Journal of Electrical Engineering & Technology (2024) 19:2859–2874

https://doi.org/10.1007/s42835-023-01734-2

ORIGINAL ARTICLE

Optimal Sizing Strategy and Economic Analysis of PV‑ESS for Demand

Side Management
Kyeong‑Hee Cho1 · Jongyul Kim1 · Gilsung Byeon1 · Wanbin Son1

Received: 11 September 2023 / Revised: 14 November 2023 / Accepted: 22 November 2023 / Published online: 21 December 2023
© The Author(s) under exclusive licence to The Korean Institute of Electrical Engineers 2023

Abstract
We propose a method to determine the optimal capacity of a photovoltaic generator (PV) and energy storage system (ESS)
for demand side management (DSM) and review its economic revenues. The calculation procedure for determining the opti-
mal capacity of PV-ESS is complicated because it includes the estimation of load and power generation patterns, selection
of candidate capacities, simulation of PV-ESS operation, calculation of revenue and cost, economic analysis, and selection
of optimal capacity. Therefore, we address and generalize the methods which are the entire procedure for selecting the
optimal capacity of PV-ESS, the method for selecting the initial candidate capacity of PV-ESS and the method for selecting
the optimal capacity of PV-ESS. In numerical simulation, we applied this methodology to four real industrial customers in
South Korea. When we analyzed the optimal capacity results calculated for customers with different load patterns, the ESS
c-rate differed depending on the load pattern. Under the simulation conditions, the economic efficiency tended to improve
when there was a peak reduction, the ESS capacity was small, and the PV capacity was large. This is because the peak rate
is high, and the cost of PV-ESS is expensive compared to the electricity rate. Under the current electricity rate and PV-ESS
cost conditions, we determined the optimal capacity of PV-ESS for DSM of actual industrial customers. We confirmed that
it is an effective method that can be applied generally.

Keywords Energy storage system · Photovoltaic generator · Optimal sizing strategy · Economic analysis

1 Introduction is needed. Therefore, it can compensate for the intermittent

power output of renewable energy and prevent constructing
Globally, renewable energy targets are rising in line with additional transmission and distribution lines [2]. Sodium,
energy transition policies, with the EU recently increasing all-solid-state, and redox flow batteries, in addition to fire-
its renewable energy target to 42.5% by 2030 [1]. With the prone lithium-based ESS, have been steadily developed
increase in renewable energy, the importance of energy stor- and commercialized, thus improving the safety of ESS,
age system (ESS) in the power system is increasing to coun- and according to BloombergNEF(BNEF), the domestic and
teract the variability of renewable energy. ESS is a system international ESS markets are expected to continue to grow.
that can store electrical energy, which can be charged when The global ESS market is also expected to expand rapidly
surplus generation occurs and discharged when stored power [3].
This study attempts to develop a design technique
* Kyeong‑Hee Cho for photovoltaic generator (PV)-ESS that connects PV
kx1004xh@keri.re.kr resources and ESS. PV resources are expected to be
Jongyul Kim installed the most among renewable energy sources in
jykim@keri.re.kr South Korea [4, 5], and ESS can mitigate the volatility
Gilsung Byeon of PV power generation. The PV-ESS business model is
gsbyeon@keri.re.kr similar to the microgrid business model, which can effi-
Wanbin Son ciently manage the energy rate of business sites and pro-
wanbin@keri.re.kr vide a flexibility in emergencies. Because ESS is expen-
sive and requires a separate installation site, the need
1
Energy Platform Research Center, Korea Electrotechnology and importance of optimal capacity selection techniques
Research Institute, Gwangju, Republic of Korea

13
Vol.:(0123456789)
2860 Journal of Electrical Engineering & Technology (2024) 19:2859–2874

are increasing. Therefore, studies have been conducted Based on the comprehensive analysis of the previous stud-
on the capacity selection methods of various distributed ies, we have found a non-convexity problem in finding the
resources according to the load type of the customer where optimal capacity considering the trade-off between operating
they are to be installed. revenue and total cost in resource design, and candidate capac-
A previous study examined ESS capacity design for edu- ities are selected by calculating all combinations or assum-
cational service customers [6]. Here, candidate capacities ing several scenarios. However, there is no clear rationale for
were selected based on savings targets to design the capacity. selecting these scenarios. To overcome these weaknesses, this
A study was also conducted on the calculation of the optimal study includes a method for selecting the initial and candidate
ESS size, including reliability constraints of microgrids [7]. capacities that can maximize the operating profit.
Another study investigated the optimal capacity determi- The PV-ESS capacity design procedure proposed here
nation of EV and PV for residential loads [8]. A study on differs from the previous studies in the following ways.
the estimation of optimal PV-BESS capacity was conducted First, the PV-ESS optimal capacity design procedure
from an economic and environmental perspective [9]. A includes load pattern forecasting, PV modeling, operation
study was conducted on the optimal design of distributed scheduling optimization functions for demand side manage-
resources considering local energy networks [10]. Previous ment (DSM), and economic index calculation. Furthermore,
studies have investigated optimization methods to find the each concept is explained, and the overall flow is introduced.
optimal point between the operating revenue generated by Therefore, the procedures and formulas can be organized for
the installation and the overall cost required when designing universal application.
resources. MIP, MILP, and PSO algorithms were mainly Second, we proposed an algorithm for selecting an initial
employed as optimization functions. Candidate capacities capacity and candidate capacities of PV-ESS, which is the
were selected based on savings targets or scenarios. most important for optimal capacity design. This algorithm
Furthermore, there are studies on how to solve non-con- has a step-by-step process: the initial capacity selection
vexity problems in optimization. A study proposed a meth- algorithm, followed by the ESS and PV capacity selection
odology for jointly optimizing the sizing and power man- algorithms. In other words, this study introduces a method
agement of PV household-prosumers, namely, PV power, for designing the capacity in sequences to maximize profit
electric vehicle charging load (EVCL), household consump- when designing a combination of two or more resources.
tion load (HCL), battery bank (BB), and power converters Third, we simulated the optimal capacity selection for
[11]. The total cost and revenue were calculated with the each load pattern utilizing the proposed strategy for four real
teaching–learning-based optimization (TLBO) algorithm, industrial (manufacturing) sites and analyzed the economic
assuming the expected scenario. efficiency results.
There was a study that provided a methodology to assess The reason for applying to industrial (manufacturing)
the techno-economic performance of PV household-pro- sites is that industrial electricity usage accounted for 53% of
sumers that jointly provide self-consumption/sufficiency the total electricity usage in South Korea in 2021, of which
enhancement (SCSE) and frequency curtailment reserve manufacturing usage accounted for 49%, the highest share
(FCR) [12]. The optimal capacity for the converter-BESS [15].
configuration is determined based on the techno-economic Section 2 describes the procedure and necessary tech-
assessment. For all combinations, the capacity that mini- niques for estimating the optimal capacity of the PV-ESS.
mizes the levelized cost of the PV household prosumer was In Sect. 3, numerical simulation is performed with the pro-
determined while considering the household electric load, posed procedure and equations. The optimal capacity is
state-of-charge (SOC), and life cycle. calculated with actual loads and conditions in four manu-
There is a study on the optimal selection of capacitor facturing sites, and the economic efficiency of PV-ESS is
banks in electrical AC distribution systems for minimiz- analyzed according to customer load patterns and changes
ing the cost of energy losses based on a discrete version of in local conditions. Section 4 discusses the contributions of
the vortex search algorithm (DVSA) [13]. The initial value the proposed method and suggests improvements to promote
of the DVSA is the center of the minimum and maximum the widespread utilization of PV-ESS.
bounds of the decision variables, and a Gaussian distribution
is employed to generate candidate values.
A study addressed the optimal reconfiguration problem 2 Methodology
of DC distribution networks by proposing a novel mixed-
integer nonlinear programming (MINLP) formulation [14]. 2.1 Optimal Sizing Procedure of PV‑ESS
The general algebraic modeling system (GAMS) software
was utilized to analyze the power loss and voltage for several This study proposes an optimal sizing procedure for PV-
scenarios to select the optimal scenario. ESS for customers who use the time-of-use electricity tariff

