Received September 24, 2020, accepted October 5, 2020, date of publication October 16, 2020, date of current version

October 28, 2020.

Digital Object Identifier 10.1109/ACCESS.2020.3031591

Practical Power Management of PV/ESS

Integrated System
SEOUNG UK JEON , (Member, IEEE), JINAH NOH , (Graduate Student Member, IEEE),
SEOKJU KANG , (Graduate Student Member, IEEE), AND
JUNG-WOOK PARK , (Senior Member, IEEE)
School of Electrical and Electronic Engineering, Yonsei University, Seoul 03722, South Korea
Corresponding author: Jung-Wook Park ([email protected])
This work was supported in part by the National Research Foundation of Korea (NRF) funded by the Ministry of Science and ICT (MIST),
Korea Government under Grant 2020R1A3B2079407, and in part by the Human Resources Program in Energy Technology, Korea Institute
of Energy Technology Evaluation and Planning (KETEP), Ministry of Trade, Industry and Energy, South Korea, under Grant
20194030202420.

ABSTRACT This study proposes the practical power management of integrated system with photo-
voltaic (PV) and energy storage system (ESS) to solve the line congestion and voltage problems in a
distribution system while considering the electricity price and state of charge (SOC) of ESS. In addition,
the proper power management of PV/ESS integrated system (PEIS) enables to increase the hosting capacity
of renewable energies in a distribution system and to maximize the profit of independent power producer
(IPP). In this study, the real power management of PEIS of 100 kW is implemented in practice and tested with
the actual measurements for 3 days. The results show that the proposed method deals with the line congestion
problem effectively and efficiently while increasing the profit of IPP by 11%. In particular, the reactive power
control based on the international electrotechnical commission (IEC) standard, IEC 61850-90-7 is applied
to the PEIS, and it can successfully mitigate the variations of voltage in a distribution system.

INDEX TERMS Distribution system, energy storage system, integrated system, photovoltaic system, power
management, reactive power control, renewable energy.

I. INTRODUCTION On the other hand, every distribution line has the power
To deal with the climate change due to greenhouse gas limit, above which is not allowed to flow, due to its thermal
emission, many countries have been strongly pursuing the capacity. This problem is called as the line congestion in
green growth policies while focusing on renewable energies a distribution system [4], [5]. For example, the real power
such as photovoltaic (PV) and wind, etc. Accordingly, many flowing in a general distribution line is limited to 10 MW at
renewable energy-based generators have been connecting in maximum in South Korea [6]. As the result, the other PV sys-
the transmission and distribution systems of existing power tems cannot be newly connected to this line because it already
grid. The PV generation has an infinite energy source with- has many renewable energy-based generators, of which total
out making environmental pollution. Its characteristic shows output power reaches to 10 MW. Even though the increase of
the pattern of general power consumption, and therefore hosting capacity with the installation of new lines and trans-
this can make helpful supplying the power. In other words, formers can be a fundamental solution, it requires the high
it mainly generates power during the daytime when the most cost. To figure out this problem, the PV and energy storage
electricity is consumed. Thus, it can be strategically used system (ESS) integrated system (PEIS) can be applied [2],
for system operator to resolve the power shortage during [3]. In other words, even when the PV system can generate
the peak demand. However, its output power is affected by its maximum power according to the good weather condition,
weather conditions. This results in increasing the variability some amount of power is charged in the ESS depending on
and uncertainty in its power generation [1]–[3]. its capacity. This can prevent the maximum power from PV
system flowing in distribution line. As the result, it enables to
The associate editor coordinating the review of this manuscript and solve the line congestion problem, and therefore increase the
approving it for publication was Chandan Kumar . hosting capacity of renewable energies.

This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
VOLUME 8, 2020 189775
 S. U. Jeon et al.: Practical Power Management of PV/ESS Integrated System

