Content uploaded by Md. Rashedul Alam
Author content
All content in this area was uploaded by Md. Rashedul Alam on Feb 09, 2024
Content may be subject to copyright.
JOURNAL OF SUSTAINABLE ENERGY VOL. 14, NO. 2, DECEMBER, 2023
ISSN 2067-5534 © 2023 JSE 124
A FEASIBILITY STUDY OF SOLAR PHOTOVOLTAIC POWER
SMOOTHING USING FUZZY LOGIC APPROACH
RASHEDUL A.*, IFTEKHAR U.B.**, NUR M.***
* Sustainable and Renewable Energy Development Authority (SREDA), Dhaka-1000, Bangladesh
** Institute of Appropriate Technology, Bangladesh University of Engineering and Technology (BUET),
Dhaka-1000, Bangladesh
*** Department of Electrical and Electronic Engineering, Chittagong University of Engineering and
Technology (CUET), Chittagong-4349, Bangladesh
ad.solar@sreda.gov.bd, rashed4912@gmail.com
Abstract - Penetration of Variable Renewable Energy
(VRE) is a challenge for safe grid integration. The short-
term variation of solar irradiance, which is initiated by
moving clouds, causes fluctuations in Solar Photovoltaic
(SPV) power generation and can jeopardize grid stability.
The fluctuations in the output power of the SPV plants are
the reason for the dynamic change of load flow in the
interconnection area of the utility network. To assess the
short-term variation of solar irradiance, 1-year time-series
solar irradiance data have been collected from a Solar
Irradiance Measurement Station; located at Chittagong
University of Engineering and Technology (CUET),
Chittagong, Bangladesh. The collected data from the case
study site reveals that the short-term variation of solar
irradiance is significant especially from April to
September. Furthermore, a feasibility study of SPV power
smoothing has been conducted using the Fuzzy Logic
approach to identify the requirement of the Energy
Storage System (ESS) as well as to minimize the solar ramp
rate and ramp level. An 8 MWh ESS with an 8 MW power
capacity has been identified as the capacity of the ESS
support system for smoothing a 20 MWp solar plant. The
daily support amount and the surplus amount have been
calculated for solar power smoothing and that found
identical in terms of energy and power. Although this
feasibility study gives a directive on grid integration
aspects before establishing a large utility-scale SPV plant,
the actual scenarios may be slightly different due to the
geographical dispersion, cloud enhancement and similar
other effects.
Keywords: Solar Irradiance, Variable Renewable Energy,
Short-term Variation, Solar Power Smoothing, Fuzzy Logic,
Feasibility Study, Energy Storage.
1. INTRODUCTION
Solar irradiance is variable in nature due to environmental
conditions like cloud patterns and their movement [1,2]. As it
is the input of the Solar Photovoltaic (SPV) power plant, the
output power of the plant depends on it [3,4]. The short-term
variability of solar irradiance, originating from moving clouds,
causes fluctuations in the SPV power generation and can
negatively affect grid stability [5]. The Ramp-rate (RR)
statistics, a quantifying parameter of solar power variability; is
widely used, most common and practical quantities [6]. The
power ramp-rate control (PRRC) strategy is employed to limit
the fluctuation rate in the photovoltaic (PV) output power
under dynamically changing irradiance conditions [7]. It has
been observed that more than 50% of the days in a year in
Bangladesh experience high solar irradiance short-term
variability [8].
To address this short-term variability, grid codes of
several countries have incorporated ramp-rate limitations to
inject variable renewable energy (VRE) like SPV plants
[9,10]. In Bangladesh, there are no such specific requirements
of ramp-rate limitations in the conditions of the power
purchase agreement of utility-scale solar power plants as well
as in the national grid code of Bangladesh [11]. The
fluctuations in the output power of the plant are the reason for
the dynamic change of load flow in the interconnection area of
the utility network [12]. There are some issues like cloud
enhancement and geographical dispersion but still, such
intermittency poses significant challenges [13–15]. Increasing
penetration in the interconnected power network impacts
system frequency response. Therefore, it is difficult for the
Transmission System Operator (TSO) to address the frequency
regulation. Understanding the nature of this intermittency is
important as it can unstable system inertia and stability. Some
important issues are discussed below.
1.1. Geographical Dispersion
The geographic dispersion of solar-photovoltaic panels
reduces variability in energy production. A study result was
published in [16] that characterized some plants’ power output
variability based on minute-averaged irradiation data from
each plant and the output from 390 inverters. The result of that
study was observed maximum ramp rates of 0.7, 0.58, 0.53,
and 0.43 times the plant’s capacity for 5, 21, 48, and 80 MW
Alternating Current (AC) plants respectively due to
geographical dispersion. The study was conducted by
simulating a step-by-step increase in the plant size at the same
location [16]. The study was based on the United States of
America (USA) and Canada although the scenario may be
different in Bangladesh with some positive effects of
geographical dispersion.
JOURNAL OF SUSTAINABLE ENERGY VOL. 14, NO. 2, DECEMBER, 2023
ISSN 2067-5534 © 2023 JSE 125
1.2. Cloud Enhancement
The solar irradiance can exceed the level of expected clear
sky irradiance for a very short time during partly cloudy days.
This is known as cloud enhancement (CE) or over irradiance
or irradiance enhancement. The CE phenomenon can be
observed worldwide but the amount may vary from place to
place. According to [17], solar irradiance on such days may be
increased to 1.5 to 1.6 times more for a short time than the
clear weather days level. It also increases the output power
fluctuation of the SPV power plant.
1.3. Application of Energy Storage to Minimize this
Variation
The power electronic inverters are capable of operating
with grid-friendly features like volt-VAR control, ramp-rate
control, high-frequency power curtailment, and event ride-
through [18]. To minimize the ramp rate of the SPV power
plant, it is essential to provide support from alternative sources
[9,19]. Utility-scale storage Li-Ion Battery is widely used for
this purpose due to their first response [20]. Some other storage
technologies like large Vanadium Redox flow batteries,
Polysulphide Bromine flow batteries, and Zinc Bromine flow
batteries are parallelly used [21,22]. Recently, energy storage,
Superconductive magnetic energy storage, Sensible thermal
energy storage, Latent-phase change material,
Thermochemical storage, and Pumped hydro storage have also
been explored as utility-scale storage [23,24]. However, the
cycle life of energy storage is a challenge for ramp
management of SPV applications [25]. Moving Average,
Exponential Moving Average, First Order Low-Pass Filter,
Second-Order Low-Pass Filter, and Fuzzy Logic Controller
are investigated as control technologies to minimize the ramp
rate [26–28]. The fuzzy logic controller is familiar among
them and comprehensively studied in [29,30]. Energy storage
may be a good solution to deal with the intermittency however,
the associated cost and its disposal mechanism and
maintenance are a few major concerns [31,32].