13
Journal of Electrical Engineering & Technology (2024) 19:2859–2874 2861

linked to the grid. PV and wind turbine generators (WT)

are the most installed renewable energy sources in Korea;
however, we focus on PVs because WTs are economically
unviable due to insufficient wind speed. Figure 1 shows a
grid connection diagram of the target PV-ESS connected
to a customer.
By installing a PV-ESS, consumers can use energy using
DSM to reduce electricity bills [16]. Here, DSM referred to
installing PV-ESS in consumer residences to supply elec-
tricity to consumers with PV generation amount, reduce
electricity bills, and control demand using an ESS. In other
words, the ESS can be charged during low-rate hours and
discharged during high-rate or peak hours to reduce electric-
ity bills.
Figure 2 shows the necessary technologies for optimal
sizing of PV-ESS. The required technologies include ana-
lyzing load patterns at target workplaces, PV modeling
technology, selecting candidate capacities for PV and ESS,
establishing optimal operation plans, calculating economic Fig. 2  Necessary technology for optimal sizing of PV-ESS
indices by determining profits and costs for each candidate
capacity, and selecting the optimal capacity.
The flowchart shown in Fig. 3 summarizes the PV-ESS monthly working days, holidays, Saturdays, and peak load
optimal sizing procedure and the interaction of the technolo- days using the load data for the past year. If past load data
gies described in Fig. 2. This procedure is explained in detail cannot be collected, load patterns should be estimated based
in the following subsections. on the workplace, electronic device specifications, and usage
time.
2.2 Necessary Techniques for PV‑ESS Optimal Sizing In this study, the average monthly load pattern (hereaf-
ter referred to as the average load pattern) by working type
This section explains the concepts and methods for load pat- was calculated by classifying working days, holidays, and
tern analysis, photovoltaic pattern analysis, scheduling opti- Saturday monthly patterns using load data for the past year.
mization, revenue and cost calculation, economic analysis, The PV-ESS operation plan was simulated using the aver-
and optimal capacity selection. They are necessary for the age load pattern, and the one-year profit was calculated by
PV-ESS optimal capacity calculation. multiplying the number of monthly days by the work type.

2.2.1 Load Pattern Analysis

The load prediction value obtained through load pattern

analysis can be used to calculate the average load pattern of

Fig. 1  PV-ESS system for DSM Fig. 3  Optimal sizing procedure of PV-ESS

13
2862 Journal of Electrical Engineering & Technology (2024) 19:2859–2874

2.2.2 Photovoltaic System Modeling can be reduced using the PV supplying load to consum-
ers, and the ESS can be charged during off-peak hours
PV modeling predicts PV power generation using the PV and discharged during peak hours to earn a profit differ-
specifications to be installed and the operational data in ence. The ESS output has a negative value when the ESS
the presence of the PV. In general, the rated output, panel is charging and a positive value when discharging. The
inclination angle, and power conditioning system (PCS) profit difference is calculated by adding the total charge
efficiency can be set for the PV specifications. By setting and discharge hours.
the amount of solar radiation, wind speed, cloud speed, and Additionally, peak reduction profits can be obtained by
south-central altitude, which are the meteorological informa- discharging the ESS or generating power using a PV sys-
tion of the installation [17], the amount of PV power genera- tem during peak hours. Therefore, the PV-ESS revenue is
tion can be predicted following (1) and (2): calculated by adding the PV energy charge saving, ESS
Meteorological information can be collected directly energy charge saving, and ESS peak reduction revenues.
using sensors, or measurement information from a nearby Each revenue can be determined using (3)–(5).
meteorological agency may be used.
( ) ∑
24
( )
PPV = CPV × SRdh × cos 𝜃 × 𝜂PV (1)
RPV
m = PPV
mh × 𝜌mh × Nm (3)
dh
h=1