There are several studies on the integrated PV-ESS system the particular distribution line, and it generates its maxi-
[7], [8], which describes the integrated PV-ESS system with mum power output according to the good weather condition,
the DC bus. It supplies the AC power depending on whether it might excess the power limit (10 MW in South Korea) of
there is a load consumption in DC bus or a generation from line due to its thermal capacity in some sections, as shown
PV system. Because the DC bus is separated from AC system, in Fig. 1(a), even though the ESS exists in the other separate
and the associated power management system (PMS) handles location [6]. To figure it out, the PEIS is applied as shown
the power fluctuation of PV generation and DC load, it can in Fig. 1(b). In this case, even if the PV system generates the
contribute to improving the stability of distribution system. same maximum power, some amount of power can be charged
However, because the DC bus voltage mainly depends on in the ESS. Then, this results in avoiding line congestion
the state of charge (SOC) of ESS, it does not change signifi- problem. Also, this energy stored in ESS of PEIS can be
cantly while the power generation from PV system is subject used when the electricity price is high at peak demand
to fluctuate. Therefore, the DC/DC converter is required to and/or the subsidy supported from government is also high
adjust the DC voltage and transfer the power. As the result, at nighttime [12].
the additional cost for DC/DC converter and the loss from
power conversion reduce the independent power producer
(IPP)’s revenue. In contrast, the proposed PEIS does not
require the DC bus. Also, its PMS can be easily applied to the
conventional PV and ESS system. Moreover, there is no need
for additional DC/DC converter or DC bus system. It requires
the only line measurement device.
The South Korea government has been carrying out the
strong renewable energy promotion program while giving
more weight (that is, higher subsidy) to the time schedule
operation of ESS. Then, the PEIS gives the chance to sell its
stored energy at nighttime at the higher cost determined by
this weight. Therefore, it can reduce the investment payback
period although it requires the high initial cost [2], [3]. Even
though there are several studies to prevent the line congestion
by estimating the margin of line capacity and improving the
optimal power flow method for microgrid [9]–[11], there are
no researches to directly solve the line congestion problem
while increasing the IPP’s revenue.
This paper proposes the new practical PMS applied to the
PEIS with the inverter (INV) for PV system and the power
conditioning system (PCS) for ESS, which basically supports
the time schedule to consider the economic profit while solv-
ing the line congestion problem in a distribution system, and FIGURE 1. Line congestion in a distribution line.
therefore increasing the hosting capacity of renewable ener-
gies. This paper is organized as follows. Section II explains B. PRACTICAL PV/ESS INTEGRATED SYSTEM OF 100 KW
the line congestion problem in a distribution line. It also The structure of practical PEIS of 100 kW is shown in the
describes the structure of PEIS with its communication pro- Fig. 2. It consists of the INV for PV system and the PCS
cess. Section III gives the full details for proposed PMS. Also, for ESS in hardware. The measurements from the INV is
it introduces the renewable energy policy of South Korea with firstly transmitted to the PMS. Also, the PCS measures the
the focus on subsidy. In Section IV, the case studies are carried input DC voltage, output AC voltage/current, and SOC of
out to verify the performance of PMS applied to the PEIS battery. Then, they are sent to the PMS in every second based
of 100 kW. Thereafter, the experimental results tested for on the recommended standard (RS) 485 communication. The
3 days are given in Section V. Finally, a conclusion is given power flow from the PEIS to line is measured by the power
in section VI. protection & monitoring equipment (PPME), and this is trans-
mitted to the PMS. Note that the PPME, which is added to
II. THE PROPOSED PV/ESS INTEGRATED SYSTEM the conventional PEIS, can also perform block function of
A. LINE CONGESTION PROBLEM reverse power flow [6]. The power flow monitoring equip-
As mentioned in above, the PV system can be strategically ment (PFME) measures the power flowing through the line,
used because its output pattern is similar to that of load and send it to the PMS.
during daytime [1]–[3]. However, it is characterized with The communication method used to implement the pro-
the variability and uncertainty depending on weather con- posed PEIS is the RS485 communication, which is more
ditions. In particular, when the PV system is connected to resistant to noise than the other serial communications such

189776 VOLUME 8, 2020

S. U. Jeon et al.: Practical Power Management of PV/ESS Integrated System

of distribution system based on the grid code including the

line congestion problem.
The existing PMS does not consider the real-time vari-
ous conditions of distribution line, and therefore it does not
actively manage the output powers from PV and ESS systems
to solve the line congestion and voltage stability problems,
etc [15]–[18]. In other words, it has simply made the only
charging and discharging operations of ESS in their time
schedule modes. In particular, when the line capacity is still
insufficient, this line congestion problem will be more aggra-
vated if the ESS discharges its energy at the same time [13].
Then, the proposed PMS can actively handle this with the
real-time information of distribution system while solving the
line congestion problem.

B. VOLTAGE CONTROL
By the IEC 61850-90-7, the PCS of renewable energy-
based generator must be able to carry out the smart inverter
FIGURE 2. Structure of PEIS with its communication process. functions, which can deal with the frequency-real power,
frequency-reactive power, voltage-real power, and voltage-
as RS232, RS442. Also, it is more preferably used for reactive power relationships, etc. In addition, they support the
the long-distance communication. In particular, the PFME operation for giving the economic profit to the IPP. Among
(which acquires the information of line) is inevitably far the above functions, which can be implemented by the virtual
from the PMS. For this case, it can transmit the informa- droop control, this study focuses on the voltage control by
tion more stably. For these reasons, the RS485 communi- reactive power compensation, as shown in Fig. 3. That is,
cation has been used in many industry applications. The the PEIS provides or absorbs the reactive power to control the
important components of proposed PEIS, which are INV voltage by the INV for PV system and/or PCS for ESS while
for PV, PCS for ESS, PPME, and PFME, are using the regulating the power factor within ±0.9 in South Korea [6].
RS485 communication. In addition, in order to prevent the The values from V1 to V4 in Fig. 3 might be different
accident by communication error, the acknowledge signal is depending on the system condition and grid code in each
periodically transmitted and received to check whether the country. When the value of voltage is lower than that of
communication is stable. If a communication error occurs, V2 (which is 372.4 V), the INV supplies the positive (over-
the proposed PEIS is stopped. It is set to 19200 bps, which excited) reactive power to increase the voltage. In contrast,
is the communication speed, which can transmit the most if it is higher than that of V3 (which is 372.4 V), the negative
data while being able to perform the stable communication. (under-excited) reactive power is compensated to decrease
There is no interference between the AC system of 60 Hz voltage. For this study, this compensation starts right after the
and the PWM frequency of 7.2 kHz. Moreover, the device voltage is above ±2% of the rated voltage. When it reaches
equipped with a terminating resistor is used in the RS485 to ±5% of the rated voltage (which are corresponding to
communication module to prevent the signal reflection. Also, V4 and V1, respectively), the PEIS makes the maximum
the proposed PMS can be easily applied to the conven- reactive power compensation up to 48.4% of rated power so
tional PV and ESS system because there is no need for that the power factor becomes ±0.9.
additional DC/DC converters or DC bus systems, which are For the reactive power compensation, when the line voltage
required in [7] and [8]. exceeds its limit, the proposed PMS calculates the amount of
On the other hand, the reactive power must be effectively reactive power required to restore the voltage. Then, it adjusts
compensated to keep the voltage stability in distribution the power factor of PEIS until it supplies the required reac-
system. In practice, the only INV performs this function. tive power. In this case, the power factor is limited not to
This study proposes the new reactive power compensation drop below 0.9 according to the grid-code regulation in [6].
method with the smart inverter functions based on the inter- In some cases, when the reactive power compensation cannot
national electrotechnical commission (IEC) standard, IEC be fully made due to the lack of PV generation or SOC of ESS,
61850-90-7 [14]. Then, its effectiveness is verified by the this control can be accomplished inside of the dark triangle
practical hardware experimental test. in Fig. 3.