The short-term variation of solar irradiance, generated
from moving clouds, is very much location-specific and
depends on the weather conditions of that area. Therefore, a
detailed grid integration study is essential for a large-scale
SPV plant to understand the integration effect on the utility
interconnection point. In this paper, a feasibility study of SPV
power smoothing has been conducted using the Fuzzy Logic
approach. This analysis is based on a proposed capacity of 20
MWp SPV plant for the location of Chittagong University of
Engineering and Technology (CUET), Bangladesh, where a
solar radiation resource monitoring station is present. The
study is divided into a few parts. Those are, (a) Solar Irradiance
data collection for the case study site, at least for a year; (b)
Data checking, filtering and finalizing the dataset for the study;
(c) Developing a fuzzy model for power smoothing analysis;
(d) Apply this model with input dataset and find out the power
and energy scenario of each day; (e) Finalize the results and
make some recommendation to overcome the challenges. As
the input data set is site or region-specific, a such study in
Bangladesh is pivotal and it’s the demand of the future. The
smoothing level may be adjusted by changing the degree of
membership function. Therefore, it can be used anywhere with
necessary modifications.
2. DATA SOURCE
2.1. National Solar Radiation Resource Assessment
Station, Chittagong
The Sustainable and Renewable Energy Development
Authority (SREDA), the nodal agency of Bangladesh,
installed eight solar irradiance resource measurement stations
in different locations of Bangladesh under a Global
Environment Facility (GEF) funded project ‘Sustainable
Renewable Energy Power Generation (SREPGen)’ to promote
Renewable Energy in the country. Locations of the solar
resource monitoring sites were Rangpur (BRUR), Rajshahi
(RUET), Mymensingh (BAU), Sylhet (SUST), Kushtia (KPL),
Khulna (KU), Patuakhali (PSTU), and Chittagong (CUET).
Out of these eight solar radiation resource measurement
stations, the Chittagong division’s site (CUET) was selected
for this analysis. The latitude and longitude of the site are
22.463998 and 91.973298 respectively. Global Horizontal
Irradiance (GHI), Diffuse Horizontal Irradiance, and some
weather data were recorded in the resource monitoring station.
The solar irradiance data were recorded using pyranometers
and a datalogger. The data sampling frequency was 10 seconds
and averaging frequency was 4 minutes. The maximum,
minimum, and standard deviation data of global horizontal
irradiance also were present with the average global horizontal
irradiance, which shows the more precise scenario of solar
irradiance variation. The location of the selected site is shown
below in Fig. 1 from the geographical map of Bangladesh.
Fig. 1. Location of the data source
JOURNAL OF SUSTAINABLE ENERGY VOL. 14, NO. 2, DECEMBER, 2023
ISSN 2067-5534 © 2023 JSE 126
3. MATERIALS AND METHODS
3.1. Block diagram of a utility-scale SPV power plant
The utility-scale SPV plant consists of one or multiple
blocks and each block is connected to a central substation.
There should be one or more, central or string inverters in each
block where each inverter has one or more input terminals with
a Maximum Power Point Tracking (MPPT) unit. The number
of PV modules will be connected in series and create a sting
according to the voltage of the input terminal of the inverter.
Similarly, the number of the string will be connected in parallel
and creates an array to achieve the input current and power of
each MPPT input terminal of the grid-tied solar inverter.
Synchronization and power quality will be ensured in each
inverter of the SPV plant by following international standards
where the output voltage of an inverter is normally less than
1kV. Alternating Current (AC) terminals of each inverter of a
block are connected to a low-voltage busbar. The voltage level
will be stepped up to 11kV or 33kV level through a power
transformer and it will be connected to a central substation of
the SPV plant. A central substation may have step-up
transformers according to the voltage level of the integration
point of the utility network. At least one energy meter will be
placed to measure the power, energy and related parameters
that will be used for billing and other purposes. A Power Plant
Controller (PPC) may be installed in the control room with the
necessary data communication infrastructure which will be
connected to each inverter of the SPV plant. There should be
at least one solar irradiance measurement station in a utility-
scale SPV plant. A block diagram of a utility-scale SPV plant
is shown in Fig. 2.
Fig. 2. Block diagram of a utility-scale solar power plant
The output power of an inverter depends on the input solar
irradiance on solar modules that are connected to this inverter.
Similarly, the combined output of all inverters will represent
the output power of the SPV plant. In this study, variation of
input solar irradiance has been identified to determine the
variation of output power and develop necessary
recommendations to reduce the variation for safe grid
integration of SPV plant. A case study site, mentioned in the
data source section, has been selected for this study.
3.2. Fuzzy Logic Approach
The solar irradiance data, mentioned in the data source
section, were used to determine the variation analysis. The
dataset was checked by the MATLAB program to find out the
out-of-range and unexpected values. The ramp rate is the
change rate of PV output per unit of time. As it is a feasibility
study before installing the SPV plant, the solar irradiance
dataset is the foundation, precise per unit time series solar
irradiance data is not available in Bangladesh; a study has been
conducted considering the existing resources to present the
current scenarios. A fuzzy logic-based model was used to
determine the energy and power support amount by
minimizing the output power variation of a proposed 20 MWp
CUET solar power plant in Bangladesh. This solar power
variation minimization support could be delivered from the
Energy Storage System (ESS) but the only tentative
requirement was assessed to reduce the solar ramping. The
required support power and energy amount were calculated on
a daily basis to understand the case study. Detailed working
procedures of the fuzzy model and its calculation are discussed
below.
3.3. Fuzzy Model
Fuzzy logic has two different meanings. In a narrow
sense, fuzzy logic is a logical system, which is an extension of
multivalued logic. However, in a wider sense, Fuzzy Logic
(FL) is almost synonymous with the theory of fuzzy sets, a
theory that relates to classes of objects with unsharp
boundaries in which membership is a matter of degree. Fuzzy
logic differs both in concept and substance from traditional
multivalued logical systems.
The collected solar irradiance dataset mentioned above
was utilized to create a model using a fuzzy logic approach to
obtain the optimum charging and discharging rate of storage
considering the minimization of output power variation of the
solar power plants. MATLAB R2018a was used to develop the
fuzzy model and calculate the power and energy support
analysis to reduce the solar ramp. A triangular membership
function was selected to develop this fuzzy model. Using the
developed model based on the solar resource variation and its
patterns, the required battery storage capacity was suggested
to optimize the output power variation of the solar power plant
by solar ramp management. The required investment amount
was identified from the suggested battery storage capacity.
