Hybrid Technique-based Optimal Energy Management in Smart Home Appliances
ShirleyC. P.1,*
RA BDepaa 2
PriyaA.3
SaralaR.4
SolankiRajdeep Singh5
K. VMalini 6
ReddyCh. Venkatakrishna7
-
( Assistant Professor, Department of Computer Science and Engineering, Karunya Institute
of Technology and Sciences, Coimbatore, India shirleycp@karunya.edu)
-
( Deputy HOD and Associate Professor, Department of Civil Engineering, Dr. M.G.R Educational
and Research Institute, Maduravoyal, Chennai 95, India depaa.civil@drmgrdu.ac.in)
-
( Associate Professor, School of Architecture, Koneru Lakshmaiah Educational Foundation,
Guntur, Andhra Pradesh, India ar.priyajohnson15@gmail.com)
-
( Assistant Professor, Department of Computer Science and Engineering, Velammal College
of Engineering and Technology, Madurai Rameshwaram High Road, Viraganoor-625009, Madurai,
India srr@vcet.ac.in)
-
( Assistant Professor, Department of Computer Science, Medi-caps University Indore,
India rajdeepslnk@yahoo.co.in)
-
( Assistant Professor and HOD, Department of Electrical and Electronics Engineering,
Sri Sairam College of Engineering Anekal, India malini.eee@sairamce.edu.in)
-
( Assistant professor, Department of Electrical and Electronics Engineering, Chaitanya
Bharathi Institute of Technology, Hyderabad -75, India chkrishnareddy_eee@cbit.ac.in)
Copyright © The Institute of Electronics and Information Engineers(IEIE)
Keywords
Air conditioning, Demand response (DR), Solar, Home energy management system (HEMS), Battery energy storage system, Electric vehicle
1. Introduction
Electricity demand in residential areas is increasing faster than the constant expansion
in energy-producing facilities, expanding the supply–demand mismatch [1,2]. In a traditional grid, public services work to close the deficit by expanding electricity
generation. On the other hand, these half-solutions or so-called temporary solutions
may affect the security restrictions of the electrical system, resulting in some serious
scenarios, such as the largest load capacity issue during peak hours [3,4]. As a result, the total cost of electricity has also increased because of the higher
costs related to the charges used during on-peak hours [5]. To manage such a situation, utilities advise their customers to reduce electricity
costs by avoiding peak usage associated with high pricing during on-peak hours and
develop effective power company planning and management functions designed to encourage
end-users to modify their position and energy usage model [6]. Nevertheless, utilities are forced to deal with these issues because of a lack of
consumer awareness on estimating the peak hours and time-varying rates. Many demand
response solutions have been proposed to handle the problem of load management during
peak hours in domestic sectors in recent years [7]. These solutions have the same goals: lessening the electricity consumption costs,
peak-to-average ratio, and improving the waiting time or customer comfort [8-10]. In exchange, utility providers use various pricing rates, the most prevalent of
which are the time-of-use (TOU), critical-peak-price, day-ahead pricing, and real-time
pricing [11,12].
Several studies have presented smart home energy management using different methods
[13,14]. Some of them are described below,
ur Rehman et al. [15] have provided a holistic strategy for optimizing the use of various household appliances
depending on the schedule and preferences of the prosumer. Bahmanyar et al. [16] suggested the multi-objective arithmetic optimization algorithm (MOAOA) as a multi-objective
variant of the arithmetic optimization algorithm (AOA). Also, a newly discovered meta-heuristic
method was used to determine the optimum scheduling of household appliances. Erenoğlu
et al. [17] have introduced a mixed-integer linear programming (MILP) problem that was utilized
to model a scenario-related energy management system (EMS). The objective of the EMS
was to reduce energy losses on the network while accounting for the stochastic characteristics
of PV and wind sources. Mohammad et al. [18] have introduced an approach to solving a multi-objective optimization issue that
incorporates power cost and system PAR. Tostado-Véliz et al. [19] have developed a stochastic many-objective solution framework based on the mixed
integer linear programming formulation.
A literature review showed that the lack of user information on estimating the peak
hours and the time-varying costs has forced utility companies to manage these issues.
Charging utilities use advanced metering infrastructure (AMI) to implement residential
DR by providing end-users with the financial incentives and time-dependent pricing
to undertake direct or indirect load control (DLC or ILC). Direct load control involves
end users granting LSEs remote control of their devices in exchange for incentives,
primarily spike mitigation or frequency regulation. On the other hand, DLC may cause
reluctance based on loss of control and privacy concerns. Real-time optimization of
energy management is offered to decrease the cost of energy usage in peak hours. The
above-mentioned limits inspired this research work.