III. THE PROPOSED POWER MANAGEMENT SYSTEM C. RENEWABLE ENERGY PRICE POLICY
A. POWER FLOW CONTROL The South Korea government is operating the renewable
Each renewable energy-based generator must be connected energy certificate (REC) program to support and expand
in practice after considering the allowable hosting capacity the renewable energy business. The system operator sets the

VOLUME 8, 2020 189777

 S. U. Jeon et al.: Practical Power Management of PV/ESS Integrated System

FIGURE 3. Voltage control by reactive power compensation by PEIS.

FIGURE 5. Power flows by the proposed PEIS to get the REC price in
different time periods.

grid from16:00 to 10:00 in the next day. If the IPP follows

this time schedule, he can get the REC 5.0. If it charges
in the other times, the IPP receives the REC 1.2, which is
FIGURE 4. REC program and its corresponding electricity price obtained the same as the case by the only PV system. In particular,
by the PEIS. when the ESS receives the power from the grid (not the
power generated by the PV system), all RECs are impounded.
electricity price by the annual bidding. This fixed bidding Moreover, if this happens repeatedly, the system is forcibly
price consists of the system marginal price (SMP) and REC cut off. Thus, the IPP is recommended to strictly follow the
prices. The SMP is the price of electrical energy (kWh) pro- above suggested time schedule while preventing the reverse
duced by power plants, and it varies depending on the energy power flow.
sources such as fossil fuels, nuclear, and liquefied natural gas The power flows by the proposed PEIS to get the REC price
(LNG), etc. Then, REC price is determined by subtracting in different time periods is shown in Fig. 5. When the power
the SMP from the fixed bidding price. The details of REC from PV of proposed PEIS flows to the grid before 10:00 as
program are shown in Fig. 4. The weight of REC is differ- shown in Fig. 5(a), it gets the REC 1.2, not the REC 5.0 (see
ent depending on the type of renewable energy source. For the yellow period about 8:00∼10:00 in Fig. 4). In contrast,
example, when compared to other renewable energy sources the power flow from the PCS for ESS is monitored and the
such as wind and geothermal power, the PV system is easy to ESS discharges the energy before 10:00 with the REC 5.0.
install, and its generation time is more predictable. Therefore, As shown in Figs.5 (b) and (c), all powers generated from the
the weight of 1.2 (which is the REC 1.2) is applied to the PV is charged in the ESS from 10:00 to 16:00. If the proposed
electricity from PV system. In summary, the IPPs with the PEIS sells the power to the grid in this period, it still gets the
PV system can get both SMP and 1.2 · REC when they sell REC 1.2. After 16:00, the proposed PEIS stops charging the
the power to the grid. On the other hand, the PV system with ESS, and it still sells the power from the PV with the REC
the ESS can store the energy, and they can sell the power any 1.2 until the PV stops generating power (see the yellow period
time they want while contributing to the grid management. about 16:00∼18:00 in Fig. 4). Thereafter, the ESS discharges
Thus, the weight of 5 (which is REC 5.0) is applied to this the energy until 10:00 in the next day with REC 5.0, as shown
system when the ESS charges the energy from PV system in Fig. 5 (d). Note that the REC price is paid as a subside in
from 10:00 to 16:00, and it sells the power to the grid by every 1,000 kWh.
discharging the energy between 16:00 and 10:00 in the next In the first half of 2020, the fixed bidding price of South
day. In conclusion, the proposed PEIS monitors the power Korea is 160 Korean won/kWh. Then, the values of SMP and
flow of the PCS for ESS to reflect the REC price as follows. REC price for 3 days in March are shown in Fig. 6, where
Firstly, the amount of energy charged from 10:00 to 16:00 is the time intervals between 10:00 and 16:00 to get the REC
measured. Thereafter, this accumulated energy is sold to the 5.0 in each day are indicated in the grey area [18]. In this

189778 VOLUME 8, 2020

S. U. Jeon et al.: Practical Power Management of PV/ESS Integrated System

FIGURE 6. Values of SMP and REC price for 3 days in March.

period, the SMP is generally high. In contrast, the REC price

is relatively low. In particular, the sudden drop in the SMP
(and the sudden increase in the REC price) occurs because
the power consumption is temporarily decreased between
12:00 and 13:00 during lunch time. If the PEIS discharges
(and sells the power) on around 3:00 at night, when the REC
price is the highest among the day, the IPP can earn the much
bigger economic profit.