This will optimize the variability of the output power of the
solar power plant at the grid interconnection point.
3.4. Input and output selection
The daily solar irradiance data and supporting storage
capacity were the inputs of the fuzzy model. The output was
the marginal stable equivalent of solar irradiance data after
ramp management support. Sometimes surplus energy is
needed to be stored in the ESS for smoothing. Similarly,
sometimes energy support will be needed from ESS to manage
the deficit solar irradiation. Energy calculations were done by
the MATLAB program. ESS is considered a full charge during
the initial study whereas it will be reduced gradually with its
discharge. Input and Output of the model including levels are
PV ModulePV Module
PV ModulePV Module
PV ModulePV Module
Inverter
Low Voltage Busbar
1st Block N th Block
High Voltage Busbar (11kV/33kV)
Extra High Voltage (132kV/230kV)
Step-up
transformer
Inverter Inverter
Cable
Cable
Cable
ACDC AC DC ACDC
PV ModulePV Module
PV ModulePV Module
PV ModulePV Module
Inverter
Energy Meter
of Offtaker
Low Voltage Busbar
Step-up
transformer
Step-up
transformer
Inverter Inverter
Cable
Cable
ACDC AC DC ACDC
Cable
JOURNAL OF SUSTAINABLE ENERGY VOL. 14, NO. 2, DECEMBER, 2023
ISSN 2067-5534 © 2023 JSE 127
given below:
INPUT: Solar Radiation Data (W/m2)
INPUT: Battery Storage Capacity (MWh)
OUTPUT: Output Power Equivalent Solar Radiation Data
(W/m2)
Radiation Difference = Input Solar Radiation Data -
Output Power Equivalent Solar Irradiance Data
Where,
VL [1] = Very Low
LOW [2] = Low
MID [3] = Medium
HIGH [4] = High
VH [5] = Very High
Fig. 3. GUI of Input and Output of the Fuzzy model
Fig. 3, a Graphical User Interface (GUI) of fuzzy logic,
represents the Mamdani-type Fuzzy Logic Designer which is
used to design the fuzzy model. Two inputs, the irradiance gap
(difference between the previous level and present level) and
storage availability are shown on the left. The
charging/discharging rate is displayed on the right side of the
Mamdani block. Each input and output block containing the
membership functions can be displayed by clicking on the
block of MATLAB Fuzzy Toolbox as displayed in Fig. 3.
3.5. Membership functions
A membership function for a fuzzy set-A on the universe
of discourse X is defined as µA:X → [0,1], where each element
of X is mapped to a value between 0 and 1 for the membership
function µA. This value, called membership value or degree of
membership, quantifies the grade of membership of the
element in X to the fuzzy set A.
Fig. 4. Triangular membership function
Membership functions allow us to graphically represent a
fuzzy set. The x-axis represents the universe of discourse,
whereas the y-axis represents the degrees of membership in the
[0,1] interval. The triangular membership function is described
below as it was used in the developed fuzzy model.
Triangular membership function: defined by a lower
limit ‘a’, an upper limit ‘b’, and a value ‘m’ as mentioned in
Fig. 4, where a < m < b. Each position of the x-axis gives a
membership value defined in the y-axis.
3.6. Rules of the fuzzy model
To develop the rules, the following Table 1 was developed
according to the two input projections.
Table 1. Fuzzy Rule Development
Based on Table 1, 25 rules were developed considering the
relation between the irradiance difference and ESS conditions.
Irradiance difference has been classified into five groups,
starting from Very Low (VL) to Very High (VH). Similarly,
ESS capacity has been classified into five groups, starting from
VL to VH. Based on the condition of 2 inputs, 25 situations
have been identified like Very Quick Change (VQC), Quick
Change (QC), Medium (MID), Flexible Change (FLC), Very
Flexible Change (VFLC), etc. A GUI image of the MATLAB
Fuzzy Rule Editor is shown in Fig. 5.
0.2
a m b
0.4
0.6
0.8
1.0
x ≤ a
a < x ≤ m
m < x < b
µ
A
(x) =
x ≥ b
0,
0,
,
(x-a)
(m-a)
,
(b-x)
(b-m)
IRR DIF
ESS CAP
VL [1] LOW [2] MID [3] HIGH [4] VH [5]
VL [1] VQC VQC QC QC MID
LOW [2] VQC QC QC MID MID
MID [3] QC QC MID MID FLC
HIGH [4] QC MID MID FLC FLC
VH [5] MID MID FLC FLC VFLC
JOURNAL OF SUSTAINABLE ENERGY VOL. 14, NO. 2, DECEMBER, 2023
ISSN 2067-5534 © 2023 JSE 128
Fig. 5. GUI of Fuzzy Rule Editor
3.7. Calculation and Model Development using MATLAB
The function evalfis() was used to execute the fuzzy model
in MATLAB. Based on the result of the fuzzy model, an
estimated output level was identified. Energy deficit and
surplus were calculated using the MATLAB program.
Maximum deficit power and maximum surplus power were
also recorded to identify the required maximum
discharge/storage rate. Finally, a summarized record table was
generated to get the evaluated results. A block diagram of the
calculation process is shown below.
Fig. 6. Block diagram of the data analysis process
In Fig. 6, ‘Rin’ is the input time series solar irradiance data
and ‘R’ is the modified output power equivalent solar
irradiance level after giving support according to the fuzzy
model.
4. RESULTS AND DISCUSSIONS
Solar irradiance is variable in nature as the weather
conditions may be different in each place. Therefore, a whole-
year assessment is essential to understand the actual output
including seasonal variations. According to a yearly solar
irradiance assessment of two sites in Bangladesh, more than
50% of days in a year contain high solar irradiance variation, a
few days have washout days and only 25-30% of days in a year
were clear weather days. The result of the case study site is
shown in Fig. 7 and more details can be found in [8].
Fig. 7. Yearly summary on daily solar irradiation, SRRA
St., CTG
4.1. Assessment of Solar Irradiance
The variation of solar irradiance creates an effect on the
output power of the SPV plant as well as the effect of load flow
in the interconnection point of the utility network. The cloud
enhancement effect is partly reflected in the solar irradiance
data but the geographical dispersion effect is not included in
the recorded solar irradiance data. Although considering the
geographical dispersion effect, the variation is still present in
the output of solar power plants in a significant amount. The
amount of additional energy support is essential to decrease the
sharp power variations that are calculated from the solar
irradiance data set. A daily solar irradiance data reflecting a
high solar irradiance variable scenario is shown in Fig. 8.
