-
tion, storage unit, and so on. The measurement of energy computation, information, and control systems can be inte-
grated efficiently.
41
2.3.2 |Management system of smart energy development
By enhanced IoT networks growth, terminal facilities, adding smart homes, smart districts, and so on may construct
monitoring platforms of energy-efficient, offer individualized management of energy/storage service and feel the smart
customization and flexible energy transaction.
42
2.3.3 |Energy market of emerging as well as business model cultivation
The multi-energy system can achieve automatic remote collection of gas, water, heat, and electricity in 1 m through IoT
system. The benefits of the IoT-based energy system are that it enables an efficient and seamless sharing and trading of
energy resources through an integrated site, enabling an opened and shared market, which adds small and micro users
like individuals and families.
MUTHUBALAJI ET AL.7of26

3|MATHEMATICAL MODELING
The objective function of SO
2
SA method is formulated. In this proposed work, the total cost of electricity should not be
correctly the most tremendous restriction characterized that is formulated below
maxXpixið1Þ
X
n
i¼1
ωixi≤Cð2Þ
here p
i
refers every items profit, x
i
implies that binary number 1 and 0 represents devices ON/OFF condition, ω
i
indicates
every item weight. The major aim is for decreasing electrical costs through increasing profit that is articulated as below
Min X
T
t¼1X
m
i¼1
PTγit
ðÞ ð3Þ
Subject to,
X
T
t¼1X
m
i¼1
PTγitðÞλtðÞ<Cð4Þ
And,
τsch ¼τot ð5Þ
At time interval t, the ON/OFF conditions of the device indicated γ
i
(t), the start and end times of device operation
indicated as τ
sch
and τ
ot
,Cdenotes maximal permissible power consumed. Maximum generation and system stability
problems are given increases while conquering these challenges.
3.1 |Modeling of PV system
The electrical power in PV system is directly proposed to the loads and based on the objective function the energy
source and load priority are available. The power generated through PV system is computed as below
Ppv ¼RtηpvApv ð6Þ
here output power of PV panels as P
pv
, solar radiation in the incline plane module as R
t
,η
pv
represents the efficiency of
the PV panel, A
pv
represents an area of the PV panel.
43
3.2 |Modeling of WT system
In the WT enabled for system IoT, it connects the matrix converter to the power grid and controls the output power
which is connected to the grid.
44
The given equations could be utilized to compute the WT output power in the sub-
system of wind energy,
Pwt ¼
0ifWS≤WSCI
αWS3βPrifWS
CI <WS<WS
R
Prif WSR≤WS < WSCO
0ifWS≥WSCO
8
>
>
>
<
>
>
>
:
ð7Þ
8of26 MUTHUBALAJI ET AL.

here WS as speed of wind, WS
CI
as wind cut-in speed, WS
R
as rated wind speed, WS
CO
as wind cut-off speed, and P
r
as
rated power.
3.3 |Modeling of FC
The FC is one of the generation units that is used to convert chemical energy into electrical energy. In FC, natural gas
is considered as an input unit and electrical and hydrogen gas as an output.
45
The input and output of FC is described
as follows
Hfc
h¼RTH pfcH2
hþpfc
hþpa
 ð8Þ
here R
th
is denoted as the thermal ratio for electric energy. In FC unit, the power generated is denoted through the
interval H. The auxiliary power device is described as p
a
. The limited level of hydrogen in the tank is given as follows
HtMin ≤Hth≤HtMax ð9Þ
In FC, the charging and discharging rate of hydrogen is denoted as follows
0≤Pfch2≤uhH:Pfch2
Max ð10Þ
0≤Pfch2≤uhDis
H:Pfch2
dis,Max ð11Þ
where UH
h
is denoted as the charging and discharging of binary variables. At each time, the hydrogen will be charge or
discharge.
UHhþUHdisc
h≤1;UHh,UHdisc
hð12Þ
The FC cost function is described as follows
Cfc ¼X
t
H¼1
Pfc
HþPfch2
HþPa
ηh
Cng
!
þηSt Pfch2
hþPfcMax om
"#
Thð13Þ
here C
ng
is denoted as the cost of natural gas and C
pump
is denoted as the cost of hydrogen pumping.
3.4 |Modeling of CHP units
Here, two types of CHP units were described and both units have different regions of operation.
46
The first type of CHP
is expressed as follows:
Pchp
Hi Pchp
Ai Pchjp
Ai Pchp
Bi
Hchp
Bi Hchp
Bi
Hchp
Hi Hchp
Ai

≤0ð14Þ
Pchp
Hi Pchp
Bi Pchjp
Bi Pchp
Ci
Hchp
Bi Hchp
Ci
Hchp
Hi Hchp
Bi

≥1Vchp
Hi

mð15Þ
MUTHUBALAJI ET AL.9of26

Pchp
Hi Pchp
Ci Pchjp
Ci Pchp
Di
Hchp
Ci Hchp
Di
Hchp
Hi Hchp
Ci

≥1Vchp
Hi

m
0≤Pchp
Hi ≤Pchp
Ai Vchp
Hi ð16Þ
0≤Hchp
Hi ≤Hchp
Bi Vchp
Hi ð17Þ
where Pchp
Hi is described as the power produced as CHP unit iat the time H,Hchp
Hi is described as the heat produced by
the CHP unit i at the time H.
47
In each CHP, the marginal points are denoted as A,B,C, and D. In second type, CHP is
characterized as follows
Pchp
Hi Pchp
Bi Pchjp
jBi Pchp
Ci
Hchp
Bi Hchp
Ci
Hchp
Hi Hchp
Bi