The rest of the manuscript is arranged as follows: Section 1 discusses the introduction,
recent research work, and their context. Section 2 explains the configuration of the
home energy management system, Section 3 describes the proposed EOSA-SNN-based optimal
energy management in smart house appliances, Section 4 demonstrates the results and
discussion, and in Section 5, the manuscript concludes.
Fig. 1. Structure of HEMS and the current electrical loads in the home.
2. Configuration of the Home Energy Management System
The chosen AC starting set-point temperature, desired TSA operational limits, selected
electric hot water usage timings, and electric vehicle charging times preferred by
the occupants must be considered when applying information about their presence. By
specifying the possible values of the AC set point for each period, the built-in smart
thermostat improves the prediction of incoming solar radiation, user-defined occupant
presence information, and energy price data. The Load Serving Entity (LSE) transmits
DR information, energy price, and weather forecast to the home. Day-ahead load scheduling
is carried out using the home energy management system optimization algorithm to offer
self-consumption and demand response.
The bi-directional power flow among BESS, home, grid, and EV was evaluated, and battery
deterioration was measured to avoid wasteful energy arbitrage. Fig 1 portrays the
structure of HEMS and the current electrical loads in the home.
2.1 Photovoltaic Model
According to Liu and Jordan, the isotropic solar method was used in the proposed HEMS
to determine the output of photovoltaic arrays on tilted surfaces and net solar radiation.
In this method, a system operator or LSE sends the price signal and day-ahead solar
prediction for a city or other place to households.
2.2 Power-shiftable Appliances (Psas)
The battery energy storage system (ESS) is given below:
According to Eq. (1), the combined power used in the home and power supplied with the grid comprises the
discharged power of the battery ESS. According to Eqs. (2) and (3), the power of charging and discharging power for a battery ESS at a given moment
cannot be more than the rate of charging and discharging of the battery. The state-of-energy
(SoE) of the battery ESS is given in Eq. (4) and (5). According to Eq. (6), the SoE of BESS must be less than the permitted depth of discharge (DoD) and cannot
be greater than the maximal capacity of a battery [20].
2.2.1 Electric Vehicle (EV)
The EV charge and discharge are restricted by the arrival and departure periods [21].
2.3 Thermostatically Controlled
The operation of a home EMS is classified into three TCAs, like a refrigerator, AC,
and EWH. According to numerous studies, a 1st-order lumped capacitance 1R1C gray-box
method is sufficiently dependable to represent the thermal behavior of a house, a
refrigerator cabinet, and a EWH tank [22].
2.3.1 Refrigerator
The thermal model to EWH is indistinguishable from its own. However, in the case of
cooling operations, the sign of the decision variable is reversed and it becomes negative.
How the door openings affect the cabinet temperature is disregarded. The latter is
not included in the model because it has little impact on the temperature and energy
use.
2.3.2 Air Conditioner (AC)
The air conditioner method is identical to that of a refrigerator. In this case, Eq.
(7) is regarded as the AC cooling function. Cooling is accomplished using the COP. The
influence of ventilation on the interior temperature is ignored. The electrical power
consumption of the AC is represented using Eq. (9).
where $t_{t}^{out}$denotes EV home arrival; $r^{AC}$indicates the house; $p^{AC}$
is the AC; $X_{t}^{AC}$ is the decision variable between 0 and 1 defining the AC;
$sp_{t}^{\min }$ AC is the min set point temperature; $p_{t}^{\max }$is the AC max
set point temperature; $COP^{AC}$ is the AC.
2.4 Problem Formulation
The purpose of the Hybrid EMS optimization challenge is to reduce the day-to-day electricity
expenditure of a household:
where $DoD^{v}$ is denoted as the EV depth of discharge; $DoD^{b}$ is the BESS; $cyc^{b}$indicates
the BESS lifecycle; $cap^{b}$ is the EV battery capacity; $l^{v}$ is the EV lifetime
throughput energy; $l^{b}$ is the lifetime of BESS. In this case, PSAs distinguish
themselves by incorporating battery degradation costs into their acquiring and selling
prices, preventing needless energy arbitrage. As a result, $\lambda _{t}^{v,buy}$and
$\lambda _{t}^{b,buy}$are the sum of the purchase price $\lambda _{t}^{buy}$, the
price of purchasing, and $\lambda ^{v,\deg }$; $\lambda ^{b,\deg }$ is the price of
battery deterioration. The total of the selling price $\lambda _{t}^{sell}$and the
battery degradation prices ($\lambda _{t}^{v,\deg }$,$\lambda _{t}^{v,\deg }$) is
represented by $\lambda _{t}^{v,\deg }$and $\lambda _{t}^{b,\deg }$, respectively.