Fig. 8. A daily solar irradiance variation data
Surplus
Deficit
E
N
D
R
Fuz
z
y
Model
Plot/Display
Seq. Data
Store
Load Fuzzy Model
S
T
A
R
T
R
in
> R
Data Filtering
Input Data
R
in
R
Y
e
s
No
0
5
10
15
20
25
30
35
JAN
2020
FEB
2020
MAR
2020
APR
2020
MAY
2020
JUN
2020
JUL
2020
AUG
2020
SEP
2020
OCT
2020
NOV
2020
DEC
2020
Yearly summary on daily solar irradiation, SRRA St., CTG
Total Days Washout Days High Variability Days Low Variability Days Clear Days
JOURNAL OF SUSTAINABLE ENERGY VOL. 14, NO. 2, DECEMBER, 2023
ISSN 2067-5534 © 2023 JSE 129
Fig. 8 is a representation of the daily solar irradiance curve
of the case study site. The data sampling frequency of that site
was 10 seconds and the averaging frequency was 4 minutes.
There are some data other than mean solar irradiance which are
very helpful to understanding the precise variation. Those are
the maximum value, minimum value and standard deviation of
solar irradiance within a data averaging period. The blue colour
represents the mean value of GHI. It was a high solar irradiance
variability day as the short-term variation of solar irradiance
level is very high, almost 200 W/m2 to 1000 W/m2. Also, the
irradiance change rate is very sharp and the change number per
day is more. The maximum value, minimum value and
standard deviation of GHI give a clearer understanding of the
variation with a precise time scale. The very sharp decrease or
increase of solar irradiance, several times a day is not a safe
grid integration for large utility-scale solar power plants.
Additional support could reduce the effect of those solar
irradiance variations where the necessary technical
arrangement is essential.
A fuzzy model using fuzzy logic approach was used to
minimize the solar ramping using the energy support from the
ESS. Details are described in the methodology section.
MATLAB R2018a version was used to develop the fuzzy
model and data analysis was completed by the MATLAB
program. The data analysis procedure is shown in Fig. 6 and
some significant analysed figures are described below.
Fig. 9. Reducing the sharp variation by additional support
or surplus component
The deep blue line of Fig. 9 indicates the daily solar
irradiance status of the day whereas the red line represents the
marginal stable output after providing the power and energy
support according to the developed fuzzy model. The light blue
line represents the surplus electricity according to the ramp
management and the brown line represents the deficit
electricity which will be supported by the ESS. After
calculation using the MATLAB program with fuzzy model, the
result shows that the total deficit electricity was approximately
4.91 MWh on that day whereas the surplus electricity amount
was approximately 5.14 MWh for a 20 MWp solar power plant.
The maximum deficit power was approximately 6.6 MW and
the maximum surplus power was approximately 4.5 MW. It
was a medium solar irradiance variability day. Based on solar
irradiance, the total power generation capability of the 20MWp
plant was approximately 81.2 MWh on that day. Therefore, the
support or surplus energy on that day was approximately 6%
of the total generation although support energy and surplus
energy are close to each other. The equivalent irradiance
change level per second has been decreased by a good amount.
The power ramp-rate control (PRRC) strategy is employed to
limit the fluctuation rate in the PV output power under
dynamically changing irradiance conditions (e.g., passing
clouds) [7]. The ramp level, very much related to the PRRC,
decreased significantly in Fig. 9 as indicated by the mark ‘A’
and ‘B’, where mark ‘A’ is the ramp level without support from
ESS and mark ‘B’ is the reduced ramp level to be obtained after
providing support from ESS. Due to the reduction of the ramp
level to a significant level with the proposed ESS support, the
output power of the solar photovoltaic power plant will be
more stable and its fluctuation level will be reduced. Category
wise some results have been described below.
4.2 Solar Irradiance in high variability days
The number of ramps and the ramp level are high on high
solar irradiance variability days which represents the short-
term variation especially due to the cloud movement. An
example scenario is discussed below.
Fig. 10. Example-1 of high solar irradiance variability day
Fig. 10 represents a high solar variability day where solar
ramping is very high and frequent. The lower irradiance level
of ramps was around 250 W/m2 and the higher irradiance level
was around 950 W/m2. After providing support from ESS
according to the developed model, the ramping level reduces
significantly and is reduced to approximately one-third in
sample ramp levels as shown by ‘A’ & ‘B’ in Fig. 10.
Equivalent irradiance change slopes decrease as well. Similar
another example is displayed in Fig. 11.
In Fig. 11, the ramp rate, i.e., dy/dt is found to be decreased
where the y-axis represents the irradiance level and the x-axis
represents the number of records with a fixed time interval. The
comparative power falling time has been highlighted in Fig. 11
between before and after providing support from ESS. The
equivalent power falling time has been increased compared
with the before providing support and hence power falling
occurs gradually instead of sharply. The ramp level also
decreases similarly as indicated in Fig. 10.
JOURNAL OF SUSTAINABLE ENERGY VOL. 14, NO. 2, DECEMBER, 2023
ISSN 2067-5534 © 2023 JSE 130
Fig. 11. Example-2 of high solar irradiance variability day
4.3. Solar Irradiance in small variability days
The small short-term solar irradiance variability days are
close to the clear weather days but little solar ramping is
present. The ramp number and level are not that much like high
solar irradiance variability days. An example is shown below.
Fig. 12. Examples of a small irradiance variability day
Fig. 12 represents a small variability day. Most of the time
of the day, it is like clear weather but variability occurs
sometimes. Due to this sudden change in weather and solar
irradiance, the output of the SPV plant also changed.
Therefore, this change is a cause of load flow direction change
in the interconnection area of the utility network. After
providing the necessary power and energy support from ESS,
this variation can be reduced to a significant level.
4.4. Solar Irradiance on washout days
The washout day is cloud-covered and receives very little
solar irradiance in a day. In most cases, partly diffuse
irradiation is received on that day and variation may be also
less as indicated in the sample figure given below.
Fig. 13. Example of a washout day
Fig. 13 is a representation of a washout day where the total
solar irradiation received on that day is 8.16 MWh according
to the recorded data. The maximum solar irradiance received
on that day is 200 W/m2 only for a few hours and for the rest
of the hours, it is within 50 W/m2. Although it is a full cloud-
covered day, the short-term solar irradiance variation is very
less. According to the calculation, the deficit and surplus of
both energies on that day is 0.33 MWh, which is much less
compared with other days.