∑ ∑
Nm 24
∑ ∑(
Nm 24
)
PPV = PPV (2) RESS PESS (4)
m = mdh × 𝜌mh
mh dh
d=1 h=1
d=1 h=1
[ ]
where SRdh : Solar irradiance at d-day at h-hour W∕m2 ; 𝜃 :
Declination angle of PV panel [°]; ηPV ∶ PV PCS efficiency RPeakcut
m
= PPeakcut
m
×𝜀 (5)
[%]; CPV ∶ Capacity of PV[kW]; PPV dh
: Generation power of
PV at d-day at h-hour [kW]; PPV : PV average generation at where RPV m : Revenue of PV at m-month [KRW]; 𝜌mh: Energy
mh
m-month at h-hour [kW]; Nm : Number of days in m-month. charge unit price at m-month at h-hour [KRW/kWh]; RESS m :
Revenue of ESS at m-month [KRW]; PESS mdh
: ESS output
power at m-month, d-day and h-hour [kW]; 𝜀 : Demand
2.2.3 Optimization of PV‑ESS Scheduling for DSM charge unit price [KRW/kW]; PPeakcut
m : Peak-cut at m-month
[kW]; RPeakcut
m : Peak-cut revenue at m-month [KRW].
The schedule optimization function for the PV-ESS opera- The PV energy charge-saving revenue is calculated by
tion simulation can be expressed as follows [18]: multiplying the electricity rate unit price by the power
generation forecast value described in Sect. 2.2.2. The ESS
• Objective function: Minimize electricity rate energy charge-saving revenue is calculated by multiplying
• Variable: ESS charging and discharging schedule the ESS charging and discharging amounts by the electric-
• Constraint: PCS output range, power consumption limit, ity rate unit price, assuming that it is actually operated
SOC range using the operation plan values simulated in Sect. 2.2.3.
The peak reduction revenue is calculated by multiplying
The objective function maximizes the PV-ESS operating the peak reduction amount by the demand rate unit price.
profit, and the constraints include the PCS power range of Here, the peak reduction revenue can be obtained even
the ESS, which is a controllable resource, the SOC upper with the amount of PV generated. However, the amount
and lower limits, and the load upper and lower limits. of solar radiation, which is a supply resource, cannot be
The optimization was performed using a nonlinear gen- adjusted. Therefore, PV is excluded from the peak reduc-
eralized reduced gradient method owing to the nonlinear tion revenue.
nature of the ESS charging and discharging efficiencies. Monthly revenue can be calculated by adding the hourly
revenue of the day and multiplying it by the number of
monthly days, yearly revenue can be calculated by adding
2.2.4 DSM Revenue and Cost Calculation all the monthly revenues. For the economic analysis for
20 years, considering the capacity degradation rate using
In this section, a method for calculating the PV-ESS revenue the first-year revenue, revenues for 20 years are calculated,
and DSM costs is described. as shown in (6) and (7). In general, the PV capacity deg-
The structure of electricity tariffs in South Korea com- radation rate was calculated as 0.7% per year, and the ESS
prises demand charge and energy charge. Electricity bills capacity degradation rate was calculated as 2% per year.

13
Journal of Electrical Engineering & Technology (2024) 19:2859–2874 2863

∑
12
( PV ) ∑
20
CFn
RDSM = Rm + RESS Peakcut
m + Rm (6) 0= − ECtotal (10)
y
m=1 n=1
(1 + IRR)n

� � where p : Rate of change of electricity bill [%]; CFn : Pre-

⎛ RPV × 1 − DF PV ∕100) × (n − 1) + ⎞ currentization revenue for N-year [KRW]; n: Operating year;
� ⎜ m�
12
� � � ⎟
IRR: Internal rate of return [%].
ESS
RDSM
n
= ⎜ RESS
m
+ RPeakcut
m
× 1 − DF
100 ⎟ (7)
m=1 ⎜ ⎟
⎝ ×(n − 1) ⎠

where RDSM : DSM revenue for the first year [KRW]; RDSM : 2.2.6 Optimal Capacity Selection
1 n
DSM [ revenue] for n years [KRW]; DF : PV degradation
PV

rate %∕year ; DF ESS : ESS degradation rate [%∕year]. The most challenging part of the PV-ESS optimal sizing
Subsequently, a method for calculating the PV cell and strategy is the algorithm for selecting the appropriate ini-
ESS costs is described. The cost is divided into facility and tial and candidate capacities and searching for the optimal
installation costs. Moreover, the cost is calculated by multi- capacity. We have encountered various load patterns through
plying the capacity by the unit price, assuming that the cost several years of design experience and designed PV-ESS.
is proportional to the capacity. The unit price per MW for the Therefore, based on this experience, we summarize the fac-
PV, PCS, and battery (BAT) is shown below. Here, the unit tors that must be considered to calculate the optimal capacity
price of BAT uses lithium-ion batteries. For DSM facilities, of a PV-ESS and explain how to select the initial capacity
operation and maintenance (O & M) costs are assumed to and candidate capacity algorithmically. In addition, we pro-
be 1% of the total cost. pose a method for determining the optimal capacity using
the calculated candidate capacities. In this paper, when
• ECPV = 1,200,000,000 [KRW/MW] selecting the initial capacity, the PV is not considered and
• ECESSPCS = 230,000,000 [KRW/MW] is calculated using the ESS. The reason is that the PV system
• ECBAT = 400,000,000 [KRW/MWh] cannot control its supply resources, so its output is uncertain.
When selecting the ESS candidate capacity, an appropri-
Total cost is calculated that the unit price per MW for PV, ate capacity must be selected considering the load pattern of
PCS, and BAT multiplied by the capacity is shown in (8). In the customer. The ESS can shift the load from off-peak load
addition, maintenance costs can be added as 1% of the total times to peak-load times through charging and discharging
cost for demand management facilities. and can enable peak shaving by discharging at peak loads.
( ) ( Thus, the amount of load shifting or peak shaving may vary
)
total CPV × ECPV + CESSPCS 𝛼 depending on the load pattern.
EC = × 1+
×EC ESSPCS + CBAT × EC BAT
100 For example, when analyzing high-voltage option (A) for
(8) industrial use among Korea’s electricity rates, the revenue
where ECPV : Equipment cost of PV [KRW/MW]; CESSPCS : price obtained when shifting the load is up to 135 KRW/
Capacity of PCS[kW]; ECESSPCS : Equipment cost of ESS kWh per hour. Moreover, the revenue price obtained when
PCS [KRW/MW]; CBAT : Capacity of battery[kWh]; ECBAT : reducing the peak load is 8320 KRW/kW per month. The
Equipment cost of Battery [KRW/MWh]; α : Management Korean electricity rate system has a large revenue unit price
cost rate [%]; ECtotal : Equipment cost of PV-ESS [KRW]. when the peak load is reduced, and the rate is applied for one
year. Therefore, the initial capacity is calculated by maxi-
2.2.5 Economic Analysis mizing the amount of peak reduction possible. The initial
capacity calculation method of ESS to maximize peak reduc-
The internal rate of return (IRR) was calculated to compare tion is shown in Fig. 4.
and analyze the capacity of each economic feasibility. The Monthly peak load(PL) refers to the highest load during
IRR is determined by calculating the pre-commercialization load pattern, and the yearly PL refers to the highest load
revenue for each year using (9) and subsequently incorporat- among the monthly PLs. The yearly PL can be calculated
ing the total cost into (10). In other words, the IRR index using (11).
refers to the annual rate of return and is used as a standard ( )
peak
for determining optimal capacity. PLy = max Lm,h
m = 1 ∼ 12 (11)
CFn = RDSM × (1 − p)n (9) h = 1 ∼ 24
n
where PLy : Yearly peak load [kW]; Lm,h : Load of peak day
peak

at M-month at h-hour [kW].