Thus,
·Only V2G or B2G are considered if the battery degradation cost exceeds the merits
of selling electricity to the grid.
·B2H and V2H are considered if the merits of purchasing energy for later use at home
outweigh the expense of battery degeneration.
Degradation cost is accounted for in EV and BESS prices to optimize load scheduling,
subtracting it from the purchasing and selling expenses.
2.5 Power Balance
The HEMS power balance is computed as follows:
where$p_{t}^{h}$refers to the power consumption of inflexible and flexible uses, excluding
Power-Shiftable appliances.
where $p_{t}^{PV,used}$is the power of photovoltaic, which distributes energy demand
of inflexible and flexible uses; $p_{t}^{PV,2g}$ is the power transfer PV to the battery.
Similarly, $p_{t}^{V,used}$ and $p_{t}^{b,used}$refer to the powers applied to the
flexible and inflexible demand of energy uses and PSA charging because of V2B, V2H,
B2H, and B2V operations.
The production of photovoltaics is used mainly in self-consumption and energy storage.
The excess production is transferred to the grid $p_{t}^{b,2g}$is the power transfer
of the battery to the grid.
The system operator induces the power limit of the power introduced to or from the
grid
3. Proposed Ebola Optimization Search Algorithm (EOSA)-spiking Neural Networks (SNN)-based
Optimal Energy Management in Smart House Appliances
A hybrid method is proposed for the optimal energy management of smart home appliances.
The EOSA method is used to manage the air-conditioning and maintenance of thermal
comfort, and SNN is used to predict the optimal control signal of the system.
3.1 Proposed EOSA Algorithm
The EOSA is one of meta-heuristic model that is based on the propagation mechanism
of Ebola virus disease. This proposed algorithm was introduced recently to identify
Ebola virus propagation among humans and calculate vaccinated, infected, hospitalized,
recovered, death infected, and death [23]. Fig. 2 presents the proposed EOSA flowchart.
Fig. 2. Flow chart of the Proposed EOSA.
Step 1: Initialization
Initiate the input-parameters, such as the occupant presence, solar radiation, and
electricity prices.
Step 2: Random Generation
After initialization, the input-parameters are created randomly:
(27)
$
R=\left[\begin{array}{l}
\left[X_{p,v,c,l}^{11}\left(t\right)\,\,\,X_{p,v,c,l}^{12}\left(t\right)\,\,\ldots
\,\,\,\,\,X_{p,v,c,l}^{1k}\left(t\right)\right]\\
\left[X_{p,v,i,l}^{21}\left(t\right)\,\,\,X_{p,v,c,l}^{22}\left(t\right)\,\,\,\ldots
\,\,\,\,X_{p,v,c,l}^{2k}\left(t\right)\right]\\
\,\,\,\,\,\,\,\,\vdots \,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\vdots
\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\\
\left[X_{p,v,c,l}^{n1}\left(t\right)\,\,X_{p,v,c,l}^{n2}\left(t\right)\,\,\,\ldots
\,\,\,\,\,X_{p,v,c,l}^{nk}\left(t\right)\right]
\end{array}\right]
$
where $R$ is the random generation, and X is the gain parameters of the FOPID controller;
$X_{p,v,c,l}\left(t\right)$ represents the gain parameters of the PI controller
Step 3: Fitness Function
The fitness is selected based on the objective function.
Step 4: Search Space Estimation
In the EOSA, the number of people exposed to the Ebola virus is calculated as
where $\alpha $is the scalar quantity of the individuals; $ZIn_{_{i}}^{t+1}$and $ZIn_{i}^{t}$
denote the original and updated positions of the individuals, respectively. $Mov\left(In\right)$
represents the movement rate of individuals; each function was performed with the
respective time $t+1$ and $t$.
Step 5: Exploration Phase
In the exploration phase, to calculate the doubted persons
Step 6: Exploitation Phase
In the exploitation phase, the neighborhood range is to calculated.
Step 7: Termination Criteria
Check the termination criteria, and if the optimum solution is obtained, then the
process stops; otherwise, repeat the step 3.