4.5. Solar Irradiance on clear weather days
Clear weather days are cloud-free days where solar
modules receive maximum solar irradiance from the sun. In
most of the cases, short-term variation of solar irradiance was
absent or may present very little. An example is shown in Fig.
14.
Fig. 14. Example of a clear weather day
Fig. 14 is an example of a clear weather day where the
short-term variation of solar irradiance is close to zero. Also,
the daily maximum solar irradiance is close to 1000 W/m2 and
the total energy of that day is 143 MWh according to the
recorded data. Solar energy generation on this day is more than
17 times higher than the energy on washout day mentioned in
Fig. 13.
Solar Irradiance (W/m²)
JOURNAL OF SUSTAINABLE ENERGY VOL. 14, NO. 2, DECEMBER, 2023
ISSN 2067-5534 © 2023 JSE 131
4.6. Full-year assessment results
A one-year daily solar irradiance pattern was analysed
with the developed fuzzy model. Daily maximum power
support from ESS depends on the irradiance change level.
Daily energy support from ESS depends on the irradiance
change level and the number of ramps. Using the MATLAB
program, daily energy support from ESS, Surplus energy
amount for ramp management, and maximum power for deficit
or surplus energy were calculated. Based on the calculation, a
full-year daily ramp management result has been analysed and
a summary is shown in Table 2.
Table 2. A sample of the daily variation status of the output power of a 20MWp Solar Power Plant using CUET irradiance
data
1
2
3
4
5
6
7
8
9
10
11
DATE
Plant
Capa
city
(MW
p)
Irradiati
on
(kWh/m
2/day)
Irradia
tion
after
RM
(kWh/
m2/day
)
Plant
Expected
Output
(MWh)
RMO
(MWh)
PEO-
SURE
(MWh)
Deficit
Energy
(MWh)
Surplu
s
Energy
(MWh)
Max
Deficit
Power
(MW)
Max
Surplus
Power
(MW)
01-01-2020 20 3.04 3.00 60.82 59.92 57.00 2.92 3.82 3.71 4.22
02-01-2020 20 1.98 1.95 39.55 38.95 36.83 2.12 2.72 2.27 4.79
03-01-2020 20 1.48 1.46 29.67 29.13 27.35 1.78 2.32 1.56 3.35
˜ ˜
˜ ˜
˜ ˜
˜ ˜
˜ ˜
˜ ˜
˜ ˜
˜ ˜
˜ ˜
˜ ˜
˜ ˜
29-12-2020 20 4.53 4.53 90.62 90.61 89.60 1.00 1.01 0.33 0.62
30-12-2020 20 4.67 4.67 93.44 93.43 92.41 1.02 1.02 0.32 0.66
31-12-2020 20 4.64 4.64 92.75 92.74 91.74 0.99 1.01 0.33 0.82
Average
4.51 4.52 90.25 90.43 86.19 4.24 4.06 4.60 4.59
Maximum
7.18 7.16 143.66 143.29 141.52 12.14 11.51 11.65 12.73
Minimum
0.41 0.41 8.16 8.16 7.83 0.33 0.33 0.28 0.27
Sum(352d)
1588 1592 31769 31831 30339 1491 1430 - -
Eq.Sum(365d)
1647 1650 32943 33006 31460 1546 1483 - -
In Table 2, column 1 represents the date and column 2
represents the capacity of the SPV plant that has been
considered for the case study. Column 3 represents the actual
GHI solar irradiation received in the case study site in
kWh/m2/day measured by the pyranometer. Column 4
indicates the modified output equivalent input GHI solar
irradiance data in kWh/m2/day after using the fuzzy model.
Columns 5 & 6 indicate the energy output of that day’s before
and after ramp management support respectively. Column 7 is
the difference between column 5 and column 9 which
represents the expected energy output of the SPV plant other
than surplus energy. Columns 8 & 9 are the daily
deficit/support energy and surplus energy for ramp
management arrangement calculated according to the fuzzy
model mentioned in the methodology section. Columns 10 &
11 are the calculated maximum deficit and surplus power
respectively according to the ramp management support. A full
one-year assessment has been conducted where the first 3 days
and last 3 days are mentioned in the table and the rest of the
days (dotted) are hidden to make it simple. However, all data
should be reflected in the figures shown below. Finally, the
average, maximum, minimum, and summation records of the
year have been calculated in the few last rows.
The daily deficit energy values mentioned in column 8 of
Table 2 are plotted and shown in Fig. 15.
Fig. 15, it is shows that the first few months and the last
few months of the year have required a small amount of ramp
management support energy and it is less than 2 MWh/day
except on some days. This support requirement is increasing in
the summer season and goes up to 12 MWh/day energy support
requirements. Similarly, to avoid the sharp rise of the SPV
plant's output and minimize the ramp rate, some surplus
energies have been identified. It is mentioned in column 9 of
Table 2.
Fig. 15. Daily deficit energy for Ramp management, 2020
The plotted values of this result are shown in Fig. 16.
JOURNAL OF SUSTAINABLE ENERGY VOL. 14, NO. 2, DECEMBER, 2023
ISSN 2067-5534 © 2023 JSE 132
Fig. 16. Daily surplus energy from Ramp management,
2020
Fig. 16 looks similar to Fig. 15. It is shown that the first
few months and last few months of the year have required small
ramp management surplus energy and it is less than 2
MWh/day except for some days. This surplus energy is
increasing in the summer season and goes up to 12 MWh/day.
The comparison analysis of ramp management deficit energy
versus surplus energy on a daily basis has been analysed and
shown in Fig. 17 below.
Fig. 17. Comparing the daily deficit energy versus surplus
energy
It can be observed from Fig. 17 that the daily deficit energy
and surplus energy for ramp management are almost close to
each other with a small difference. Also, it is observed that all
days of the summer season are not similar. Some days require
high ramp management support whereas the next day’s
requirement is not that much. It depends on the weather
condition of that area, especially cloud pattern and movement;
and it’s a location-specific issue.
The energy and power are proportional to each other but a
high-power requirement could be involved with more power
electronic equipment where the cost will be significantly
involved. The maximum deficit and surplus power have been
analysed daily and shown in columns 10 and 11 of Table 2. The
plotted figure of those values is shown in Fig. 18.
Fig. 18. Comparing the daily maximum deficit power
versus surplus power
The power requirement according to Fig. 18 is limited to
12 MW of which 4 – 8 MW is the common scenario for most
days. There is some difference between the daily maximum
deficit and surplus power although the maximum one needs to
be counted. Finally, the deficit and surplus power and energy
scenario calculated by the model is presented in
Table 3.