≤0ð18Þ
Pchp
Hi Pchp
Ci Pchjp
jCi Pchp
Di
Hchp
Ci Hchp
Di
Hchp
Hi Hchp
Ci

≤0ð19Þ
Pchp
Hi Pchp
Di Pchjp
Di Pchp
Ei
Hchp
Di Hchp
Ei
Hchp
Hi Hchp
Di

≥1x2HðÞmð20Þ
Pchp
Hi Pchp
Ei Pchjp
Ei Pchp
Fi
Hchp
Ei Hchp
Fi
Hchp
Hi Hchp
Ei

≥1x1HðÞmð21Þ
0≤Pchp
Hi ≤Pchp
Ai Vchp
Hi
0≤Pchp
Hi ≤Pchp
Ci Vchp
Hi
x1Hþx2H¼Vchp
Hi ð22Þ
Hchp
Hi Hchp
Ei

≤1x1HðÞm
Hchp
Hi Hchp
Ei

≤1x2HðÞmð23Þ
where E,Fis described as the marginal points of the second type in CHP. x
1
H,x
2
Hdenotes the operational state of
CHP. The below equation expresses the total operational cost of CHP.
CP
chp
H,Hchp
H

¼APchp
H

2þbPchp
HþCþDHchp
H

2þEHchp
HþFHchp
HPchp
Hð24Þ
here A,B,C,D,E, and Frepresents CHP cost function coefficients.
4|ENERGY MANAGEMENT IN SMART ENERGY SYSTEM USING SO
2
SA
In this section, EMS for DS utilizing a hybrid method is explained. The proposed hybrid method is a consolidated
implementation of SOA and OSA and hence it is named as SO
2
SA technique.
48,49
SOA search behavior is enhanced by
OSA technique.
10 of 26 MUTHUBALAJI ET AL.

As they are subject to a variety of optimization approaches: Genetic Algorithm (GA), Particle Swarm Optimization
(PSO), Bat Algorithm (BA), other optimization algorithms may not be modified for providing a general impact-based
valve point, forbidden operation zones, quadratic cost operation per parts. Optimization indicates the procedure of dic-
tating the best probable solutions for exact problems. The aforementioned optimization systems are mostly determinis-
tic and suffer from one major problem: local optimal trapping. This makes it enormously incompetent to solve real
problems. This indicates the incentive of such operations to establish novel algorithms stimulated by nature for compet-
ing with other optimization algorithms. Here, a new surrogate-assisted meta-heuristic is introduced for shape optimiza-
tion. SOA is utilized to solving the shape optimization of vehicle bracket. The design problem is to find structural shape
while minimizing structural mass and meeting a stress constraint. Function evaluations are carried out using finite ele-
ment analysis and estimated using a Kriging model. Also, a new nature-inspired optimization approach called OSA that
is used to solve the global optimization issues. OSA is the population-based method in terms of owls hunting mecha-
nism at dark. The SO
2
SA technique is verified on benchmark issues generally utilized in optimization field.
Both of those optimization models have unique advantages and disadvantages by doing it separately. By SOA inde-
pendently, the optimal control parameters are exhibited rely upon advantage, similar to capacity for moving toward
global optimal stages, joining a ton of cost operations, quicker implementation, and so on. Aside from the disadvantages
of SOA are each solution is getting just a concern objective, people can hold fast to local optima of individual objectives.
So as to improve the customary SOA, the searching behavior of SOA is improvised by OSA. In OSA, βdenotes linearly
reducing constant 1.9 to 0. Here, βexperiences huge modifications and encourages search space exploration. These
differences are minimized to encourage exploitation as the algorithm progresses. The SO
2
SA technique contains a user-
defined parameter (β), but GA, PSO, and BA having the maximum count of parametric settings. Subsequently, a consol-
idated procedure called SO
2
SA strategy is utilized to avoid shortcomings caused to the singular utilization of these two
optimization strategies. In this way, it is sensible to effective energy management in smart grids utilizing IoT.
4.1 |Steps of SO
2
SA technique
Step 1: Population initialization.
Initialization of seagull's population is the first step. In the initialization process, the generated output demand
response, power generated with source, cost function, limits of associated generation is the input, and randomly gener-
ate the demand response using the following equation
DR ¼
Y11 Y12
Y21 Y22
…Y1n
…Y2n
.
.
..
.
.
Yn1Yn2
.
.
..
.
.
 Ynn
2
6
6
6
6
6
4
3
7
7
7
7
7
5
ð25Þ
here the randomly created demand response is DR and Y
ij
implies random solutions.
Step 2: Fitness function.
The optimal energy is demonstrated depending on the electrical cost and profitability of every appliance at the smart
home. This is considered to be the fitness function. The fitness function is diagnosed by Equation (1).
Step 3: New Position Updating Using OSA.
Each owl's position is directly considered the related fitness value of intensity information obtained by the ears. So,
the better owl gets the maximum intensity (for maximum complications) because this is closest to the vole. To update
the position the generalized ith owl intensity information is employed
50
:
Ii¼fiw
bwð26Þ
here b¼max
k1,…,nfk,w¼min
k1,…,nfk, intensity information of ith owl can be indicated as I
i
. Each owl and prey distance
information has been calculated based on the following equation
MUTHUBALAJI ET AL.11 of 26

FIGURE 2 (A) Flowchart of SO
2
SA strategy. (B) Formulation of the proposed technique
12 of 26 MUTHUBALAJI ET AL.