13
2864 Journal of Electrical Engineering & Technology (2024) 19:2859–2874

ESS capacity can be calculated using (16) and (17) using

the CD and PL patterns. This is expressed as an algorithm
as shown in Fig. 5.
8 ( )
∑ peak
CESSPCS,UL = min CD −Lm,t )∕Toffpeak (16)
m=1∼12
t=1

Fig. 4  Calculation process of initial capacity 8 (

∑ )
peak
CESSBAT,UL = min CD −Lm,t )∕DOD (17)
m=1∼12
t=1
Subsequently, the yearly PL can be used to calculate the
where CESSPCS,UL : Upper limit of PCS capacity [kW];
demand charge. In particular, the lower limit of the peak load
CESSBAT,UL : Upper limit of battery capacity [kWh]; Toffpeak :
is 30% of the contract demand (CD). The available peak cut
Period of off-peak in the morning [hour].
(APC) is calculated based on the following condition (12):
Here, if (15) is larger than (17), APC cannot be sat-
If PL > CD × 0.3 isfied, so the BAT initial capacity must be recalculated
APC = PLy − CD × 0.3 by lowering the peak reduction amount. Additionally, the
larger value between (14) and (16) should be selected as
else (12)
the initial PCS capacity. This is expressed as an algorithm
APC as shown in Fig. 6.
end The initial capacity calculated using Algorithm Figs. 4,
5 and 6 becomes the candidate capacity. The ESS initial
where CD: Contract Demand [kW]; APC: Available Peak
capacity was calculated using the algorithm shown in
Cut [kW].
Fig. 7 to simulate the possibility of peak load reduction.
The upper limit of demand power (DPUL) is the value
If an optimal solution cannot be found, the peak-shaving
obtained by subtracting the APC from the yearly PL in (13).
test is repeated while increasing the battery capacity; if an
The PCS initial capacity can be calculated using the calcu-
lated DPUL and PL patterns in (14). Moreover, the initial
battery capacity can be calculated by adding the average load
patterns on workdays, Saturdays, and holidays to the load
exceeding the DPUL in (15).
DPUL = PLy − APC (13)

peak
If Lm,t > PUL
Fig. 5  Calculation process of upper limit capacity
( )
peak
CESSPCS,ini = max Lm,t − DPUL
m = 1 ∼ 12 (14)
t = 9 ∼ 22
( 22 )
∑ ( peak )
CESSBAT,ini = max Lm,t − DPUL ∕DOD (15)
m=1∼12
t=9

End
Where DPUL: Upper limit of demand power
[kW]; CESSPCS,ini : Initial value of PCS capacity[kW];
CESSBAT,ini : Initial value of battery capacity[kWh];
DOD = SOCmax − SOCmin[%].
If the yearly PL is low and no peak reduction capacity
exists, there may be no initial capacity of ESS. In this
case, the maximum capacity of the ESS can be calculated
to maximize load shifting. Therefore, the upper limit of
Fig. 6  Check algorithm of upper limit of capacity

13
Journal of Electrical Engineering & Technology (2024) 19:2859–2874 2865

Fig. 7  Test process of candidate capacity

optimal solution is found, this solution is selected as the

candidate capacity.
The initial capacity is the candidate capacity to opti-
mize the schedule for the average load pattern. If the opti-
mal solution cannot be found, the battery capacity and
number of retests are increased. If the optimal solution
is found, the candidate capacity has the smallest capac- Fig. 8  Procedure for selecting the optimal PV capacity
ity while maximally reducing the peak. By changing the
peak reduction amount, the candidate capacity can be cal- 3 Numerical Simulation
culated, and the economic feasibility can be analyzed to
determine the most economical capacity. In this section, we analyze the actual load patterns of four
This approach described above is required because the Korean industrial (manufacturing) customers and perform
economic indices of the PV-ESS are minimal to determine numerical simulations using the technologies previously
the optimal capacity considering the reduction in energy explained in Sect. 2 to calculate the optimal capacity of
charges. Moreover, the reason for selecting the candidate the PV-ESS for DSM for each customer and analyze the
capacity is to determine whether the peak can be reduced. economic feasibility. An Excel program was used for the
Figure 8 shows the algorithm used to select the opti- simulation of per pattern, the number of variables is 24 to
mal PV capacity. The maximum PV capacity that can be use to be optimized in each simulation and the optimization
installed for a customer must be considered when selecting method used Nonlinear generalized reduced gradient due to
the customer. Considering the area available for PV instal- the nonlinear characteristics of ESS charging and discharg-
lation and investment costs, we determine the upper limit ing efficiency.
of the PV capacity such that the power generation output
relative to the load does not exceed that required to prevent • Optimization method: Nonlinear generalized reduced
backflow during DSM. gradient
Furthermore, an economic analysis was performed • Program: Microsoft Office Professional Plus 2016 Excel
while increasing the PV capacity at regular intervals up to
the upper limit for the target consumers. The IRR index is First, the average load pattern is calculated, and each load
used to determine economic feasibility, and when the IRR pattern is analyzed.
index increases, the IRR is calculated by increasing the
maximum PV capacity. In the case of a convexity graph,
where the IRR increases and then decreases, the largest 3.1 Pattern Forecasting
IRR is selected as the optimal PV capacity. In contrast,
in the case of a non-convexity graph, the time when the Table 1 presents the primary information of the four indus-
IRR increase rate is the highest is selected as the optimal trial customers. The main information includes the installa-
PV capacity. tion position of each costumer, contracted power, rate plan,

13
2866 Journal of Electrical Engineering & Technology (2024) 19:2859–2874

Table 1  Information regarding the four actual customers

Customer no. 1 2 3 4

Target city Chungcheongnam-do Seoul

Contract demand [kW] 6100 4500 4000 3100
Tariff type Industrial (B) High-voltage
(A) Option II
Average load [kWh] 3103 1026 396 829
Rate applied power [kW] 4118 1488 1200 1459
Electricity bill [million KRW/year] 3849 808 473 813
PV upper limit [kW] 300 500 800 500

Fig. 10  Customer 1 load pattern on peak days each month

Fig. 9  Customer 1 load pattern (August)