D. PROCEDURE TO IMPLEMNT THE PROPOSED PMS

The most important feature of proposed PMS is the use of line
and SMP information so that it avoids the line congestion and
voltage stability problems while maximizing the economic
profit. The procedure to implement the proposed PMS is
given in Fig. 7. Firstly, the PMS acquires the data from the
PEIS and external route. That is, the output powers, Ppv
and Pess and power factors, pfpv and pfess from the INV
and ESS, respectively, are measured. Also, the SOC of ESS
is obtained, and it is transmitted to the PMS. In addition,
the line capacity, Pl.cap and the line voltage, vl are measured
by the PFME. Then, with the SMP and time information
from the external source, the PMS determines the REC
price and its weight information (ωREC ), as shown in Fig. 4.
By using the above information, the reference power flow,
Pref for ESS and reference power factor, pfref for PEIS are
calculated. In particular, the Pref is used as a variable of
optimization problem. The measured Pess can be expressed
with the difference between the charging, Pc and discharging,
Pd powers as
Pess = Pc − Pd (1)
Depending on the charging time in Fig. 4, the ωREC is
selected to 1.2 or 5. Then, it is important that the ESS must be
FIGURE 7. Procedure to implement the proposed PMS.
charged when the ωREC is 5.0 to maximize the profit of IPP.
Therefore, during the period from 10:00 to 16:00, the PMS
controls to charge the ESS fully such that the reference charg- 10:00 in the next day, the reference discharging power, Pref .d
ing power, Pref .c becomes 100 kW as (2). On the contrary, is computed as (3). The PMS receives the SOC information in
when the ωREC is 1.2 for the time intervals from 16:00 to real-time, and it determines the reference discharging power.

VOLUME 8, 2020 189779

 S. U. Jeon et al.: Practical Power Management of PV/ESS Integrated System

In other words, it can be calculated by multiplying the rated (DCP) [19] such that the objective function of Jp + Js is
capacity of PCS (which is 100 kW) and the ratio of 1SOC minimized as
and 1Time.
min Js + Jp .


Pref .c = 100 (2) s.t Ppv − Pref ≤ Pl.cap
1SOC
Pref .d = 100 · (3) Pref ≤ Ppv . (9)
1Time
where 1SOC is the difference between the present SOC The minimization of combination of Jp in (4) and Js in (6)
and the constant SOC of 10%, which is the recommended becomes a convex optimization problem and it consists of
as the minimum limit of ESS. And, 1Time is the remain- a quadratic objective function and two linear constraints.
ing time until 10:00 in the next day from the present Then, it can be solved by the DCP while finding the proper
time. After calculating the Pref .c by (2) and Pref .d by (3), value of Pref . For solving this quadratic equation with upper
the objective function, Jp , for the power flow of ESS is and lower bounds, the DCP is transformed into a quadratic
defined as program (QP) which can be solved with a standard QP

2 solver [20].
Jp = Pref − Pref .c orPref .d (4) The DCP is implemented in MATLAB and solves the
Firstly, the PMS determines the value of Pref to minimize optimization problem with least-squares gradient estimation
the objective function, Jp in (4), which makes the reference as follows. Firstly, the proposed PMS updates the measure-
charging and discharging powers, Pref .c or Pref .d to operate ment, and it assigns the variables to the optimization problem.
as (2) and (3), respectively, while reflecting the renewable Thereafter, the gradient method starts to find the optimal
energy policy. That is, if the value of Pref is different from that value within the available range. If the value of objective
of Pref .c or Pref .d , the value of Jp increases. In contrast, if they function of Jp + Js is decreased after the Pref is decreased in
are same, the value of Jp becomes zero. To prevent diverging a present iteration, then the Pref will be decreased in the next
its value to the negative infinity, the Jp is set as a quadratic iteration with the update in search of finding its minimum
equation. Thereafter, with the ωREC , the total selling price value. In the opposite direction, if the value of objective
(SP) can be obtained as function of Jp + Js is decreased after the Pref is increased in
a present iteration, then the Pref will be increased in the next
SP = SMP + ωREC · RECprice(ωREC = 1.2, 5) (5) iteration with the update in search of finding its minimum
value. This iteration performs the gradient method to find
Next, the second objective function, Js reflecting the price
the proper value until the change in the objective function
information is defined as the product of SP in (5) and
1(Jp + Js ) is less than one. As mentioned above, the PMS is
Pref as
implemented in the DCP under electricity price, measurement
Js = SP · Pref (6) information and grid-code regulation.
Next, as shown in Fig. 3, the required reactive power, Qreq
Because the value of Js becomes negative only when the depending on the variation of vl , is compensated. In other
ESS discharges, it is recommended that the PEIS controls words, if the vl is within ±2% of the rated voltage, there
to discharge the ESS more than the Pref .d when the SP is no need to compensate. Therefore, the Qreq is zero, and
is high. As shown in the Fig. 6, the REC price becomes the pfref becomes 1. Otherwise, the Qreq has non-zero value.
different depending on the time. For example, even though Thereafter, it is compared with the actual reactive power from
the SMP is the lowest at around 3:00, the corresponding the PEIS, QPEIS , which is defined as the sum of the reactive
REC price is the highest among the day. The Js reflects this powers from INV for PV and PCS for ESS as
price information to the PMS. There are some constraints to q q
solve the optimization problems in (4) and (6). Firstly, if the 1 − pf 2pv 1 − pf 2ess
PEIS receives the power from the grid, the profit supported QPEIS = Ppv · + Pess · (10)
by the REC is reduced [6]. Thus, the PEIS must charge pf pv pf ess
the power less than the Ppv . The associated constraint is The PMS keeps decreasing the pfref by 0.001 in step from
formulated as 1 to 0.9 until the values of QPEIS and Qreq become same.
Pref ≤ Ppv (7) According to [6], if the pfref is lower than 0.9, its value is
set to 0.9. Note that when the value of Qreq is positive and
Also, because the power flowing to the grid, Ppv − Pref , negative, the PEIS performs the lagging and leading pfref
cannot exceed the Pl.cap to avoid the line congestion problem, controls, respectively.
the following constraint is required to consider.
IV. SIMULATION RESULTS
Ppv − Pref ≤ Pl.cap (8)
To evaluate the performance of PMS applied to the PEIS
Finally, the optimization problem to find the proper of 100 kW, the case study is carried out by using the real
value of Pref is solved by the disciplined convex program measurements for 3 days, as shown in Fig. 8.