TABLE 1 Simulation data of
sources Sources Minimum power Maximum power
PV 0 25
WT 0 20
FC 1 40
Battery 20 20
CHP 2 55
FIGURE 3 Analysis of load demand for 24 h
FIGURE 4 Load demand analysis to wind
FIGURE 5 Load demand analysis to 24 h periods of
photovoltaic
MUTHUBALAJI ET AL.13 of 26

Ri¼Oi,V
kk
2ð27Þ
here the prey location that is reached through the fittest owl can be indicated as V,ith owl can be specified as O
i
, each
owl and prey distance information can be specified as R
i
. At forest there is only one vole (global optimum) is consid-
ered. Owls move toward the prey they make silent fly. Therefore, they obtain a changed intensity subject to the sound
intensities inverse square law. The intensity modify to ith owl could be received as:
Ici¼Ii
R2
i
þRandom noise ð28Þ
In Equation (28), R2
iis used instead of 4πR2
iand to create the mathematic model much realistic the random noise of
environment is also assumed. Voles are active in real world, so its movements force the owls to silently changing its
FIGURE 6 Best allocation of microgrid power sources and grid in summer using the existing technique in the first scenario
FIGURE 7 Best allocation of microgrid power sources and grid in summer using proposed and existing techniques in the second
scenario
14 of 26 MUTHUBALAJI ET AL.

present position. The prey movement is planned utilizing probability, so novel owls positions could be acquired through
the given mechanism of position updating:
Otþ1
i¼Ot
iþβIciαVOt
i
,ifpvm <0:5
Ot
iβIciαVOt
i
,ifpvm ≥0:5
(ð29Þ
here probability of vole movement can be specified as p
vm
, uniformly distributed random number at range [0, 0.5] as α,
linearly minimizing constant 1.9 to 0 as β.
Step 4: Compute the fitness value of the new location.
Step 5: Make sure the termination criteria. If it is satisfied, they save the optimal solution or go to step 2. Figure 2A
represents the flowchart of the proposed hybrid method. The SO
2
SA technique formulation is delineated in Figure 2B.
FIGURE 9 Best allocation of microgrid power sources and grid in winter using proposed and existing techniques in Scenario 1
FIGURE 8 Best allocation of micro grid power sources and grid in summer using proposed and existing techniques in Scenario 3
MUTHUBALAJI ET AL.15 of 26

5|RESULT AND DISCUSSION
In this section, the EMS depending on SO
2
SA method for distribution system with IoT using the hybrid system is pres-
ented. The proposed method is used to optimize managing the distribution system power and resources through contin-
uous monitor the data as IoT-depend communication system. At SO
2
SA method, every device of home are connected to
a module of data acquisition, which is a unique IP address of an IoT object as a result of a network of large mesh wire-
less devices. Here, IoT-based communication framework will be used to facilitate EMS of demand response improve-
ment to the distribution system. The data for the result is taken from Reference 51. Based on the proposed and existing
technique, the load demand for wind, PV and for the time period of 24 h is presented and the cost comparison for the
proposed and existing technique is also presented. The summer and winter scenarios and 1 week were presented. In
addition, the proposed technique is responsible for meeting the total supply and energy requirement. Finally, the
FIGURE 10 Best allocation of microgrid power sources and grid in winter using proposed and existing techniques in Scenario 2
FIGURE 11 Best allocation of microgrid power sources and main grid in winter using proposed and existing techniques in Scenario 3
16 of 26 MUTHUBALAJI ET AL.

FIGURE 12 Analysis of load power with and without ESS
FIGURE 13 Analysis of load power with ESS
FIGURE 14 Analysis of load power without ESS
MUTHUBALAJI ET AL.17 of 26

proposed technique will be actualized at the worksite of MATLAB/Simulink. The simulation data of sources are tabu-
lated in Table 1.
Figure 3 portrays the analysis of load demand for 24 h. During January month, power for 24 h flows from 70 KW
and it flows up to 90 KW and high power is achieved on 100 KW at the time period of 20 s. During the month of July,
the power for 24 h flows from 69 KW and it reaches the maximum power up to 98 KW at the time period of 18 s and m
is reduced up to 90 KW. Figure 4 shows the analysis of load demand for wind. In summer, during the month of
January, the power flows from 5 KW and it flows up to 8 KW and high power achieved on 9 KW at the time period
of 20 s. In winter during the month of July, the power flows from 15 KW and it attains the maximal power up to
19 KW at 24 s time interval. Figure 5 portrays the analysis of load demand for 24 h' time period of PV. In summer, dur-
ing the month of July, the power flows from 0 KW and it flows up to 19 KW and it reduced to the power of 0 KW. In
winter during the month of January, the power flows from 0 KW and reaches the maximum power up to 12 KW at 7–
17 s time intervals. Figure 6 portrays the best allocation of the power of MG sources and grid in summer to the existing
technique in the first scenario. Here, the power value for FC is optimal at 45 KW, the CHP is increased up to 97 KW,
the wind is increased up to 99 KW, the PV is increased up to 100 KW, the battery is increased up to 102 KW and the
main grid is 53 KW.
Figure 7 shows the best allocation of the power of MG sources and grid in summer to proposed and existing tech-
nique in the second scenario. Here, the value of power for FC is optimal at 45 KW, the CHP is increased up to 97 KW,
the wind is increased up to 99 KW, the PV is increased up to 100 KW, the battery is increased up to 102 KW, and the
main grid is at 53 KW. Figure 8 shows the best allocation of the power of MG sources and grid in summer to proposed
and existing technique in Scenario 3. Here, the power values for FC are optimal at 45 KW, the CHP is increased up to
FIGURE 15 Cost comparison of proposed and existing
methods
FIGURE 16 Demand graph for 1 week
18 of 26 MUTHUBALAJI ET AL.