In
Table 3, the ITEM/RANGE row is the range of Power
(MW) or Energy (MWh). The rest of the values represent the
number of days on which the power or energy amount was
within this range. It is noted that it could be varied with the
support instruction provided by the model and final output
shaping will change accordingly. More support will provide
more stable output and less support will provide less stable
output, but the cost is associated with it accordingly. The
summarized results are shown in Fig. 19 and Fig. 20 for the
power and energy scenarios separately.
Table 3. Summary of Power Variation and Energy Variation of a Solar Power Plant, in number of days in a year (352d)
ITEM/RANGE (MW/ MWh) 0-2 2-4 4-6 6-8 8-10 >10
Maximum Surplus Power
Number of
days
99 55 71 75 41 11
Max Deficit Power 103 49 67 79 48 6
Surplus Energy 112 79 71 54 27 9
Deficit Energy
120
69
54
66
29
14
JOURNAL OF SUSTAINABLE ENERGY VOL. 14, NO. 2, DECEMBER, 2023
ISSN 2067-5534 © 2023 JSE 133
Fig. 19. Energy deficit and surplus statistics in a year
Fig. 20. Maximum deficit and
surplus power statistics in a year
From the above Fig. 19 and Fig. 20, a minimum 8 MWh
ESS with 8 MW charging/discharging capacity can be
considered for solar ramp management support. Also, we
observed that the deficit and surplus energy and power are
nearly close to each other. Therefore, around 50% of the
storage can be charging state, and the remaining half can be
discharging state up to the safe discharging level, then it could
be swapped. The battery type must be capable of dynamic
charging and discharging for solar ramp management. The
lithium-ion battery technology is the most popular grid-scale
stationary energy storage technology as it has a fast response
capacity and high specific energy [20]. Small energy storage
units close to the source of power quality disturbance are cost-
effective and offer excellent potential for widespread
implementation in the low-voltage distribution grid [33]. The
effect on the cycle life of energy storage needs to be analysed
for ramp management support purposes that could be discussed
in a separate paper.
4.7. Cost Scenarios
Gravity energy storage, a novel energy storage system,
compares its performance with alternative energy storage
systems used in large-scale applications such as Pumped-hydro
energy storage (PHES), Compressed air energy storage
(CAES), Sodium Sulfur (NaS), and Li-ion batteries (Li-ion).
The GES is the most cost-effective large-scale energy storage
technology for storage capacities of more than 1 GWh. In
addition, for a 1 GW power capacity and 125 MWh energy
capacity system, gravity energy storage (GES) has an attractive
LCOS of 202 $/MWh [34]. The Liner Electric Machine-based
GESS is about 26% more cost-effective than the currently
competitive flywheel energy storage technology whereas this
technology is more sensitive in terms of capital expenditure,
efficiency, discount rate and discharge duration [35].
The comprehensive review of the paper [36] shows that
the lithium-ion battery fits both low and medium-size
applications with high power and energy density requirements
in the electrochemical storage category. According to the
report [37] of National Renewable Energy Laboratory (NREL)
of the USA, the fixed-Tilt utility-scale PV benchmark cost was
$0.89/WDC and the one-axis tracker utility-scale PV
benchmark cost was $0.96/WDC in the first quarter of 2020.
Similarly, the fixed-Tilt utility-scale PV benchmark cost was
$0.83/WDC and the one-axis tracker utility-scale PV
benchmark cost was $0.89/WDC in the first quarter of 2021
where the assessment was conducted considering 100 MWDC
plant capacity [38].
Li-Ion Utility-Scale Storage and PV-Plus-Storage Model
have been calculated by NREL in the report [39]. In this model,
the cost of a Stand-alone 100-MWDC PV system with one-axis
tracking was $89 million. The cost of a Stand-alone 60-MWDC
/240-MWh Usable, 4-hour-duration energy storage system was
$90 million. For DC-coupled PV (100-MWDC) plus storage
(60-MWD/AC/240-MWhUsable, 4-hour-duration) system, the cost
was $168 million, whereas AC-coupled PV (100-MWDC) plus
storage (60-MWD/AC/240-MWhUsable, 4-hour-duration) system,
the cost was $167 million. In those cases, the cost difference
between DC-coupled and AC-coupled utility-scale energy
storage systems were not a significant amount. Finally, the
investment cost of PV (100-MWDC) and storage (60-
MWD/AC/240-MWhUsable, 4-hour-duration) systems sited in
different locations was $179 million. The Li-ion battery
cabinet cost is approximately 59% of the total storage cost
(4hr) whereas it will be approximately 44% for a 1hr duration
with similar storage capacity. A storage cost scenario from this
report is shown below.
Fig. 21. U.S. utility-scale Li-ion battery stand-alone
storage costs for durations of 0.5–4.0 hours (60 MWDC),
Q1 2021 [39]
Number of Days
JOURNAL OF SUSTAINABLE ENERGY VOL. 14, NO. 2, DECEMBER, 2023
ISSN 2067-5534 © 2023 JSE 134
In Fig. 21, $/kWhusable and $/kWhnameplate data are presented
for 0.5-hour, 1-hour, 2-hour, and 4-hour conditions. Here, a
Lithium-ion battery cabinet holds the maximum percentage of
cost in each category. However, due to the increasing charging
and discharging power capacity, lower hour scenarios have
more battery central inverter and similar other costs, known as
power electronic cost. According to the approximate cost
projection of our proposed capacity, the 20 MWDC solar
project’s cost should be around $18 million. Whereas Li-ion
energy storage cost of 8 MWh with 8 MW power capacity (1
hour) should be around $5 million, which is 28% of the solar
project cost.
However, the volume-weighted price of lithium-ion
battery packs across all sectors averaged $152/kWh according
to the assessment of BloombergNEF in 2022 [40]. The energy
storage systems whose total costs are dominated by power
component costs ($/kW) are better suited for longer-term
energy storage and those dominated by energy (storage)
component costs ($/kWh) should be used for shorter-term
energy storage [41]. For the promotion of wide-scale Li-ion
energy storage, the key challenges are fire safety and recycling,
instead of capital cost, battery cycle life, or
mining/manufacturing challenges [42]. Round-trip efficiency,
the ratio of useful energy output to useful energy input, is
identified as 86% and the 2022 Annual Technology Baseline
(ATB) adopts this value [43]. The cost and performance of the
battery systems are based on an assumption of approximately
one cycle per day. The fixed operation and maintenance costs
include battery replacement costs, based on assumed battery
degradation rates that drive the need for 20% capacity
augmentations after 10 and 20 years to return the system to its
nameplate capacity [39].