189780 VOLUME 8, 2020

S. U. Jeon et al.: Practical Power Management of PV/ESS Integrated System

discharging. Most IPPs have to sell the stored power from

16:00 to 10:00 in the next day. In other words, they discharge
the ESS in a short time (about 2 hours) among this long-time
duration, and they stop the PCS for ESS in the rest time, when
it cannot participate in the compensation. Therefore, it does
not contribute to solve the line congestion problem.

B. CASE B: PROPOSED PMS METHOD WITH VOLTAGE

FIGURE 8. Line capacity and output power from PV system for 3 days.
CONTROL ONLY
In this case, the proposed PMS method performs the charging
and discharging operation of ESS based on the scheduled
mode by the conventional PMS method (like the case A) with
the only voltage control.
As shown in Figs. 11(a) and 11(b), the ESS charges the
power from PV system in the period of time with the REC
5.0. After that, the PCS discharges the stored energy with
FIGURE 9. Frequency measurement of line. the rated capacity of 100 kW for about 2 hours, and it stops
until the next charging time. Also, it receives the voltage
information of line, and it makes the reactive power com-
pensation. It is possible to perform the reactive compensation
by its voltage control when the PV system is generating,
or the ESS is discharging. However, even if the voltage drops
below the lower limit of 372.4 V, the PEIS often cannot
participate in the compensation when there is no generation
from the PV system, and the ESS does not discharge, as
shown in Fig. 12(a).

FIGURE 10. Voltage measurement of line and required reactive power

compensation.

In this case study, the Pl.cap , which is indicated by the blue

line, is assumed to be the difference between the power of
100 kW and load consumption in the range from few kW to
100 kW at maximum. The Ppv is represented by the red line.
Then, the line congestion problem might occur when the Ppv
is greater than the Pl.cap , as shown in the light green circles
in each day. Also, the practical measurement of frequency in
the line is shown in Fig. 9. It is observed that the frequency
of system is kept stable within the proper range from 59.5 Hz
to 60.5 Hz required by the grid code. In contrast, it is known
from the measurement shown in Fig. 10(a) that the vl drops
below the lower limit of 372.4 V required by the grid code
whenever the load varies seriously (for example, at around
9:00 in every morning). Then, the corresponding Qreq can be
calculated, as shown in Fig. 10(b).

A. CASE A: CONVENTIONAL PMS METHOD

In this case, the PV system and the ESS operate in the
scheduled mode by the conventional PMS method [15]–[18].
The ESS is charged in the recommended time period between
10:00 and 16:00. At the other times, it sells the power by FIGURE 11. Results of line congestion for each case.