97 KW, the wind is increased up to 99 KW, the PV is increased up to 100 KW, the battery is increased up to 102 KW,
and the main grid is at 53 KW. Figure 9 shows the best allocation of the power of MG sources and grid in winter to
SO
2
SA and existing technique in Scenario 1. Here, values of power for FC are optimal at 40 KW, the CHP is increased
up to 80 KW, the wind is increased up to 85 KW, the PV is increased up to 88 KW, the battery is increased up to
100 KW, and the main grid is at 63 KW.
Figure 10 depicts best allocation of power of MG sources and grid in winter to proposed and existing technique in
Scenario 2. Here, the power value for FC is optimal at 40 KW, the CHP is increased up to 80 KW, the wind is increased
up to 85 KW, the PV is increased up to 88 KW, the battery is increased up to 100 KW, and the main grid is at 63 KW.
Figure 11 shows the best power allocation of MG sources and major grid at winter to proposed and existing technique
in Scenario 3. Here, the power value for FC is optimal at 40 KW, the CHP is increased up to 80 KW, the wind is
increased up to 85 KW, the PV is increased up to 100 KW, the battery is increased up to 102 KW, and the main grid
is at 63 KW. Figure 12 shows the analysis of with and without ESS. Here the power flows from 57 KW and flows up to
80 KW and high power is achieved on 85 KW in a period of 18 s. At the time period of 20 s, the power is reduced
to 68 KW.
Figure 13 shows the analysis of load power with ESS is presented. Here, the value of power for FC is optimal at
40 KW, the CHP is increased up to 80 KW, the wind is increased up to 85 KW, the PV is increased up to 100 KW, the
battery is increased up to 102 KW, and the main grid is at 63 KW. Figure 14 shows the analysis of load power without
ESS is presented. Here, the value of power for FC is optimal at 40 KW, the CHP is increased up to 80 KW, the wind is
increased up to 85 KW, the PV is increased up to 100 KW, and the main grid is at 63 KW. Figure 15 indicates the cost
comparison of proposed with existing techniques. Here, the cost of proposed technique reaches the maximum value of
0.8 s. The cost for the existing techniques of SOA is 0.9, OSA is 1, QOCSOS RF is 1, RF FA is 1, ALO is 1.1, and FA is
1.2. Compared to existing techniques the cost for the proposed method is less.
FIGURE 17 Power generation for 1 week by (A) photovoltaic, (B) wind turbine, (C) micro turbine, and (D) battery using the proposed
technique
MUTHUBALAJI ET AL.19 of 26

Figure 16 depicts the load requirement of SO
2
SA technique for 1 week. The investigation stage of the system is
defined by varying load demands. For example, the load demand is prolonged from 2620–2820 KW. Figure 17 repre-
sents the power generation of the MG sources as wind turbine, PV, micro turbine, and battery. During 1 week, the max-
imum PV power (700 KW) is generated using the proposed approach which is appeared in Figure 17A. In Figure 17B,
FIGURE 18 Cost comparison of proposed vs. existing systems. (A) Proposed versus FA. (B) Proposed versus ALO. (C) Proposed versus
RF-FA. (D) Proposed versus QOCSOS-RF. (E) Proposed versus OSA. (F) Proposed versus SOA
20 of 26 MUTHUBALAJI ET AL.

during the first and second day of the week, the high generated WT power is lower than 600 KW. By then the generated
power varies subject to the wind speed. In Figure 17C, the high generated power of MT is lower to 400 KW which is
generated subject to the battery operation discharge mode. In Figure 17D, the high generated power using the battery
is lower than 900 KW under the charge mode when the time interval t=15–20 h.
The cost of each MG resource with proposed and existing approaches is represented in Figure 18, which is analyzed
under charge and discharge mode during 1 week. The subplot Figure 18A–F depicts the cost comparison of the pro-
posed versus FA, proposed versus ALO, proposed versus RF-FA, proposed versus QOCSOS-RF, proposed versus OSA,
and proposed versus SOA techniques, respectively. The efficiency of the recent strategy is gathered by comparing the
outcomes with other approaches like FA,
52
ALO,
53
RF-FA,
52
QOCSOS-RF,
54
OSA, and SOA strategy. The battery is
operated normally under the charging mode when the time interval t=1–12 h. In these plots, operating costs and the
overall cost of SO
2
SA technique are low rather than the other techniques. The cost comparison of the SO
2
SA and exis-
ting techniques are plotted in Figure 19. Figure 19 clearly shows the efficiency of the proposed system depends on less
cost. The other existing technique provides a higher cost. For example, FA technique on the first day of the week gives
the cost value of nearly 390$. But our proposed technique gives the best result, that is, 300$.
Table 2 tabulates the total cost comparison using various techniques. Total cost is described to numerous sorts of
load requirement situation in this table. Then, MG statistical analysis is depicted in Figure 20A–C that includes mean,
median, and standard deviation of methods like proposed, SOA, OSA, and QOCSOS-RF system. The outcomes get as
statistic analysis displays that the SO
2
SA technique contains minimal deviation while compared to existing techniques.
Table 3 tabulates the percentage of cost accuracy, execution time in seconds, and cost of buying and selling the pro-
posed and existing techniques. By utilizing the given equation the accuracy percentage of cost is shown:
FIGURE 19 Cost comparison of the proposed and existing techniques
TABLE 2 Total cost comparison
using various techniques Load SO
2
SA SOA OSA QOCSOS-RF
3 0.401 0.418 0.425 0.436
4 0.695 0.705 0.709 0.712
5 0.713 0.728 0.735 0.743
6 0.95 0.98 0.99 1
7 0.98 0.99 1.001 1.030
9.6 0.985 0.989 0.95 1
10.82 0.9 0.94 0.97 1.001
11.53 1.8 1.88 1.96 2
14.61 1.8 1.86 1.94 2
MUTHUBALAJI ET AL.21 of 26