Fig. 11  Customer 2 load pattern (August)

average load, applied power, one-year electricity rate, and
PV upper limit.
The hourly load data for one year stored on the power
company site was available. This data was used to obtain
the monthly average load pattern, and the average load pat-
terns for workdays, Saturdays, and holidays were calculated
for the simulation input [19]. To analyze the power usage
characteristics of each consumer, the average load patterns
for working days, Saturdays, and holidays in August were
drawn. The results are shown in Figs. 9, 10, 11, 12, 13, 14,
15 and 16.
As shown in Fig. 9, consumer 1 has flat and large load
patterns, which facilitate the load supply with PV power
generation. The base load value shows a flat load pattern,
similar to that of the average load. Therefore, the peak load Fig. 12  Customer 2 load pattern on peak days each month
during load shifting could be increased.
As shown in Fig. 10, the annual peak load is 4122 kW,
higher than 1830 kW, which is 30% of the contract demand, Therefore, the expected peak reduction is slight, and thus,
and is advantageous for obtaining peak reduction revenues. the revenue is expected to be small.
As shown in Fig. 11, the workday load pattern of con- As shown in Fig. 13, in the workday pattern of consumer
sumer 2 shows a pattern that increases sharply at 9:00 AM 3, the load is high during working hours and drops slightly
and then decreases at 12:00, 15:00, and 20:00. On Saturdays, during lunch hours. Moreover, the load is rarely used on Sat-
employees work only in the morning, and on holidays, they urdays and holidays. As shown in Fig. 14, the peak load of
rarely work. As shown in Fig. 12, the peak load is 1492 kW, 888 kW is lower than 1200 kW, 30% of the contract demand.
similar to 1350 kW, which is 30% of the contract demand. Therefore, no peak reduction in revenue is expected.

13
Journal of Electrical Engineering & Technology (2024) 19:2859–2874 2867

Fig. 13  Customer 3 load pattern (August) Fig. 16  Customer 4 load pattern on peak days each month

Table 2  PV specifications

Spec. Value Unit

Angle of inclination 50 Degree

PCS efficiency 95 %
PV system type Stationary
Panel direction South
Life 20 Years

3.2 Common Specifications and Parameters

Fig. 14  Customer 3 load pattern on peak days each month This section describes the photovoltaic specifications,
ESS parameters, unit price of an electricity bill, and unit
cost of equipment for installing PV-ESS to be entered dur-
ing operational simulations. The PV generation patterns
are calculated by assigning the specifications for the PV
system type, PCS efficiency, and angle of inclination as
shown in Table 2. A target area is selected, and the past
solar irradiance data is stored [20]. Using the specifica-
tions in Table 2 and historical data from the target area, the
daily and average monthly power generation are predicted
and used in operational simulations.
ESS specifications vary according to the type of bat-
tery. This paper focuses on lithium-ion batteries, and the
specifications of the entered lithium-ion batteries are sum-
marized in Tables 3 and 4. Here, the lithium-ion battery
cycle efficiency indicates that the battery decreases to 80%
Fig. 15  Customer 4 load pattern (August) of its rated capacity when charged and discharged 3,500
times. The charge/discharge efficiencies of the lithium ion
batteries in Table 4 were linearized and applied during
The workday load pattern of customer 4 in Fig. 15 charge/discharge.
shows that the load during working hours is high, imply- The DSM revenue was calculated using the electricity
ing that some work is performed on Saturdays, and almost unit price and considering the industrial (B) high-voltage
no work is performed on holidays. As shown in Fig. 16, (A) option (II) type commonly used by the target consum-
the peak load is 1460 kW, higher than 930 kW, which is ers in Table 1.
30% of the contract demand. Therefore, this customer can The energy charge unit price of the industrial (B) high-
receive peak reduction revenues. voltage (A) rate type is shown in Table 5, and the seasonal

13
2868 Journal of Electrical Engineering & Technology (2024) 19:2859–2874

Table 3  ESS specifications

Spec. Value Unit

Battery type Lithium-ion battery

Euro efficiency 95 %
Cycle efficiency 3500 Cycles
80 %
Battery usage year 0 Year
Minimum SOC 20 %
Maximum SOC 90 %
Initial SOC 20 %
Life 10 years

Table 4  Battery efficiency

Fig. 17  Seasonal energy charge unit price graph of industrial (B)
Discharge duration [h] Discharge current [C] Efficiency [%] high-voltage (A) option (II)
5 0.2 100
1 1 99 Table 6  Demand charge table for industrial service (B) high-voltage
0.2 5.2 98 (A)
0 10 95 Classification Energy
charge
[KRW/kW]
Table 5  Energy charge table for industrial service (B) high-voltage High-voltage (A) Option I 7,220
(A)
Option II 8,320
Classification Energy charge [KRW/kWh] Option III 9,810
Time Summer Spring/fall Winter
period
Table 7  Economic parameters
High-volt- Option I Off-peak 92.8 92.8 99.8
age (A) Value Unit
Mid-peak 145.7 115.3 145.9
On-peak 227.8 146.0 203.4 Currency discount rate 4.5 %/year
Option II Off-peak 87.3 87.3 94.3 Electricity bill increase rate 3 %/year
Mid-peak 140.2 109.8 140.4 O & M cost ratio 1 %/year
On-peak 222.3 140.5 197.9 PV power degradation 0.7 %/year
Option III Off-peak 86.4 86.4 93.7 ESS power degradation 2 %/year
Mid-peak 139.6 108.5 139.8
On-peak 209.9 132.2 186.7

Table 8  Unit price of equipment cost

energy charge unit price graph for Option II is shown in Value Unit
Fig. 17 [18]. PV cost (included PCS cost) 1200 KRW/MW
Here, industrial (B) indicates that the contract demand of ESS battery cost 400 KRW/MW
the customer is 300 kW or more. Similarly, high-voltage (A) ESS PCS cost 120 KRW/MW
indicates that the supply voltage of the customer is greater
than or equal to 3300 V and less than or equal to 66,000 V
[16]. Additionally, the basic unit price of the industrial (B)
high-voltage (A) rate plan is shown in Table 6. rate, electric unit price increase rate, power degradation
The DSM revenue was calculated by applying the energy rate per resource, and O & M cost rate.
charge rate and demand charge rate unit prices in (3)–(5). Additionally, the PV-ESS investment cost was calcu-
The parameters used for calculating the economic feasi- lated using the cost unit price per resource in Table 8. The
bility index are shown in Table 7. The parameters used to unit cost is the unit price including the installation and
calculate the NPV and IRR include the currency discount facility costs, and is assumed to increase with capacity.