3.2 Spiking Neural Network
The word SNN is used for work encompassing any form of neural network incorporating
the concept of time. The SNN algorithm has the capability to autonomously adjust the
SNN [24] throughout the learning phase. SNN is effective in balancing exploration and exploitation
in diverse applications. Neural coding methods, such as population, rate, and temporal
encoding, represent real numbers using spike trains. It is assumed that sequence-related
temporal events are suppressed in a temporal brain coding system that is crucial for
applications involving sequence modeling and forecasting. It transmits data into the
network continuously at the exact moment of spikes. The firing time series symbolizes
spike trains, which consist of the input and output of neurons that spike. Recent
spikes have a larger impact on the possibility of net action than previous spikes.
The neuron fires the spike if the integrated sum of the incoming spikes is larger
than a predefined value. The SNN is considered a critical component for online load
forecasting because it is a dynamic system.
4. Results and Discussion
This section evaluates the performance of the proposed method based on the simulation
results. This manuscript utilizes the hybrid EOSA-SNN approach to lessen the cost
and energy management in smart house appliances. The performance of the proposed method
is assessed using the MATLAB platform and is compared with existing Particle Swarm
Optimization (PSO), Radiofrequency Thermal Ablation (RFA), and Border Collie Optimization
(BCO) methods. The cost reduction was 0 to 13% lower than the existing approaches.
Fig. 3 presents the analysis results of the dry bulb temperature and solar radiation data
for (a) irradiation and (b) temperature. Subplot 3 (a) shows the irradiation. The
value started at 0W/m$^{2}$ at one hour and ended at 0 W/m$^{2}$ at 25 hours. Subplot
3(b) shows the temperature starting at 24.5 $^{\circ}$C at zero hours and ending at
24.6 $^{\circ}$C at zero to 25 hours. Fig. 4 presents the analysis of the daily presence of occupants in the household. In the
sleep hours, the value started at 3.9＄/KWh at zero hours and ended at 0.9＄/KWh at
six hours. When all were at home, the value was 4＄/KWh in 24 hours. Fig. 5 presents smart thermostat input parameters, depicting variations in Electricity Price
and Solar Radiation. Subplot 5(a) shows Electricity Price ranging from 0.078 to 0.25
over 25 hours, while Solar Radiation decreases from 0.7 to 0.07. Subplot 5(b) displays
a consistent Temperature set-point of 23.56 over 24 hours.
Fig. 3. Analysis of the dry bulb temperature and solar radiation data for (a) irradiation; (b) temperature.
Fig. 4. Analysis of the daily presence of occupants in the household.
Fig. 5. Analysis of the smart thermostat input parameters of (a) Amplitude; (b) Temperature in the proposed technique.
Fig. 6. Household’s daily base-load profile analysis uses the proposed method excluding: (a) battery storage; (b) load scheduling.
Fig. 6 depicts the Households daily base-load profile analysis uses the proposed method
excluding (a) battery storage and (b) load scheduling. Subplot 6(a) shows PV production
starting at 0 kW at 7 hours and increasing to 0.2 kW at 23 hours, while the electric
water heater begins at 0.3 kW at one hour and ends at 0.8 kW at 24 hours. Inflex load
remains at 2.2 kW at 13 hours, and PV reaches 2.7 kW at 13 hours. Subplot 6(b) displays
varying power consumption for the air conditioner, dishwasher, refrigerator, index
loads, and electric water heater without load scheduling. Fig. 7 shows the Household’s daily base-load profile analysis excluding load scheduling,
battery storage, and self-consumption. BESS Charge ranges from -1 kW to 0 kW from
7 to 24 hours, and EV charge/discharge starts at 3.5 kW and decreases to 0.35 kW over
25 hours. Fig. 8 presents the analysis results of AC operation based on set-point and pre-cooling.
In Subplot 8(a), AC usage begins at zero and increases to 0.01 by 25 hours. Subplot
8(b) shows AC temperature, starting at 23.5$^{\circ}$C and ending at 22.5$^{\circ}$C
by 25 hours, while the Minimum set-point temperature starts at 23$^{\circ}$C and decreases
to 22.5$^{\circ}$Cover the same time frame. Fig. 9 shows the analysis results of the temperature change inside the EWH tank. In Subplot
9(a), the EWH tank temperature remains constant at 45$^{\circ}$C from zero to 25 hours.