Accuracy %ðÞ¼
CBCW
CB
100 ð30Þ
Here, the best cost value can be represented as C
B
,C
W
represents worst cost value. In Equation (30), the proposed accu-
racy (%) and different methods are solved, as proved in Table 3. Table 3 specifies the author assumed the selling cost of
the proposed and different methods to be higher while compared to buying cost. Here, the SO
2
SA technique comprises
better-selling and the buying cost. Similarly, SO
2
SA technique accuracy is 9.91. At existing techniques, the SOA accu-
racy (%) is 9.89; OSA is 9.88, and QOCSOS-RF is 9.876. This outcome exhibit the SO
2
SA technique contains great
FIGURE 20 Statistical analysis (A) mean, (B) median, and (C) standard deviation
TABLE 3 Performance comparison of proposed and existing techniques
Method
Winter Summer
Accuracy Execution time
Selling Buying Selling Buying
SO
2
SA 1.08 1.02 0.95 0.72 9.91 3.58
SOA 1.06 0.98 0.90 0.71 9.89 4.0
OSA 1.04 0.95 0.84 0.70 9.88 4.1
QOCSOS—RF 1.02 0.91 0.79 0.68 9.876 4.206
22 of 26 MUTHUBALAJI ET AL.

accuracy while compared to existing approaches. Furthermore, the execution time of the SO
2
SA technique is great
when compared to other existing techniques.
Table 4 shows the proposed and existing techniques statistical analysis during 50 and 100 trials. For evaluating the
total performance of the SO
2
SA system is compared to existing systems. At 50 count of iterations, the range of RMSE,
MAPE, and MBE of SO
2
SA technique represents 5.63, 0.90, 1.035, and the existing system as RMSE, MAPE, and MBE
value of FA represents 18.91, 6.76, and 2.68. The RMSE, MAPE, and MBE value of ALO indicates 23.43, 13.54, and 5.73.
The RMSE, MAPE, and MBE value of RF-FA indicates 10.65, 1, and 2.87. The RMSE, MAPE, and MBE value of
QOCSOS-RF denotes 8.52, 0.99, and 1.78. In RMSE range, the OSA implies 7.63, MAPE implies 0.93, and MBE implies
1.23. In RMSE range, the SOA indicates 6.99, MAPE indicates 0.92, and MBE indicates 1.11. At 100 count of iterations,
the range of RMSE, MAPE, and MBE of SO
2
SA technique denote 6.25, 0.41, 2, and the existing system as RMSE, MAPE,
and MBE of FA denotes 21.57, 7.78, and 5.58. The RMSE, MAPE, and MBE of ALO are 26.78, 14.67, and 8.89. The
RMSE, MAPE, and MBE of RF-FA are 13.67, 2.56, and 5.75. The RMSE, MAPE, and MBE of QOCSOS-RF are 9.36,
1, and 3.87. The RMSE range of OSA, MAPE, and MBE is 9.36, 1, and 3.87. The RMSE range of OSA, MAPE, and MBE
is 8.65, 0.67, and 2.69. The RMSE range of SOA, MAPE, and MBE is 7.67, 0.55, and 2.17. The proposed technique is
compared to existing techniques based on specificity, sensitivity, and accuracy. The formula for the sensitivity, specific-
ity, and accuracy is referred from Reference 55. Table 5 displays the performance comparison of proposed with existing
techniques for 50 and 100 number of trials that include recall, specificity, accuracy, and precision. In 50 numbers of tri-
als, the SO
2
SA approach contains 1.77 accuracy, 1.99 specificity, recall is 2, and 1.74 precision. At the above existing
techniques, the accuracy of 0.75, specificity of 0.7, recall of 0.8, and precision of 0.72 in FA. At ALO, the accuracy of
0.85, specificity of 0.8, recall of 0.9, and precision of 0.82. At RF-FA, the accuracy of 0.90, specificity of 0.9, recall
TABLE 4 Metrics comparison for
50 and 100 number of trials
Metrics
50 trails
FA ALO RF-FA QOCSOS-RF OSA SOA Proposed
RMSE 18.91 23.43 10.65 8.52 7.63 6.99 5.43
MAPE 6.76 13.54 1 0.99 0.93 0.92 0.90
MBE 2.68 5.73 2.87 1.78 1.23 1.11 1.035
Metrics
100 trails
FA ALO RF-FA QOCSOS-RF OSA SOA Proposed
RMSE 21.57 26.78 13.67 9.36 8.65 7.67 6.25
MAPE 7.78 14.67 2.56 1 0.67 0.55 0.41
MBE 5.58 8.89 5.75 3.87 2.69 2.17 2
TABLE 5 Performance comparison of proposed and existing techniques for 50 and 100 number of trials
Performance measures
50 trails
FA ALO RF-FA QOCSOS-RF OSA SOA Proposed
Accuracy 0.75 0.85 0.90 0.91 0.97 1.23 1.77
Specificity 0.7 0.8 0.9 0.94 0.99 1.54 1.99
Recall 0.8 0.9 1 1.53 1.66 1.98 2
Precision 0.72 0.82 0.9 0.96 0.98 1.54 1.74
Performance measures
100 trails
FA ALO RF-FA QOCSOS-RF OSA SOA Proposed
Accuracy 0.7 0.8 0.9 0.93 1 1.43 1.5
Specificity 0.63 0.75 0.85 0.9 1 0.95 1.05
Recall 0.6 0.85 0.8 0.85 0.9 0.92 1.03
Precision 0.65 0.72 0.95 0.98 1 1.43 2
MUTHUBALAJI ET AL.23 of 26