13
Journal of Electrical Engineering & Technology (2024) 19:2859–2874 2869

3.3 Results of PV‑ESS Sizing From initial capacity, the ESS candidate capacity was cal-
culated by randomly gradually reducing the peak reduction
In this section, we describe the results of calculating the amount. The economic feasibility determined by operational
optimal PV-ESS capacity using the proposed algorithm. simulations is shown in Table 9. Assuming the lifespan of
The initial and candidate capacities were selected for each the ESS to be 10 years, the candidate capacity in Case 1–5
customer using the algorithm in Figs. 4, 5, 6 and 7, and the was not economically feasible. Therefore, PCS and battery
economic feasibility was calculated using (3)–(10). of 50 kW and 230 kWh, respectively, were determined as the
First, the economic results according to the calculated optimal capacity of ESS and bold in the Table 9. Computa-
candidate capacity for Consumer 1 are shown in Table 9. tion time for Case 6 is 71 s.
According to (12), the maximum APC for Customer 1 was In the case of Customer 1, economic feasibility was not
2464 kW and the DPUL calculated using (13) was 1830 kW. when there was no peak reduction or when there was large-
The initial capacity calculated using (14) and (15) was PCS peak reduction. Even if the peak cut is reduced, a smaller
2292 kW. BAT was 41,633 kWh. According to calculations ESS capacity indicates better economic feasibility. This may
of the upper limit of the ESS capacity using (16) and (17), be attributed to the fact that DSM revenues are small com-
PCS and BAT were 2455 kW and 28,059 kWh, respectively. pared to ESS unit prices.
A comparison indicates that the initial capacity that satisfies The maximum APC for Customer 2 was calculated
the maximum APC exceeds the upper limit of ESS capacity; as 142 kW, which required a battery capacity of at least
thus, the peak reduction amount cannot be satisfied. 660 kWh. The minimum ESS candidate capacity required
Therefore, when the APC was lowered to 1000 kW and for reducing the peak was calculated and selected as the
DPUL was increased to 3294 kW while calculating the ini- candidate capacity. An economic feasibility analysis was
tial capacity, the PCS and BAT capacities were estimated as performed, as shown in Table 10. In the case of Customer
2156 kW and 18,500 kWh, respectively, which were satisfied 2, peak reduction was required for economic feasibility.
below the ESS upper limit. Smaller peak reduction amounts corresponded to better

Table 9  Economic analysis results of Customer 1 for installing ESS

Case Capacity Assumption Investment cost [100 Million Revenue [100 Million KRW] Economic index
Num- KRW]
ber
PCS [kW] Battery [kWh] Peak reduction [kW] Equipment cost O&M cost Load shifting Peak shaving IRR [%] Payback
time
[year]

1 2156 18,500 1000 78.96 15.79 30.4 12.01 – > 20

2 1000 8000 500 34.30 6.86 15.01 6.01 – > 20
3 450 3700 300 15.84 3.17 7.25 3.60 – > 20
4 250 2000 200 8.58 1.72 4.01 2.40 – > 20
5 100 700 100 3.03 0.61 1.48 1.20 – > 20
6 50 230 50 1.04 0.21 0.55 0.60 2.36 9.90

Table 10  Economic analysis results of Customer 2 for installing ESS

Case Number Capacity Assumption Investment cost [100 Mil- Revenue [100 Million Economic index
lion KRW] KRW]
PCS [kW] Battery [kWh] Peak reduction Equipment cost O&M cost Load shifting Peak shaving IRR [%] Payback
[kW] time
[year]

1 142 660 142 2.97 0.59 1.24 1.51 – > 20

2 100 360 100 1.67 0.33 0.69 1.07 0.41 > 20
3 80 215 80 1.04 0.21 0.42 0.85 5.03 8.83
4 60 160 60 0.78 0.16 0.32 0.64 5.14 8.78
5 40 75 40 0.39 0.08 0.15 0.43 10.33 7.08
6 20 55 20 0.27 0.05 0.11 0.21 4.68 8.95

13
2870 Journal of Electrical Engineering & Technology (2024) 19:2859–2874

economic efficiency, and a peak reduction by 40 kW was For Customer 4, economic feasibility comes out when the
the most economical. Bold in the Table 10 means optimal peak reduction is decreased. According to IRR of case 4–6,
capacity by analysis. Computation time for Case 5 is 74 s. the optimal capacity was determined as PCS 100 kW and
For Customer 3, the peak load remained within 30% battery 250 kWh which was bold in the Table 12. Computa-
of the contract demand. Therefore, there was no peak tion time for Case 4 is 90 s.
cut amount, and the peak cut revenue of (5) could not be The optimal ESS capacity was selected in Tables 9, 10, 11
obtained. In this case, profits may be obtained from reduc- and 12, and the optimal PV capacity for each customer was
ing electricity bills as much as possible by load shifting calculated according to the proposed algorithm in Fig. 8.
that moves the load during the off-peak load time to the For all the customers, the economy improved considerably
peak load time. when PV was included.
Therefore, the PCS capacity was fixed at 40 kW, which is The results of the economic feasibility analysis while
10% of the average load of 396 kW, and the battery capacity increasing PV capacity for Customer 1 are shown in
was increased to 2 to 7 times the PCS capacity. During eco- Table 13. A higher PV capacity corresponds to higher IRR,
nomic feasibility analysis, there was no economic feasibility but lower increasing rate of IRR. The graph of IRR change
for all candidate capacities. PCS 40 kW-Battery 75 kWh was according to PV capacity increase is shown in Fig. 18.
arbitrarily selected as the optimal capacity and bold in the Therefore, Case 2, which had the highest increase rate, was
Table 11. Computation time for Case 3 is 52 s. determined as the optimal capacity and bold in the Table 13.
Because the maximum peak reduction of Consumer 4 was Likewise, in the case of Customer 2, the economy
530 kW, the battery candidate capacity was calculated using improved considerably when PV was included. Table 14
the algorithm of Fig. 5 while randomly gradually decreasing shows the results of analyzing the economic feasibility while
the peak reduction at 500 kW. The results of the economic increasing PV capacity.
analysis according to the candidate capacity are presented The graph of IRR change according to PV capac-
in Table 12. Cases 1 to 3 were not economically feasible. ity increase of customer 2 is shown in Fig. 19. The IRR

Table 11  Economic analysis results of Customer 3 for installing ESS

Case Capacity Assumption Investment cost [100 Million Revenue [100 Million KRW] Economic index
Num- KRW]
ber
PCS [kW] Battery [kWh] Peak reduction [kW] Equipment cost O&M cost Load shifting Peak shaving IRR [%] Payback
time
[year]

1 40 80 0 0.41 0.08 0.16 0 – > 20

2 40 120 0 0.57 0.11 0.23 0 – > 20
3 40 160 0 0.73 0.15 0.30 0 – > 20
4 40 200 0 0.89 0.18 0.37 0 – > 20
5 40 240 0 1.05 0.21 0.42 0 – > 20
6 40 280 0 1.21 0.24 0.47 0 – > 20

Table 12  Economic analysis results of Customer 4 for installing ESS

Case Capacity Assumption Investment Cost [100 Mil- Revenue [100 Million KRW] Economic index
Num- lion KRW]
ber
PCS [kW] Battery [kWh] Peak reduction [kW] Equipment cost O&M cost Load shifting Peak shaving IRR [%] Payback
time
[year]