Subplot 9(b) displays EWH tank usage, starting at zero and increasing to 0.04 by 24
hours. Fig. 10 provides an analysis of temperature changes inside the refrigerator cabinet in the
proposed techniques. Subplot 10 (a) shows the refrigerator cabinet temp. The value
started at 3.8 $^{\circ}$C at zero hours and ended at 4.04 $^{\circ}$C at 25 hours.
Subplot 10 (b) shows the refrigerator cabinet usage. The value started at zero at
zero hours and ended at 0.21 at 24 hours.
Fig. 7. Household’s daily base-load profile analysis excluding load scheduling, battery storage, and self-consumption.
Fig. 8. Analysis of AC operation based on the pre-cooling and set-point: (a) usage; (b) temperature in the proposed technique.
Fig. 9. Analysis of the temperature change inside EWH tank: (a) temperature; (b) usage in the proposed technique.
Fig. 10. Analyses of the temperature change in the refrigerator cabinet: (a) Temperature; (b) Usage in the proposed technique.
Fig. 11. Load profile analyses of dissimilar household types (A).
Fig. 11 provides a Load profile analyses of dissimilar household types (A). Subplot 11(a)
shows variations in the water heater and air conditioner power usage. Subplot 11(b)
displays power usage for the washing machine, dishwasher, and clothes dryer. Subplot
11(c) illustrates power usage for the refrigerator and air conditioner. Subplot 11(d)
shows power fluctuations in the washing machine and dishwasher. Subplot 11(e) presents
changes in the water heater, refrigerator, and air conditioner power usage. Subplot
11(f) shows power variations for the washing machine, dishwasher, and clothes dryer.
Fig. 12 presents the load profile analyses of dissimilar household types (B). Subplot 12(a)
shows variations in power usage for the water heater, refrigerator, and air conditioner.
Subplot 12(b) illustrates power fluctuations in the washing machine, dishwasher, and
clothes dryer. Subplot 12(c) displays changes in the water heater, refrigerator, and
air conditioner power usage. Subplot 12(d) shows power variations for the washing
machine, dishwasher, and clothes dryer. Subplot 12(e) presents fluctuations in the
water heater, refrigerator, and air conditioner power usage. Subplot 12(f) shows power
changes in the washing machine, dishwasher, and clothes dryer.
Fig. 12. Load profile analyses of dissimilar household types (B).
Fig. 13. Analysis of the dynamic feed-in tariff (FIT) rates and modified RTP of Turkey in the proposed technique.
Fig. 14. Analysis of load profile of smart home using dynamic FIT and modified RTP rates (TCAs + TSAs + EV+ PV + BESS) in the proposed technique.
Fig. 15. Analysis of temperature change in the household: (a) Temperature; (b) Usage in the proposed method.
Fig. 16. Comparison of cost reduction for the proposed and existing approaches.
Fig. 13 depicts the Analysis of the dynamic feed-in tariff (FIT) rates and modified RTP of
Turkey in the proposed technique. The BESS buying prices start at ＄0.19/kWh and peak
at ＄0.20/kWh over 24 hours. The TOU rate ranges from ＄0.50/kWh at hour 0 to ＄0.09/kWh
at hour 25. EV selling prices vary from ＄0.15/kWh at hour 0 to ＄0.15/kWh at hour 23,
while EV buying and BESS selling prices are both -＄0.19/kWh at hours 0 and 25 and
-＄0.15/kWh at hours 0 and 25, respectively. Fig. 14 presents the Analysis of load profile of smart home using dynamic FIT and modified
RTP rates (TCAs + TSAs + EV+ PV + BESS) in the proposed technique. Subplot 14(a) illustrates
power usage in the water heater, refrigerator, and air conditioner. The water heater
load starts at 0.5 kW and ends at 0.6 kW over 24 hours, while the refrigerator power
ranges from 0.5 kW at one hour to 0.7 kW at 23 hours. The air conditioner usage varies
from 0.5 kW at one hour to 0.8 kW at 24 hours. Subplot 14(b) shows power fluctuations
in the washing machine, dishwasher, and clothes dryer, with varying values at different
hours.
Fig. 15 presents an analysis of the temperature change inside the household. Subplot 15 (a)
presents the temperature. The temperature in the refrigerator cabinet started at 3.8
$^{\circ}$C at zero hours after a continuous cycle; the value ended at 4.5 $^{\circ}$C.
The minimum cabinet temperature was constant. The maximum cabinet temperature was
constant. Subplot 15 (b) presents the data. The temperature started at ${-}$2 $^{\circ}$C
at zero hours and ended at ${-}$2.9 $^{\circ}$C at 25 hours.