VOLUME 8, 2020 189781

 S. U. Jeon et al.: Practical Power Management of PV/ESS Integrated System

32:00 and 56:00 in the morning, when the line congestion

becomes rapidly serious, the ESS is charged despite it is not
the time period of REC 5.0. It is shown in Fig. 12 that the
reactive power is supplied for the voltage control in all cases
except for the case A. The compensation result for case B
is shown in Fig. 12(a). The PEIS cannot participate in the
compensation at around 22:00, 46:00, and 70:00, when there
is no generation from the PV system, and the PCS does not
discharge the ESS. By considering the SOC and time to use
the discharging power properly as the cases C and D, the PEIS
can participate in the reactive power compensation for the
entire time, as shown in Figs. 12(b) and 12(c). At around
10:00 and 32:00 for both cases, the amount of compensation
is less than the required reactive power because the real power
is insufficient, and the power factor is limited by 0.9.
Table I summarizes the comparison results of all case
studies. The line congestion time represents the time, when
the line congestion problem might occur, if there is no specific
constraint reflecting the power flow information of line. It is
observed that the PEIS might not participate in reactive power
FIGURE 12. Required reactive power and PEIS voltage compensation compensation if the PMS discharges the ESS in a short time,
output for each case. and it do not distribute the discharging power appropriately.
The charging and discharging operations for the case B are
the same as those for the case A. However, the profit in the
C. CASE C: PROPOSED PMS METHOD WITH VOLTAGE case B is slightly reduced than that in the case A because the
AND POWER MANAGEMENT CONTROLS active power decreases as much as the PMS participates in
In this case, in order not to discharge the ESS too quickly, the the reactive power compensation for voltage control in the
power from ESS is uniformly produced by (3) during the dis- case B. Also, it is proved from profits in the cases C and D that
charge time. By the consideration of SOC and time, the PMS the proper distribution of discharging powers can increase the
makes the ESS discharging with the small amount of power profit of IPP because the PMS discharges the ESS in the time
from 16:00 to 10:00 in the next day, as shown in Fig. 11(c). period, when the REC price is expensive.
Even if there is no PV generation, the PEIS enables to par-
ticipate in all compensations for the entire time, as shown TABLE 1. Comparison of PMS performance in all cases.
in Fig. 12(b). However, the PMS does not consider the line
capacity and the electricity price information for this case.
Therefore, the line congestion problem might occur, and the
discharging power will not change significantly.

D. CASE D: PROPOSED PMS METHOD BY OPTIMIZATION

WITH CONSIDEREATION OF ELECTRICITY PRICE
In this case, the power reference of ESS is optimized, and
it applied to the PMS with the consideration of economic
profit for the IPP as well as line congestion problem and In particular, the profit in the case D becomes the high-
the reactive power compensation. Firstly, the PMS adjust the est because it has some adjustment reflecting the price
output power from the PEIS to avoid the line congestion information.
problem in all operations, as shown in Fig. 11(d). That is,
if the PEIS operates without considering the power flow V. EXPERIMENTAL RESULTS
information of line as in the cases A–C, its output can exceed A. IMPLEMENTATION WITH HARDWARE PROTOTYPE
the line capacity, as shown in the light green circles of To verify the relation between the reactive power and the line
Figs. 11(a)–(c). voltage, the power factor of PEIS is regulated [21]. The power
When compared to the discharging operation in the cases capacity of implemented PEIS is 100 kW, and it is too small
C and D, it is observed that amount of selling power for when compared to the practical distribution line of 10 MW.
the IPP is adjusted to maximize the profit. In other words, Therefore, the line voltage will not change noticeably by the
it is increased at around 3:00, when the REC price is the reactive power compensation of 100 kW PEIS. To overcome
most expensive, as shown in Fig. 11(d) because the SMP and this problem, the experimental test in Fig. 13 is carried out.
REC price are reflected to the PMS. In particular, at around In other words, the power factor of PEIS is changed to

189782 VOLUME 8, 2020

S. U. Jeon et al.: Practical Power Management of PV/ESS Integrated System

FIGURE 13. Experimental set for voltage compensation function.

FIGURE 15. PEIS of 100 kW on the real test-bed.

FIGURE 16. Line capacity and PV generation of demonstration test-bed

in 3 days.

FIGURE 14. Voltage compensation due to reactive power supply.

while implementing the proposed PMS. Because the amount
of PV generation must be charged to the ESS, both the
confirm the change of voltage by the reactive power compen- rated capacity of the INV for PV and the PCS for ESS are
sation, while the output power from PEIS is fixed. The Ppv implemented to be equal as 100 kW. Also, the rated capacity
and Pess are supplied from the PV and ESS simulator, and of ESS is designed such that it is 2.5 to 3 times bigger than that
they are fixed with the rated power of 100 kW. Also, the INV of maximum PV generation [22], [23]. Therefore, the ESS
for PV and the PCS for ESS are connected to the same of 274 kWh is connected to the PEIS.
line. However, they are isolated by transformer to prevent After carrying out the test for 3 days, the experimental
switching interference and circulating current. Then, they results are shown in Fig. 16. It is observed that the line
are supplying the reactive power up to their rated capacity congestion problem might occur when the Ppv is greater than
of 48.4% by adjusting the power factor of 0.9. the Pl.cap (see the light green circles of Fig. 16). As confirmed
The real-time results are shown in Fig. 14. It is observed in the case studies of Section IV, all functions such as voltage
from Fig. 14(a) that the PMS regulates the lagging power compensation and SOC management to solve this problem
factor, and the INV for PV and the PCS for ESS provide the are performed in the experiment.
positive reactive power. As the result, the voltage is increased. Also, it is known from the results of Fig. 17(a) that the PEIS
In contrast, it is known from Fig. 14(b) that the PMS adjusts discharges more power because its PMS reflects the price
the leading power factor, and the INV for PV and the PCS for information at around 3:00, 27:00, and 50:00 when the REC
ESS generate the negative reactive power. Correspondingly, price is the most expensive. In contrast, the output power of
the voltage is decreased. PEIS is adjusted not to exceed the line capacity even though
it is the time to sell the PV generation at around 9:00, 32:00,
B. VERIFICATION ON REAL TEST-BED and 56:00, as shown in Fig. 17(b).
To verify the practical effectiveness of proposed PEIS of As mentioned in above, the effect of reactive power com-
100 kW, it is constructed on real test-bed, as shown in Fig. 15, pensation is not high because the PEIS of 100 kW is not

VOLUME 8, 2020 189783

 S. U. Jeon et al.: Practical Power Management of PV/ESS Integrated System

demonstration test-bed. In other words, the proposed PEIS

and PMS is developed by considering the real price infor-
mation and grid-code policy, etc. Also, many real-time field
tests were carried out with actual data to verify the practical
effectiveness of proposed PEIS and PMS. Therefore, it would
be expected that the proposed PEIS and PMS over 1 MW
(with the enhanced scalability) can be expanded for the field
test on a practical distribution system of 22.9 kV without
much difficulty.