5. CONCLUSION
Global warming is a challenge considering sustainable
development and it is essential to move from the dependency
on fossil fuels to renewable energy sources, where Solar PV is
a suitable one. The intermittent nature of Solar PV energy is
one of the main challenges for the promotion of large-scale
power plants, especially for grid connected systems. For a high
share of such VRE in grid should have some auxiliary support
system to reduce this short-term variation of the output power
of the Solar PV power plant that is generated due to the short-
term variation of Solar irradiance. Additionally, the effects
associated with the utility system will be reduced which will
be helpful including frequency regulation and supply-demand
management by the transmission system operator. Therefore,
before the establishment of a large-scale Solar PV plant in a
location, grid integration effects of short-term power variation
should be conducted for the proposed interconnection point.
In this paper, a feasibility study of SPV power smoothing
has been conducted using the Fuzzy Logic approach. This
analysis is conducted for a 20 MWp SPV plant based on the 1-
year’s time series solar irradiance dataset at the case study site.
A fuzzy model has been used to determine the smoothing
amount in terms of power and energy. The smoothing level
may be adjusted by changing the degree of membership
function. According to the assessment of the whole year’s data,
a minimum 8 MWh ESS with 8 MW charging/discharging
capacity has been calculated for the ramp management support,
which will increase approximately 28% of the solar project
cost. The daily requirement of support amount and surplus
amount in terms of energy and power were found identical.
The actual scenarios may be slightly different due to the
geographical dispersion, cloud enhancement, and similar other
effects. It may be relaxed for low VRE penetration, but
essential for high VRE penetration into the grid. This is a site-
specific feasibility study of Solar photovoltaic power
smoothing using the Fuzzy Logic approach that represents a
scenario of Bangladesh. It is recommended to analyse with
real-time output power and energy data of a solar power plant
with a precise timescale.
DECLARATION OF COMPETING INTEREST
The authors declare that they have no known competing
financial interests or personal relationships that could have
appeared to influence the work reported in this paper.
ACKNOWLEDGEMENTS
We thank the financial support from the Bangladesh
University of Engineering and Technology (BUET) through
the Committee for Advanced Studies and Research (CASR).
REFERENCES
[1]. Jazayeri M, Jazayeri K, Uysal S. Generation of spatially dispersed
irradiance time-series based on real cloud patterns. Solar Energy
2017;158:977–94. https://doi.org/10.1016/j.solener.2017.10.026.
[2]. Wei C-C. Predictions of Surface Solar Radiation on Tilted Solar
Panels using Machine Learning Models: A Case Study of
Tainan City, Taiwan. Energies (Basel) 2017;10:1660.
https://doi.org/10.3390/en10101660.
[3]. Kim J, Lee S, Chong KT. A Study of Neural Network
Framework for Power Generation Prediction of a Solar Power
Plant. Energies (Basel) 2022;15:8582.
https://doi.org/10.3390/en15228582.
[4]. Bhavani M, Vijaybhaskar Reddy K, Mahesh K, Saravanan S.
Impact of variation of solar irradiance and temperature on the
inverter output for grid connected photo voltaic (PV) system at
different climate conditions. Mater Today Proc 2023;80:2101–
8. https://doi.org/10.1016/j.matpr.2021.06.120.
[5]. Martins J, Spataru S, Sera D, Stroe D-I, Lashab A. Comparative
Study of Ramp-Rate Control Algorithms for PV with Energy
Storage Systems. Energies (Basel) 2019;12:1342-1–15.
https://doi.org/10.3390/EN12071342.
[6]. Lave M, Kleissl J, Stein J. Quantifying and Simulating Solar-
Plant Variability Using Irradiance Data. Solar Energy
Forecasting and Resource Assessment, Elsevier; 2013, p. 149–
69. https://doi.org/10.1016/B978-0-12-397177-7.00007-3.
[7]. Yang Y, Kim KA, Blaabjerg F, Sangwongwanich A. Flexible
active power control of PV systems. Advances in Grid-
Connected Photovoltaic Power Conversion Systems, Elsevier;
2019, p. 153–85. https://doi.org/10.1016/B978-0-08-102339-
6.00006-3.
[8]. Alam MdR, Bhuiyan IU, Mohammad N. Variability Assessment of
Solar Irradiance for the Safe Grid Integration of Solar Photovoltaic
Power Plants. Journal of Renewable Energy and Environment 2023.
https://doi.org/10.30501/jree.2023.377735.1522.
[9]. Marcos J, Storkël O, Marroyo L, Garcia M, Lorenzo E. Storage
requirements for PV power ramp-rate control. Solar Energy
2014;99:28–35. https://doi.org/10.1016/j.solener.2013.10.037.
JOURNAL OF SUSTAINABLE ENERGY VOL. 14, NO. 2, DECEMBER, 2023
ISSN 2067-5534 © 2023 JSE 135
[10]. Schnabel J, Valkealahti S. Energy Storage Requirements for PV
Power Ramp Rate Control in Northern Europe. International
Journal of Photoenergy 2016;2016:1–11.
https://doi.org/10.1155/2016/2863479.
[11]. Bangladesh Energy Regulatory Commission (BERC).
Bangladesh Energy Regulatory Commission (Electricity Grid
Code) Regulations, 2019. Dhaka: 2020.
[12]. Meena O, Rajpoot A. Empirical Study of Solar PV Integration
in Smart Grid Analyzing Issues and Challenges. 14th
International Conference on Recent trends in Engineering,
Applied Science and Management, 2019, p. 223–9.
[13]. Marcos J, Marroyo L, Lorenzo E, García M. Smoothing of PV
power fluctuations by geographical dispersion. Progress in
Photovoltaics: Research and Applications 2012;20:226–37.
https://doi.org/10.1002/PIP.1127.
[14]. Rowlands IH, Kemery BP, Beausoleil-Morrison I. Managing
solar-PV variability with geographical dispersion: An Ontario
(Canada) case-study. Renew Energy 2014;68.
https://doi.org/10.1016/j.renene.2014.01.034.
[15]. Järvelä M, Lappalainen K, Valkealahti S. Characteristics of the
cloud enhancement phenomenon and PV power plants. Solar
Energy 2020;196:137–45.
https://doi.org/10.1016/j.solener.2019.11.090.
[16]. van Haaren R, Morjaria M, Fthenakis V. Empirical assessment
of short‐term variability from utility‐scale solar PV plants.
Progress in Photovoltaics: Research and Applications
2014;22:548–59. https://doi.org/10.1002/pip.2302.