Fig. 16 compares the cost reduction for the proposed and existing approaches. The PSO, RFA,
and BCO approaches varied from 0 to 38%, zero to 40.2%, and zero to 54%, respectively.
The proposed method varies from zero to 13%. The cost reduction of the proposed approach
is superior to the existing approaches. The proposed method provided a better result
than existing methods. Table 1 provides the Households Daily Base-Load Profile (excluding battery storage). The
table includes details such as the operation hours of PV production, an electric water
heater, an inflexible load, and self-generated PV power. For example, PV production
runs from 7 to 23 hours at 0.5 KW, while the electric water heater operates continuously
at 0.8 KW. The inflexible load runs from 11 to 13 hours at 2.5 KW, and self-generated
PV power is active from 12 to 13 hours at 2.6 KW. Table 2 presents an analysis of feed-in tariff rates and Time-of-Use (TOU) for various hours.
It includes the cost associated with BESS buying price, feed-in tariff, EV buying
price, EV selling price, TOU rate, and BESS selling price. The Table 2 illustrates how these rates vary throughout the day. For example, at hour 0, the
BESS buying price is 0.19, the feed-in tariff is 0.08, and the TOU rate is 0.09, with
similar variations for other hours. Table 3 presents a comparison table for proposed and existing methods. The proposed method
in user comfort and computational time was 8.7 and 136.8, respectively. The corresponding
data for the existing PSO, RFA, and BCO techniques were 8.1, 7.4, and 7.2 and 189.5,
227.5, and 324.3, respectively. Hence, the proposed method was better than the existing
techniques. Table 4 lists the Comparison of efficiency in the proposed and existing techniques. The efficiency
of the proposed technique is 95%. The existing PSO, RFA, and BCO techniques showed
values of 90%, 87%, and 79%, respectively. Therefore, the proposed method-based efficiency
was higher than existing techniques.
Table 1. Households Daily Base-Load Profile (excluding battery storage).
|
Appliance
|
Start Time
|
End Time
|
Power (KW)
|
|
PV Production
|
7 hours
|
23 hours
|
0.5KW
|
|
Electric Water Heater
|
1 hour
|
24 hours
|
0.8KW
|
|
Inflex Load
|
11 hour
|
13 hour
|
2.5 KW
|
|
PV (Self)
|
12 hour
|
13 hour
|
2.6 KW
|
Table 2. Analysis of feed-in tariff rates and Time-of-Use (TOU).
|
Time (hour)
|
Cost
|
|
BESS Buying Price
|
Feed in Tariff
|
EV Buying Price
|
EV Selling Price
|
TOU Rate
|
BESS Selling Price
|
|
0
|
0.19
|
0.08
|
0.16
|
-0.02
|
0.09
|
-0.06
|
|
10
|
0.24
|
0.08
|
0.2
|
-0.02
|
0.12
|
-0.06
|
|
20
|
0.24
|
0.08
|
0.26
|
-0.02
|
0.18
|
-0.06
|
|
24
|
0.19
|
0.08
|
0.16
|
-0.02
|
0.09
|
-0.06
|
Table 3. Comparison table for proposed and existing methods.
|
Solution Techniques
|
User comfort
|
Computational time
|
|
Proposed Technique
|
8.7
|
136.8
|
|
PSO
|
8.1
|
189.5
|
|
RFA
|
7.4
|
227.5
|
|
BCO
|
7.2
|
324.3
|
Table 4. Comparison of efficiency in the proposed and existing techniques.
|
Solution Techniques
|
Efficiency (%)
|
|
Proposed Technique
|
95%
|
|
PSO
|
90%
|
|
RFA
|
87%
|
|
BCO
|
79%
|
5. Conclusion
This study suggested a method to lessen the total daily electricity costs in households
using the EOSA-SNN approach. The proposed method scheduled all types of electrical
loads using a comprehensive system. The proposed method was tested using the MATLAB
platform and was compared with existing methods. Various scenarios, such as optimum
and random scheduling, and the intricate EOSA algorithm were used to assess the proposed
method. The proposed method incorporating a photovoltaic model, thermostatically controlled
appliances, electric water heater, electric vehicle, power-shiftable appliances, air
conditioner, and problem formulation effectively achieved cost minimization while
also considering the optimal demand response and solar self-consumption. Furthermore,
the proposed strategy outperformed other optimization techniques significantly. This
suggested that the EOSA-SNN approach may be a valuable tool for energy management
in smart home appliances, providing cost savings and efficient resource utilization.