REFERENCES
[1] A. A. Almehizia, H. M. K. Al-Masri, and M. Ehsani, ‘‘Integration of
renewable energy sources by load shifting and utilizing value storage,’’
IEEE Trans. Smart Grid, vol. 10, no. 5, pp. 4974–4984, Sep. 2019.
[2] M. C. Di Piazza, M. Luna, G. La Tona, and A. Di Piazza, ‘‘Improving
grid integration of hybrid PV-storage systems through a suitable energy
FIGURE 17. Results of active power control by the PEIS on the real management strategy,’’ IEEE Trans. Ind. Appl., vol. 55, no. 1, pp. 60–68,
test-bed. Jan. 2019.
[3] A. Saez-de-Ibarra, V. I. Herrera, A. Milo, H. Gaztanaga,
I. Etxeberria-Otadui, S. Bacha, and A. Padros, ‘‘Management strategy
for market participation of photovoltaic power plants including storage
systems,’’ IEEE Trans. Ind. Appl., vol. 52, no. 5, pp. 4292–4303,
Sep. 2016.
[4] R. Ciavarella, M. Di Somma, G. Graditi, and M. Valenti, ‘‘Congestion
management in distribution grid networks through active power control
of flexible distributed energy resources,’’ in Proc. IEEE Milan PowerTech,
Jun. 2019, pp. 1–6.
[5] N. I. Yusoff, A. A. M. Zin, and A. B. Khairuddin, ‘‘Congestion man-
agement in power system: A review,’’ in Proc. 3rd Int. Conf. (PGSRET),
Jan. 2018, pp. 22–27.
[6] KEPCO. (Apr. 29, 2020). Res. 1, Regulations for the Electrical
Equipment for Transmission and Distribution. [Online]. Available:
http://cyber.kepco.co.kr/ckepco/front/jsp/CY/H/C/CYHCHP00601.jsp
[7] Z. Yi, W. Dong, and A. H. Etemadi, ‘‘A unified control and power man-
agement scheme for PV-battery-based hybrid microgrids for both grid-
connected and islanded modes,’’ IEEE Trans. Smart Grid, vol. 9, no. 6,
FIGURE 18. Voltage measurements and corresponding reactive power pp. 5975–5985, Nov. 2018.
compensation by the PMS on the real test-bed. [8] K. Sun, X. Wang, Y. W. Li, F. Nejabatkhah, Y. Mei, and X. Lu, ‘‘Par-
allel operation of bidirectional interfacing converters in a hybrid AC/DC
microgrid under unbalanced grid voltage conditions,’’ IEEE Trans. Power
large enough to actually affect the line voltage. Nevertheless, Electron., vol. 32, no. 3, pp. 1872–1884, Mar. 2017.
the proposed PMS effectively supplies the reactive power by [9] K. Chaudhari, A. Ukil, K. N. Kumar, U. Manandhar, and S. K. Kollimalla,
‘‘Hybrid optimization for economic deployment of ESS in PV-integrated
the power factor control according to the variations of line EV charging stations,’’ IEEE Trans. Ind. Informat., vol. 14, no. 1,
voltage in all times while solving the line congestion problem, pp. 106–116, Jan. 2018.
as shown in Fig. 18. [10] M. Khanabadi, Y. Fu, and C. Liu, ‘‘Decentralized transmission line switch-
ing for congestion management of interconnected power systems,’’ IEEE
Trans. Power Syst., vol. 33, no. 6, pp. 5902–5912, Nov. 2018.
VI. CONCLUSION [11] T. M. Masaud and E. F. El-Saadany, ‘‘Optimal wind DG integration for
This paper proposed the new practical power management security risk-based line overload enhancement: A two stage approach,’’
IEEE Access, vol. 8, pp. 11939–11947, 2020.
system (PMS) applied to the photovoltaic (PV) and energy [12] Renewable Portfolio System. (Jun. 20, 2018). [Online]. Available:
storage system (ESS) integrated system (PEIS) to solve the http://www.energy.or.kr/web/kem_home_new/new_main.asp
line congestion and voltage problems in a distribution system. [13] California ISO. (Nov. 11, 2016). What the Duck Curve Tells us
About Managing a Green Grid. [Online]. Available: http://www.caiso.
Also, it utilized the electricity price and state of charge (SOC) com/Documents/FlexibleResourcesHelpRenewables_FastFacts.pdf
of ESS information for its optimal operation. To evaluate the [14] Communication Networks and Systems for Power Utility Automation—
performance of proposed PMS, the simulations case studies Part 90-7: Object Models for Power Converters in Distributed Energy
Resources (DER) Systems, document IEC/TR 61850-90-7, Feb. 2013.
were carried out with the actual data. Moreover, the hardware
[Online]. Available: https://webstore.iec.ch/publication/6027
prototype of PEIS with the rated power of 100 kW was [15] (Sep. 17, 2014). The 4th Renewable Energy Basic Plan Ministry of
implemented and tested on the real test-bed. Both simulation Commerce Industry and Energy of Korea. [Online]. Available: http://
and experimental results verified that the proposed PEIS with climatepolicydatabase.org/index.php/4th_National_Basic_Plan_for_New_
and_Renewable_Energies_(2014-2035)
new PMS can successfully solve the line congestion problem, [16] D. H. Ko, J. Chung, K. S. Lee, J. S. Park, and J. H. Yi, ‘‘Current policy
mitigate the variations of voltage, and increase the economic and technology for tidal current energy in Korea,’’ Energies, vol. 12, no. 9,
profit of producer at the same time. p. 1807, May 2019.
[17] Korea Energy Agency. ‘‘Status and forecast of national energy,’’
In this study, the South Korea electricity price and renew- New Renew. Energy, Ministry Trade, Ind. Energy (MOTIE), Sejong,
able energy supply policy were applied to the practical test on South Korea, Tech. Rep. 11-1410000-001321-11, 2018, pp. 58–59.