1 500 3850 500 16.55 3.31 5.86 5.33 – > 20

2 400 2500 400 10.92 2.18 4.04 4.26 – > 20
3 300 1350 300 6.09 1.22 2.45 3.20 – > 20
4 200 550 200 2.66 0.53 1.08 2.13 4.68 8.95
5 100 250 100 1.23 0.25 0.50 1.07 6.10 8.44
6 50 150 50 0.72 0.14 0.29 0.53 3.36 9.44

13
Journal of Electrical Engineering & Technology (2024) 19:2859–2874 2871

Table 13  Economic analysis results of Customer 1 for installing PV-ESS

Case Candidate capacity Investment cost [100 Million Revenue [100 Million KRW] Economic index
Number KRW]
PCS [kW] Battery [kWh] PV [kW] Equipment cost O&M cost Load shifting Peak shaving IRR [%] Payback
time [year]

1 50 230 0 1.04 0.21 0.55 0.60 2.36 9.90

2 50 230 50 1.64 0.33 2.16 0.60 13.2 6.65
3 50 230 100 2.24 0.45 3.78 0.60 16.98 5.82
4 50 230 150 2.84 0.57 5.32 0.60 18.66 5.50
5 50 230 200 3.44 0.69 7.00 0.60 20.27 5.21
6 50 230 250 4.04 0.81 8.53 0.60 20.84 5.11
7 50 230 300 4.64 0.93 10.13 0.60 21.50 5.01

capacity was 100 kW, and it was determined as the optimal

capacity and bold in the Table 15.
The economic feasibility analysis results obtained by
increasing the PV capacity of Customer 4 to the upper limit
in Table 1 are shown in Table 16 and Fig. 21.
The IRR increased with the PV candidate capacity, but
the increased rate of IRR was the highest when the PV
capacity was 100 kW. According to the algorithm in Fig. 8,
the optimal capacity of Consumer 4 was determined to be
ESS 100 kW–250 kWh and PV 100 kW and bold in the
Table 16.
Fig. 18  Plot of IRR according to the PV capacity of Customer 1

4 Conclusion
increases with the PV capacity, but decreases when it
exceeds 300 kW. Therefore, the IRR was the highest when This paper proposes a strategy for designing capacity for
the PV capacity was 300 kW, and it was determined as the consumers who want to install PV-ESS for DSM.
optimal capacity and was bold in the Table 14. The PV-ESS optimal capacity design procedure includes
Table 15 shows the results of analyzing economic fea- load pattern forecasting, PV modeling, operation schedul-
sibility as the PV capacity of Customer 3 increased. As ing optimization functions for demand management, and
Fig. 20, the IRR increases with the PV capacity, attains economic index calculation. Furthermore, each concept is
its highest value at 500 kW, and then IRR falls. However, explained, and the overall flow is introduced. Therefore,
the increased rate of IRR was the highest when the PV the procedures and formulas can be organized for universal
application.

Table 14  Economic analysis results of Customer 2 for installing PV-ESS

Case Candidate capacity Investment cost [100 Million Revenue [100 Million KRW] Economic index
Number KRW]
PCS [kW] Battery [kWh] PV [kW] Equipment cost O&M cost Load shifting Peak shaving IRR [%] Payback
time [year]

1 40 75 0 0.39 0.08 0.15 0.43 10.33 7.08

2 40 75 100 1.59 0.32 2.69 0.43 17.99 5.32
3 40 75 200 2.79 0.56 5.22 0.43 18.95 5.15
4 40 75 300 3.99 0.80 7.75 0.43 19.28 5.09
5 40 75 400 5.19 1.04 10.16 0.43 19.14 5.12
6 40 75 500 6.39 1.28 12.45 0.43 18.79 5.19

13
2872 Journal of Electrical Engineering & Technology (2024) 19:2859–2874

Fig. 19  Plot of IRR according to the PV capacity of Customer 2 Fig. 20  Plot of IRR according to PV capacity of Customer 3

And we proposed an algorithm for optimal sizing entire • When economic feasibility analysis was conducted by
procedure of PV-ESS, calculation process of initial capacity, adding PV resources to the optimal ESS capacity, eco-
upper limit capacity and candidate capacities, procedure for nomic feasibility improved in most cases. As PV candi-
selecting the optimal PV capacity, which is the most impor- date capacity increases, economic feasibility improves,
tant for optimal capacity design. In other words, this study but the increase rate is low. Therefore, the IRR of the
introduces a method for designing the capacity in sequences optimal PV-ESS candidate capacity was calculated to be
to maximize profit when designing a combination of two or 10.01–17.99% according to customer load pattern.
more resources. • In conclusion, for DSM economic feasibility, using pro-
The results of the numerical simulations performed using posed method is need to find the minimum ESS capac-
the proposed method for four actual industrial customers in ity accompanied by peak reduction and PV capacity of
the Republic of Korea are summarized below. maximum increase rate according to customer load pat-
tern.
• The optimal capacity and economic index varied accord-
ing to the characteristics of the consumer load pattern. Therefore, if the government wants to revitalize the sup-
In most simulation results, the larger the peak reduc- ply of renewable energy and reduce output volatility with
tion amount, the larger the ESS candidate capacity, and ESS, it is judged that investment cost support projects or
the high cost of ESS made it uneconomical. Therefore, incentive rate systems for PV-ESS will be necessary.
economic feasibility is good when peak reduction was Future work is planned to extend the current study and
achieved by lowering the ESS capacity as possible. In implement a stochastic model for solar power generators
the result analysis, the IRR varied from 2.36 to 10.33% to more perfect the peak-saving profit calculation method.
according to customer load pattern.