The proposed technique efficiency becomes 95%. existing techniques, such as PSO, RFA,
and BCO, had efficiencies of 90%, 87%, and 79%, respectively. Finally, the efficiency
of the proposed method was higher than the existing techniques. However, in the future,
an optimal system design and a comprehensive techno-economic and life-cycle-cost analysis
(LCCA) will be incorporated into research on HEMS-operated houses. In the future,
it will be possible to investigate the cost reduction of energy consumption by combining
renewable energy sources and storage systems.
REFERENCES
I. Serban, S. Cespedes, C. Marinescu, C. A. Azurdia-Meza, J. S. Gomez, D. S. Hueichapan,
``Communication requirements in microgrids: A practical survey,'' IEEE Access, vol.
8, pp. 47694-47712, Mar. 2020.
Y. Saleem, N. Crespi, M.H. Rehmani, R. Copeland, ``Internet of things-aided smart
grid: technologies, architectures, applications, prototypes, and future research directions,''
Ieee Access, vol. 7, pp. 62962-63003, Apr. 2019.
R.A. Sanjeevkumar, P. Kavya, ``Optimization of DG Placement for Demand Side Management
in Distribution System with Demand Response,'' In 2021 IEEE PES/IAS PowerAfrica (pp.
1-5). IEEE. Aug. 2021.
M. Castaneda, S. Zapata, J. Cherni, A.J. Aristizabal, I. Dyner, ``The long-term effects
of cautious feed-in tariff reductions on photovoltaic generation in the UK residential
sector,'' Renewable Energy, vol. 155, pp. 1432-1443, Aug. 2020.
U. Zafar, S. Bayhan, A. Sanfilippo, ``Home energy management system concepts, configurations,
and technologies for the smart grid,'' IEEE access, vol. 8, pp. 119271-119286, Jun.
2020.
Z. Cheng, M.Y. Chow, ``Resilient collaborative distributed energy management system
framework for cyber-physical DC microgrids,'' IEEE transactions on smart Grid, vol.
11, no. 6, pp. 4637-4649, Jun. 2020.
[7] A.C. Duman, H.S. Erden, Ö. Gönül, Ö. Güler, “A home energy management system with
an integrated smart thermostat for demand response in smart grids,” Sustainable Cities
and Society, vol. 65, p. 102639, Feb. 2021.
G.R. Venkatakrishnan, R. Ramasubbu, R. Mohandoss, ``An efficient energy management
in smart grid based on IOT using ROAWFSA technique,'' Soft Computing, vol. 26, no.
22, pp. 12689-12702, Nov. 2022.
G.R. Venkatakrishnan, R. Rengaraj, R. Jeya Rajalakshmi, K.S. Viswanath, ``Real Time
Dynamic Home Surveillance Using Raspberry Node,'' In International Conference on Internet
of Things (pp. 14-24). Cham: Springer Nature Switzerland. Apr. 2022.
H. Anand, M. Rajalakshmi, G.R. Venkatakrishnan, R. Rengaraj, R. Jeya, ``Energy Bill
Minimisation of Dynamic Tariff Bound Residential Consumers by Intentional Load Shifting,''
In International Conference on Internet of Things (pp. 79-92). Cham: Springer Nature
Switzerland. Apr. 2022.
R.G. Babu, V. Amudha, C. Chellaswamy, K.S. Kumar, ``IoT based residential energy management
system for demand side response through load transfer with various types of domestic
appliances,'' Build. Simul. Aug. 2021.
K. Ramasamy, K. Chandramohan, D. Ghanta, ``Energy Management in Plugin Hybrid Electric
Vehicles with Hybrid Energy Storage System Using Hybrid Approach,'' Energy Technology,
vol. 10, no. 10, p. 2200355. Oct. 2022.
M.A. Ravindran, K. Nallathambi, ``A CryStAl-RDF technique-based integrated circuit
topology for fast charging station of electric vehicle (EV),'' Electrical Engineering,
pp. 1-14, 2023.
R. Saravanan, O. Sobhana, M. Lakshmanan, P. Arulkumar, ``Fuel cell electric vehicles
equipped with energy storage system for energy management: A hybrid JS-RSA approach,''
Journal of Energy Storage, vol. 72, p. 108646, 2023.