of 1, and precision of 0.9. At QOCSOS-RF, the accuracy of 0.91, specificity of 0.94, recall of 1.53, and precision of 0.96.
At OSA, the accuracy of 0.97, specificity of 0.99, recall of 1.66, and precision ranges of 0.98. At SOA, the accuracy, speci-
ficity, recall, and precision ranges are 1.23, 1.54, 1.98, and 1.54. In 100 number of trials, the SO
2
SA technique contains
1.5 accuracy, 1.05 specificity, 1.03 recall, and 2 precision. In the above existing techniques of FA has an accuracy of 0.7,
specificity of 0.63, recall of 0.6, and precision of 0.65. At ALO, the accuracy of 0.8, specificity of 0.75, recall of 0.85, and
precision of 0.72. At RF-FA, the accuracy of 0.9, specificity of 0.85, recall of 0.8, and precision of 0.95. At QOCSOS-RF,
the accuracy of 0.93, specificity of 0.9, recall of 0.85, and precision of 0.98. The OSA has 1 accuracy, 1 specificity, 0.9
recall, and1precision. The SOA has 1.43 accuracy, 0.95 specificity, 0.92 recall, and 1.43 precision. Compared to the exis-
ting techniques, the proposed technique executes more accurately. The simulation results show that the proposed tech-
nique is more efficient than other existing techniques.
6|CONCLUSION
A high-performance SO
2
SA approach is proposed based on EMS for distribution systems with IoT using a hybrid sys-
tem. The proposed method is used for optimal managing the power and distribution system resources by continuously
monitoring the data as a communication process of IoT. The proposed method is implemented in MATLAB/Simulink
site. The performance of the proposed method is compared to other existing methods. The simulation outcomes demon-
strate that the proposed method is responsible for meeting the total supply and energy requirement.
In the future, the presence of uncertainty deals with implementing the new optimization strategy that is a reactive
programming system and iteratively solves the deterministic issue by advancing the optimization horizon at each repe-
tition; consider the state of the system is upgraded as soon as the various uncertain or insufficiently precise parameters
are known, the optimal program for new out coming scenario (and optimization horizon) can be found. This strategy
deems a prediction horizon, in which it is assumed that all the uncertain parameters linked to this time horizon are
known with a few certainties (because the system under study receives feedback linked with unknown parameters) and
a horizon of control, wherever optimization decisions are used for actual prediction horizon.
DATA AVAILABILITY STATEMENT
Data sharing does not apply to this manuscript as no new data has been created or examined in this research.
ORCID
Sankaramoorthy Muthubalaji https://orcid.org/0000-0002-1898-8403
REFERENCES
1. Abdelaziz E, Saidur R, Mekhilef S. A review on energy saving strategies in industrial sector. Renew Sustain Energy Rev. 2011;15(1):150-
168. https://doi.org/10.1016/j.rser.2010.09.003.
2. Trianni A, Cagno E, Farné S. Barriers, drivers and decision-making process for industrial energy efficiency: a broad study among
manufacturing small and medium-sized enterprises. Appl Energy. 2016;162:1537-1551. https://doi.org/10.1016/j.apenergy.2015.02.078.
3. Tao F, Zuo Y, Da Xu L, Lin L, Lin Z. Internet of things and BOM-based life cycle assessment of energy-saving and emission-reduction of
products. IEEE Trans Ind Inform. 2014;10(2):1252-1261. https://doi.org/10.1109/tii.2014.2306771.
4. Bi Z, Da Xu L, Wang C. Internet of things for enterprise systems of modern manufacturing. IEEE Trans Ind Inform. 2014;10(2):1537-
1546. https://doi.org/10.1109/tii.2014.2300338.
5. Tao F, Cheng Y, Da Xu L, Lin Z, Li BH. CCIoT-CMfg: cloud computing and internet of things-based cloud manufacturing service sys-
tem. IEEE Trans Ind Inform. 2014;10(2):1435-1442. https://doi.org/10.1109/tii.2014.2306383.
6. Dupré A, Drobinski P, Alonzo B, Badosa J, Briard C, Plougonven R. Sub-hourly forecasting of wind speed and wind energy. Renew
Energy. 2020;145:2373-2379. https://doi.org/10.1016/j.renene.2019.07.161.
7. Aslam S, Iqbal Z, Javaid N, Khan ZA, Aurangzeb K, Haider SI. Towards efficient energy management of smart buildings exploiting heu-
ristic optimization with real time and critical peak pricing schemes. Energies. 2017;10(12):2065. https://doi.org/10.3390/en10122065.
8. Aurangzeb K, Aslam S, Haider SI, et al. Energy forecasting using multiheaded convolutional neural networks in efficient renewable
energy resources equipped with energy storage system. Trans Emerg Telecommun Technol. 2019;e3837. https://doi.org/10.1002/ett.3837.
9. Aslam S, Khalid A, Javaid N. Towards efficient energy management in smart grids considering microgrids with day-ahead energy fore-
casting. Electr Pow Syst Res. 2020;182:106232. https://doi.org/10.1016/j.epsr.2020.106232.
10. Siano P. Demand response and smart grids—a survey. Renew Sustain Energy Rev. 2014;30:461-478. https://doi.org/10.1016/j.rser.2013.
10.022.
24 of 26 MUTHUBALAJI ET AL.