Table 15  Economic analysis results of Customer 3 for installing PV-ESS

Case Candidate capacity Investment cost [100 Million Revenue [100 Million KRW] Economic index
Number KRW]
PCS [kW] Battery [kWh] PV [kW] Equipment cost O&M cost Load shifting Peak shaving IRR [%] Payback
time
[year]

1 40 160 0 0.73 0.15 0.30 0 – > 20

2 40 160 100 1.93 0.39 2.84 0 10.01 7.26
3 40 160 200 3.13 0.63 5.35 0 14.08 6.18
4 40 160 300 4.33 0.87 7.90 0 15.92 5.76
5 40 160 400 5.53 1.11 10.30 0 16.49 5.64
6 40 160 500 6.73 1.35 12.56 0 16.57 5.63
7 40 160 600 7.93 1.59 14.54 0 16.06 5.73
8 40 160 700 9.13 1.83 16.12 0 15.00 5.97
9 40 160 800 10.33 2.07 17.42 0 13.73 6.27

13
Journal of Electrical Engineering & Technology (2024) 19:2859–2874 2873

Table 16  Economic analysis results of Customer 4 for installing PV-ESS

Case Candidate capacity Investment cost [100 Million Revenue [100 Million KRW] Economic index
Number KRW]
PCS [kW] Battery [kWh] PV [kW] Equipment cost O&M cost Load shifting Peak shaving IRR [%] Payback
time [year]

1 100 250 0 1.23 0.25 0.50 1.07 6.10 8.44

2 100 250 100 2.43 0.49 2.98 1.07 13.47 6.30
3 100 250 200 3.63 0.73 5.47 1.07 15.60 5.81
4 100 250 300 4.83 0.97 7.95 1.07 16.63 5.60
5 100 250 400 6.03 1.21 10.44 1.07 17.23 5.48
6 100 250 500 7.23 1.45 12.90 1.07 17.58 5.41

7. Bahramirad S, Reder W, Khodaei A (2012) Reliability-con-

strained optimal sizing of energy storage system in a microgrid.
IEEE Trans Smart Grid 3(4):2056–2062
8. Boglou V, Karavas CS, Karlis A, Arvanitis KG, Palaiologou
I (2023) An optimal distributed RES sizing strategy in hybrid
low voltage networks focused on EVs’ integration. IEEE Access
11:16250–16270
9. Blasuttigh N, Negri S, Massi Pavan A, Tironi E (2023) Optimal
sizing and environ-economic analysis of PV-BESS systems for
jointly acting renewable self-consumers. Energies 16(3):1244
10. Luo X et al (2019) Optimal design and cost allocation of a
distributed energy resource (DER) system with district energy
networks: a case study of an isolated island in the South China
Sea. Sustain Cities Soc 51:101726
11. Gomez-Gonzalez M, Hernandez JC, Vera D, Jurado F (2020)
Fig. 21  Plot of IRR according to the PV capacity of Customer 4 Optimal sizing and power schedule in PV household-prosumers
for improving PV self-consumption and providing frequency
containment reserve. Energy 191:116554
12. Hernández JC, Sanchez-Sutil F, Muñoz-Rodríguez FJ, Baier CR
We also plan to generalize the characteristics of various (2020) Optimal sizing and management strategy for PV house-
battery models to calculate optimal capacity. hold-prosumers with self-consumption/sufficiency enhancement
and provision of frequency containment reserve. Appl Energy
Acknowledgements This work was supported by the Korea Institute of 277:115529
Energy Technology Evaluation and Planning(KETEP), granted finan- 13. Gil-González W, Montoya OD, Rajagopalan A, Grisales-Noreña
cial resource from the Ministry of Trade, Industry & Energy, Republic LF, Hernández JC (2020) Optimal selection and location of
of Korea (No. 2021202090053B). fixed-step capacitor banks in distribution networks using a
discrete version of the vortex search algorithm. Energies
13(18):4914
14. Montoya OD, Gil-González W, Hernández JC, Giral-Ramírez
DA, Medina-Quesada A (2020) A mixed-integer nonlinear pro-
References gramming model for optimal reconfiguration of DC distribution
feeders. Energies 13(17):4440
15. Electric power Statistics Information System (https://​epsis.​kpx.​
1. World energy market insight 2023–12, Korea Energy Economics or.​kr/​epsis​new/​selec​tEksa​AscAs​aChart.​do?​menuId=​060405)
Institute (http://​www.​keei.​re.​kr/​insig​ht) 16. KEPCO cyber homepage (https://​h ome.​kepco.​c o.​k r/​kepco/​
2. Kim J-H, Kwak C (2023) How to improve the flexibility of CY/K/​htmlV​iew/​CYKAH​P001.​do?​menuCd=​FN030​70101​01)
power system against increased fluctuations in renewable energy 17. Rahman MH, Yamashiro S (2007) Novel distributed power
generation with global energy crisis. J. Korean Prod Oper generating system of PV-ECaSS using solar energy estimation.
Manag Soc 34(3):287–307 IEEE Trans Energy Convers 22(2):358–367
3. BNEF (2023) 1H 2023 energy storage market outlook 18. Cho KH, Kim SK, Kim ES (2013) Optimal capacity determination
4. Renewable Energy Data Center Cite. Renewable energy poten- method of battery energy storage system for demand management
tial, Korea Institute of Energy Research (https://​kier-​solar.​org/​ of electricity customer. Trans Korean Inst Electr Eng 62(1):21–28
user/​map/​map_​paten​tial.​do), Access on Sep. 6 2023 19. KEPCO power planner homepage (https://​pp.​kepco.​co.​kr/)
5. Ministry of Trade, Industry and Energy (2023) The 10th basic 20. Korea meteorological administration site (http://​web.​kma.​go.​
electricity supply and demand plan. Notice No. 2023-036, Jan. kr/​eng/​index.​jsp)
13, 2023
6. Ahn JW, Oh Y-T (2023) A study on energy storage system Publisher's Note Springer Nature remains neutral with regard to
capacity design for educational service customer. J Korea Acad jurisdictional claims in published maps and institutional affiliations.
Ind Cooperation Soc 24(4):61–69

13
2874 Journal of Electrical Engineering & Technology (2024) 19:2859–2874

Springer Nature or its licensor (e.g. a society or other partner) holds Gilsung Byeon He received B.S
exclusive rights to this article under a publishing agreement with the and Ph.D degrees from the
author(s) or other rightsholder(s); author self-archiving of the accepted School of Electrical Engineer-
manuscript version of this article is solely governed by the terms of ing, Korea University in 2006
such publishing agreement and applicable law. and 2012. Currently, he works as
a principal researcher with the
energy platform research center,
Kyeong‑Hee Cho She received KERI.
B.S. and M.S. degrees from the
Gyeongsang National University
in 2010 and 2012. She received
Ph.D. degree from Pusan
National University in 2023. She
has been with Korea Electrotech-
noloty Research Institute
(KERI), since 2011. Her research
interests are renewable energy, Wanbin Son He received Ph.D.
energy storage system and in computer science from
microgrid design. POSTECH, South Korea. Cur-
rently, he works as a senior
researcher with the energy plat-
form research center, KERI. His
research interests are PV fore-
Jongyul Kim He received B.S., casting and digital twin of dis-
M.S., and Ph.D. degrees in Elec- tributed resources.
trical Engineering from Pusan
National University. He has been
with the KERI, since 2001. Cur-
rently, he works as a director
with the Energy Platform
Research Center. His research
interests are the DERs and
design & operation of microgrid
and smart grid.

13