U. ur Rehman, K. Yaqoob, M.A. Khan, ``Optimal power management framework for smart
homes using electric vehicles and energy storage,'' International Journal of Electrical
Power & Energy Systems, vol. 134, p. 107358. Jan. 2022.
D. Bahmanyar, N. Razmjooy, S. Mirjalili, ``Multi-objective scheduling of IoT-enabled
smart homes for energy management based on Arithmetic Optimization Algorithm: A Node-RED
and NodeMCU module-based technique,'' Knowledge-Based Systems, vol. 247, p. 108762.
Jul. 2022.
[17] A.K. Erenoğlu, İ. Şengör, O. Erdinç, A. Taşcıkaraoğlu,
J.P. Catalão, “Optimal energy management system for microgrids considering energy
storage, demand response and renewable power generation,” International Journal of
Electrical Power & Energy Systems, vol. 136, p. 107714. Mar. 2022.
A. Mohammad, M. Zuhaib, I. Ashraf, ``An optimal home energy management system with
integration of renewable energy and energy storage with home to grid capability,''
International Journal of Energy Research, vol. 46, no. 6, pp. 8352-8366. May 2022.
M. Tostado-Véliz, S. Gurung, F. Jurado, ``Efficient solution of many-objective Home
Energy Management systems,'' International Journal of Electrical Power & Energy Systems,
vol. 136, p. 107666. Mar. 2022.
M.G. Villalva, J.R. Gazoli, E. Ruppert Filho, ``Modeling and circuit-based simulation
of photovoltaic arrays,'' In 2009 Brazilian Power Electronics Conference (pp. 1244-1254).
IEEE. Sep 2009.
O. Erdinc, N.G. Paterakis, T.D. Mendes, A.G. Bakirtzis, J.P. Catalão, “Smart household
operation considering bi-directional EV and ESS utilization by real-time pricing-based
DR,” IEEE Transactions on Smart Grid, vol. 6, no. 3, pp. 1281-1291. Sep. 2014.
H. Golmohamadi, R. Keypour, B. Bak-Jensen, J.R. Pillai, ``Optimization of household
energy consumption towards day-ahead retail electricity price in home energy management
systems,'' Sustainable Cities and Society, vol. 47, p. 101468. May. 2019.
O.N. Oyelade, A.E.S. Ezugwu, T.I. Mohamed, L. Abualigah, ``Ebola optimization search
algorithm: A new nature-inspired metaheuristic optimization algorithm,'' IEEE Access,
vol. 10, pp. 16150-16177. Jan. 2022.
C. Liu, W. Shen, L. Zhang, Y. Du, Z. Yuan, ``Spike neural network learning algorithm
based on an evolutionary membrane algorithm,'' IEEE Access, vol. 9, pp. 17071-17082.
Jan 2021.
C. P. Shirley received B.E. degree in CSE from C. S. I. Institute of Technology,
Thovalai affiliated to Anna University, Chennai, Tamil Nadu in 2002. M.Tech Degree
in CSE from Dr. M. G. R. Educational and Research Institute, Chennai, Tamil Nadu in
2009.PhD from Anna University, Chennai, Tamil Nadu in 2020. She is currently working
in the Department of CSE, Karunya Institute of Technology and Sciences, Coimbatore,
India. Her research interests include Image Processing, Machine Learning and AI.
Depaa RA B currently working as a Deputy Hod and Associate Professor, Department
of Civil Engineering, Dr. M.G.R Educational And Research Institute, Maduravoyal, Chennai
95, India
A. Priya currently working as an Associate Professor, School of Architecture, KoneruLakshmaiah
Educational Foundation, Guntur, Andhra Pradesh, India
R. Sarala currently working as an Assistant Professor, Department of Computer Science
and Engineering, Velammal College of Engineering and Technology, Madurai Rameshwaram
High Road, Viraganoor-625009, Madurai, India
Rajdeep Singh Solank currently working as an Assistant professor, Department of
Computer Science, Medi-caps University Indore, India
Malini K. V currently working as an Assistant Professor and HOD, Department of
Electrical and Electronics Engineering, Sri Sairam College Of Engineering Anekal,
India
Ch. Venkatakrishna Reddy received Doctorate in Electrical and Electronics Engineering
from JNTUH, Hyderabad in 2021. His research areas are Power systems and Renewable
energy systems. His reasech is published in 17 international Journals, 6 international
conferences and 4 national conferences. Presently he is working as Assistant Professor
in the Department of EEE in Chaitanya Bharathi Institute of Technology, Hyderabd,
India.