11. Marzband M, Sumper A, Domínguez-García J, Gumara-Ferret R. Experimental validation of a real time energy management system for
microgrids in islanded mode using a local day-ahead electricity market and MINLP. Energy Conver Manag. 2013;76:314-322. https://doi.
org/10.1016/j.enconman.2013.07.053.
12. Marzband M, Ghadimi M, Sumper A, Domínguez-García J. Experimental validation of a real-time energy management system using
multi-period gravitational search algorithm for microgrids in islanded mode. Appl Energy. 2014;128:164-174. https://doi.org/10.1016/j.
apenergy.2014.04.056.
13. Neves D, Pina A, Silva C. Demand response modeling: a comparison between tools. Appl Energy. 2015;146:288-297. https://doi.org/10.
1016/j.apenergy.2015.02.057.
14. Ding Y, Hong S, Li X. A demand response energy management scheme for industrial facilities in smart grid. IEEE Trans Industr Inform.
2014;10(4):2257-2269. https://doi.org/10.1109/tii.2014.2330995.
15. Kondili E, Pantelides C, Sargent R. A general algorithm for short-term scheduling of batch operations—I. MILP formulation. Comput
Chem Eng. 1993;17(2):211-227. https://doi.org/10.1016/0098-1354(93)80015-f.
16. Ferrari P, Flammini A, Vitturi S. Performance analysis of PROFINET networks. Comput Stand Interf. 2006;28(4):369-385. https://doi.
org/10.1016/j.csi.2005.03.008.
17. Jansen D, Buttner H. Real-time Ethernet: the EtherCAT solution. Comput Contr Eng. 2004;15(1):16-21. https://doi.org/10.1049/cce:
20040104.
18. Erwinski K, Paprocki M, Grzesiak L, Karwowski K, Wawrzak A. Application of Ethernet Powerlink for communication in a Linux RTAI
open CNC system. IEEE Trans Ind Electron. 2013;60(2):628-636. https://doi.org/10.1109/tie.2012.2206348.
19. Figueiredo V, Rodrigues F, Vale Z, Gouveia J. An electric energy consumer characterization framework based on data mining tech-
niques. IEEE Trans Power Syst. 2005;20(2):596-602. https://doi.org/10.1109/tpwrs.2005.846234.
20. Espinoza M, Joye C, Belmans R, DeMoor B. Short-term load forecasting, profile identification, and customer segmentation: a methodol-
ogy based on periodic time series. IEEE Trans Power Syst. 2005;20(3):1622-1630. https://doi.org/10.1109/tpwrs.2005.852123.
21. Albert A, Rajagopal R. Smart meter driven segmentation: what your consumption says about you. IEEE Trans Power Syst. 2013;28(4):
4019-4030. https://doi.org/10.1109/tpwrs.2013.2266122.
22. Ahmad S, Baig Z. Fuzzy-based optimization for effective detection of smart grid cyber-attacks. Int J Smart Grid Clean Energy. 2012;1(1):
15-21. https://doi.org/10.12720/sgce.1.1.15-21.
23. Moro J, Duarte L, Ferreira E, Dias J. A home appliance recognition system using the approach of measuring power consumption and
power factor on the electrical panel, based on energy meter ICs. Circuits Syst. 2013;04(03):245-251. https://doi.org/10.4236/cs.2013.
43033.
24. Wu C, Chang R, Chan H. A green energy-efficient scheduling algorithm using the DVFS technique for cloud datacenters. Fut Gen Com-
put Syst. 2014;37:141-147. https://doi.org/10.1016/j.future.2013.06.009.
25. Faheem M, Butt R, Raza B, Ashraf M, Ngadi M, Gungor V. Energy efficient and reliable data gathering using internet of software-
defined mobile sinks for WSNs-based smart grid applications. Comput Stand Interf. 2019;66:103341. https://doi.org/10.1016/j.csi.2019.
03.009.
26. Liu Y, Yang C, Jiang L, Xie S, Zhang Y. Intelligent edge computing for IoT-based energy Management in Smart Cities. IEEE Netw. 2019;
33(2):111-117. https://doi.org/10.1109/mnet.2019.1800254.
27. Yassine A, Singh S, Hossain M, Muhammad G. IoT big data analytics for smart homes with fog and cloud computing. Fut Gen Comput
Syst. 2019;91:563-573. https://doi.org/10.1016/j.future.2018.08.040.
28. Li W, Logenthiran T, Phan V, Woo W. A novel smart energy theft system (SETS) for IoT-based smart home. IEEE Internet Things J.
2019;6(3):5531-5539. https://doi.org/10.1109/jiot.2019.2903281.
29. Tom R, Sankaranarayanan S, Rodrigues J. Smart energy management and demand reduction by consumers and utilities in an IoT-fog-
based power distribution system. IEEE Internet Things J. 2019;6(5):7386-7394. https://doi.org/10.1109/jiot.2019.2894326.
30. Ahuja K, Khosla A. Data analytics criteria of IoT enabled smart energy meters (SEMs) in smart cities. Int J Energy Sec Manag. 2019;13
(2):402-423. https://doi.org/10.1108/ijesm-11-2017-0006.
31. Sureshkumar K, Ponnusamy V. Hybrid renewable energy systems for power flow management in smart grid using an efficient hybrid
technique. Trans Inst Measure Contr. 2020;42:2068-2087.
32. Roy K, Mandal KK, Mandal AC. Application of ANFASO for optimal power flow management of MG-connected system with energy
storage. Int J Energy Res. 2020;44(7):5272-5286. https://doi.org/10.1002/er.5273.
33. Durairasan M, Balasubramanian D. An efficient control strategy for optimal power flow management from a renewable energy source
to a generalized three-phase microgrid system: a hybrid squirrel search algorithm with whale optimization algorithm approach. Trans
Inst Measure Contr. 2020;42:1960-1976.
34. Snoonian D. Smart buildings. IEEE Spectr. 2003;40(8):18-23. https://doi.org/10.1109/mspec.2003.1222043.
35. Khan ZA, Zafar A, Javaid S, Aslam S, Rahim MH, Javaid N. Hybrid meta-heuristic optimization based home energy management system
in smart grid. J Ambient Intell Humaniz Comput. 2019;10(12):4837-4853.
36. Xu Z, Guan X, Jia Q, Wu J, Wang D, Chen S. Performance analysis and comparison on energy storage devices for smart building energy
management. IEEE Trans Smart Grid. 2012;3(4):2136-2147. https://doi.org/10.1109/tsg.2012.2218836.
37. Masera M, Bompard E, Profumo F, Hadjsaid N. Smart (electricity) grids for smart cities: assessing roles and societal impacts. Proc IEEE.
2018;106(4):613-625. https://doi.org/10.1109/jproc.2018.2812212.
38. Dagoumas A. Modelling socio-economic and energy aspects of urban systems. Sustain Cities Soc. 2014;13:192-206. https://doi.org/10.
1016/j.scs.2013.11.003.
MUTHUBALAJI ET AL.25 of 26