189784 VOLUME 8, 2020

S. U. Jeon et al.: Practical Power Management of PV/ESS Integrated System

[18] Korea Power Exchange (KPX). (Mar. 23, 2020). Electric Power Statistics SEOKJU KANG (Graduate Student Member,
Information System. [Online]. Available: http://epsis.kpx.or.kr/epsisnew/ IEEE) received the B.S. degree from Yonsei Uni-
[19] S. Boyd and L. Vandenberghe, Convex Optimization. Cambridge, U.K.: versity, Seoul, South Korea, in 2016, where he
Cambridge Univ. Press, 2004. is currently pursuing the Ph.D. degree through
[20] M. Grant and S. Boyd. CVX: MATLAB Software for Disciplined Convex the combined M.S. and Ph.D. Program with the
Programming. [Online]. Available: http://cvxr.com/cvx/ School of Electrical and Electronic Engineering.
[21] U. Standard. (Jan. 2010). UL 1741—Standard for Inverters, His research interests include operation of dis-
Converters, Controllers and Interconnection System Equipment
tribution networks, renewable energies-based dis-
for Use With Distributed Energy Resources. [Online]. Available:
tributed generation systems, and voltage stability
https://standardscatalog.ul.com/standards/en/standard_1741_2,
[22] H. G. Lee, G.-G. Kim, B. G. Bhang, D. K. Kim, N. Park, and H.-K. Ahn, of power systems.
‘‘Design algorithm for optimum capacity of ESS connected with PVs under
the RPS program,’’ IEEE Access, vol. 6, pp. 45899–45906, 2018.
[23] X. Song, Y. Zhao, J. Zhou, and Z. Weng, ‘‘Reliability varying charac-
teristics of PV-ESS-based standalone microgrid,’’ IEEE Access, vol. 7,
pp. 120872–120883, 2019.

SEOUNG UK JEON (Member, IEEE) received JUNG-WOOK PARK (Senior Member, IEEE)
the B.S. degree from Yonsei University, Seoul, was born in Seoul, South Korea. He received
South Korea, in 2011, where he is currently pursu- the B.S. degree (summa cum laude) from the
ing the Ph.D. degree through the combined M.S. Department of Electrical Engineering, Yonsei Uni-
and Ph.D. Program with the School of Electrical versity, Seoul, South Korea, in 1999, and the
and Electronic Engineering. His research interests M.S.E.C.E. and Ph.D. degrees from the School
include power conversion system control, opti- of Electrical and Computer Engineering, Geor-
mal management using converter-based renewable gia Institute of Technology, Atlanta, GA, USA,
energy, and energy storage systems. in 2000 and 2003, respectively. He was a Post-
doctoral Research Associate with the Department
of Electrical and Computer Engineering, University of Wisconsin-Madison,
USA, from 2003 to 2004. He was also a Senior Research Engineer with
LG Electronics Inc., South Korea, from 2004 to 2005. Since 2005, he has
JINAH NOH (Graduate Student Member, IEEE) been with the School of Electrical and Electronic Engineering, Yonsei Uni-
received the B.S. degree from Gyeongsang versity, where he is currently a Professor. He is also the Director of the
National University, Jinju, South Korea, in 2011. Yonsei-Power System Research Center for Great Energy Transition (Yonsei-
She is currently pursuing the Ph.D. degree through PREFER) supported by the Leading Research Program (7.2M USD Grant
the combined M.S. and Ph.D. Program with the for nine years), National Research Foundation (NRL), South Korea. His cur-
School of Electrical and Electronic Engineering, rent research interests include power system dynamics, energy management
Yonsei University, Seoul, South Korea. She was systems, renewable energies-based distributed generation systems, operation
an Assistant Researcher with KD Power Company and planning of microgrid, hardware implementation of power-electronic-
Ltd., Chuncheon, South Korea, from 2012 to 2015. based inverters, and so on. He was a recipient of the Young Scientist
Her research interests include control and opera- Presidential Award from the Korean Academy of Science and Technology
tion of power systems, energy management systems, and renewable energies- (KAST), South Korea, in 2013.
based distributed generation systems.

VOLUME 8, 2020 189785