39. Liu Y, Yuen C, Ul Hassan N, Huang S, Yu R, Xie S. Electricity cost minimization for a microgrid with distributed energy resource under
different information availability. IEEE Trans Ind Electron. 2015;62(4):2571-2583. https://doi.org/10.1109/tie.2014.2371780.
40. Zhong W, Xie K, Liu Y, Yang C, Xie S. Auction mechanisms for energy trading in multi-energy systems. IEEE Trans Industr Inform.
2018;14(4):1511-1521. https://doi.org/10.1109/tii.2017.2787751.
41. Mnih V, Kavukcuoglu K, Silver D, et al. Human-level control through deep reinforcement learning. Nature. 2015;518(7540):529-533.
https://doi.org/10.1038/nature14236.
42. Hassan N, Pasha M, Yuen C, Huang S, Wang X. Impact of scheduling flexibility on demand profile flatness and user inconvenience in
residential smart grid system. Energies. 2013;6(12):6608-6635. https://doi.org/10.3390/en6126608.
43. Khouzam K, Khouzam L, Groumpos P. Optimum matching of ohmic loads to the photovoltaic array. Solar Energy. 1991;46(2):101-108.
https://doi.org/10.1016/0038-092x(91)90022-o.
44. Esfahanian V, Torabi F, Mosahebi A. An innovative computational algorithm for simulation of lead-acid batteries. J Power Sources.
2008;176(1):373-380. https://doi.org/10.1016/j.jpowsour.2007.10.052.
45. Castañeda M, Cano A, Jurado F, S
anchez H, Fern
andez L. Sizing optimization, dynamic modeling and energy management strategies of
a stand-alone PV/hydrogen/battery-based hybrid system. Int J Hydrogen Energy. 2013;38(10):3830-3845. https://doi.org/10.1016/j.
ijhydene.2013.01.080.
46. Jenkins D, Fletcher J, Kane D. Lifetime prediction and sizing of lead–acid batteries for microgeneration storage applications. IET Renew
Power Gen. 2008;2(3):191-200. https://doi.org/10.1049/iet-rpg:20080021.
47. Nojavan S, Majidi M, Zare K. Risk-based optimal performance of a PV/fuel cell/battery/grid hybrid energy system using information gap
decision theory in the presence of demand response program. Int J Hydrogen Energy. 2017;42(16):11857-11867. https://doi.org/10.1016/j.
ijhydene.2017.02.147.
48. Dhiman G, Kumar V. Seagull optimization algorithm: theory and its applications for large-scale industrial engineering problems. Knowl
Based Syst. 2019;165:169-196. https://doi.org/10.1016/j.knosys.2018.11.024.
49. Transpire Online. A new seagull optimization algorithm to solve numerical optimization problems, transpire online 2019. Available from
https://transpireonline.blog/2019/08/30/a-new-seagull-optimization-algorithm-to-solve-numerical-optimization-problems/. Accessed
January, 2020.
50. Jain M, Maurya S, Rani A, Singh V. Owl search algorithm: a novel nature-inspired heuristic paradigm for global optimization. J Intell
Fuzzy Syst. 2018;34(3):1573-1582. https://doi.org/10.3233/jifs-169452.
51. Elsied M, Oukaour A, Gualous H, Hassan R. Energy management and optimization in microgrid system based on green energy. Energy.
2015;84:139-151. https://doi.org/10.1016/j.energy.2015.02.108.
52. Rehman AU, Aslam S, Abideen ZU, et al. Efficient energy management system using firefly and harmony search algorithm. Interna-
tional Conference on Broadband and Wireless Computing, Communication and Applications; 2017; Springer, Cham. pp. 37-49. https://
doi.org/10.1007/978-3-319-69811-3_4
53. Roy K, Mandal KK, Mandal AC. Ant-lion optimizer algorithm and recurrent neural network for energy management of micro grid con-
nected system. Energy. 2019;167:402-416. https://doi.org/10.1016/j.energy.2018.10.153.
54. Chen J, Zhou Z, Karunakaran V, Zhao S. An efficient technique-based distributed energy management for hybrid MG system: a hybrid
QOCSOS-RF technique. Wind Energy. 2020;23(3):575-592. https://doi.org/10.1002/we.2443.
55. Zhu W, Zeng N, Wang N. Sensitivity, specificity, accuracy, associated confidence interval and ROC analysis with practical SAS
implementations. NESUG Proceedings: Health Care and Life Sciences Vol. 19, Baltimore, Maryland; 2010. p. 67.
How to cite this article: Muthubalaji S, Srinivasan S, Lakshmanan M. IoT based energy management in smart
energy system: A hybrid SO
2
SA technique. Int J Numer Model. 2021;e2893. https://doi.org/10.1002/jnm.2893
26 of 26 MUTHUBALAJI ET AL.