[17]. Järvelä M, Lappalainen K, Valkealahti S. Characteristics of the cloud
enhancement phenomenon and PV power plants. Solar Energy
2020;196:137–45. https://doi.org/10.1016/j.solener.2019.11.090.
[18]. Y Lakshmi PAP and. Active Power Control of Grid tied PV
system using Fuzzy Controller. International Journal for
Modern Trends in Science and Technology 2020;6:1–4.
https://doi.org/10.46501/ijmtst060801.
[19]. Liu H, Peng J, Zang Q, Yang K. Control Strategy of Energy
Storage for Smoothing Photovoltaic Power Fluctuations. IFAC-
PapersOnLine 2015;48:162–5.
https://doi.org/10.1016/j.ifacol.2015.12.118.
[20]. Wali SB, Hannan MA, Ker PJ, Rahman MA, Mansor M,
Muttaqi KM, et al. Grid-connected lithium-ion battery energy
storage system: A bibliometric analysis for emerging future
directions. J Clean Prod 2022;334:130272.
https://doi.org/10.1016/j.jclepro.2021.130272.
[21]. Fathima AH, Palanisamy K, Padmanaban S, Subramaniam U.
Intelligence-based battery management and economic analysis
of an optimized dual-Vanadium Redox Battery (VRB) for a
wind-PV hybrid system. Energies (Basel) 2018;11.
https://doi.org/10.3390/en11102785.
[22]. Gür TM. Materials and technologies for energy storage: Status,
challenges, and opportunities. MRS Bull 2021;46:1153–63.
https://doi.org/10.1557/s43577-021-00242-w.
[23]. Ud-Din Khan S, Wazeer I, Almutairi Z, Alanazi M. Techno-
economic analysis of solar photovoltaic powered electrical
energy storage (EES) system. Alexandria Engineering Journal
2022;61:6739–53. https://doi.org/10.1016/J.AEJ.2021.12.025.
[24]. Diaz VS, Cantane DA, Santos AQO, Ando Junior OH.
Comparative Analysis of Degradation Assessment of Battery
Energy Storage Systems in PV Smoothing Application.
Energies (Basel) 2021;14. https://doi.org/10.3390/en14123600.
[25]. Alam MJE, Saha TK. Cycle-life degradation assessment of
Battery Energy Storage Systems caused by solar PV variability.
IEEE Power and Energy Society General Meeting 2016;2016-
Novem:0–4. https://doi.org/10.1109/PESGM.2016.7741532.
[26]. Sukumar S, Marsadek M, Agileswari KR, Mokhlis H. Ramp-
rate control smoothing methods to control output power
fluctuations from solar photovoltaic (PV) sources—A review. J
Energy Storage 2018;20:218–29.
https://doi.org/10.1016/j.est.2018.09.013.
[27]. Sonia M, Rao KK. Application of Fuzzy Controller for
Improvement of Power Quality using UPQC. International
Journal for Modern Trends in Science and Technology
2020:27–30. https://doi.org/10.46501/IJMTST060806.
[28]. Atif A, Khalid M. Fuzzy logic controller for solar power
smoothing based on controlled battery energy storage and
varying low pass filter. IET Renewable Power Generation
2020;14:3824–33. https://doi.org/10.1049/iet-rpg.2020.0459.
[29]. Sivaneasan B, Yu CY, Goh KP. Solar Forecasting using ANN
with Fuzzy Logic Pre-processing. Energy Procedia
2017;143:727–32.
https://doi.org/10.1016/j.egypro.2017.12.753.
[30]. Kumar Rajalwal Sameena Elyas Mubeen Priyanka Shrivastava
A. A Survey on Smart Grid Load Balancing Techniques and
Challenges. International Journal of Scientific Research &
Engineering Trends 2018;4:541–4.
[31]. Li M, Shan R, Virguez E, Patiño-Echeverri D, Gao S, Ma H.
Energy storage reduces costs and emissions even without large
penetration of renewable energy: The case of China Southern
Power Grid. Energy Policy 2022;161:112711.
https://doi.org/10.1016/J.ENPOL.2021.112711.
[32]. Chai Z, Chen X, Yin S, Jin M, Wang X, Guo X, et al.
Construction of a new levelled cost model for energy storage
based on LCOE and learning curve. E3S Web of Conferences
2022;338:01049.
https://doi.org/10.1051/e3sconf/202233801049.
[33]. Smolenski R, Szczesniak P, Drozdz W, Kasperski L. Advanced
metering infrastructure and energy storage for location and
mitigation of power quality disturbances in the utility grid with
high penetration of renewables. Renewable and Sustainable
Energy Reviews 2022;157:111988.
https://doi.org/10.1016/j.rser.2021.111988.
[34]. Berrada A. Financial and economic modeling of large-scale
gravity energy storage system. Renew Energy 2022;192:405–
19. https://doi.org/10.1016/j.renene.2022.04.086.
[35]. Mugyema M, Botha CD, Kamper MJ, Wang R-J, Sebitosi AB.
Levelised cost of storage comparison of energy storage systems
for use in primary response application. J Energy Storage
2023;59:106573. https://doi.org/10.1016/j.est.2022.106573.
[36]. Kebede AA, Kalogiannis T, Van Mierlo J, Berecibar M. A
comprehensive review of stationary energy storage devices for
large scale renewable energy sources grid integration.
Renewable and Sustainable Energy Reviews 2022;159:112213.
https://doi.org/10.1016/j.rser.2022.112213.
[37]. Feldman D, Ramasamy V, Fu R, Ramdas A, Desai J, Margolis
R. U.S. Solar Photovoltaic System and Energy Storage Cost
Benchmark (Q1 2020). Golden, CO (United States): 2021.
https://doi.org/10.2172/1764908.
[38]. Cole W, Frazier AW, Augustine C. Cost Projections for Utility-
Scale Battery Storage: 2021 Update. 2021.
[39]. Ramasamy V, Feldman D, Desai J, Margolis R. U.S. Solar
Photovoltaic System and Energy Storage Cost Benchmarks: Q1
2021.
[40]. Top 10 Energy Storage Trends in 2023 | BloombergNEF n.d.
https://about.bnef.com/blog/top-10-energy-storage-trends-in-
2023/ (accessed July 21, 2023).
[41]. Augustine C, Blair N. Energy Storage Futures Study: Storage
Technology Modeling Input Data Report. 2021.
[42]. Huang Y, Li J. Key Challenges for Grid‐Scale Lithium‐Ion
Battery Energy Storage. Adv Energy Mater 2022;12:2202197.
https://doi.org/10.1002/aenm.202202197.
[43]. Mongird K, Viswanathan V, Alam J, Vartanian C, Sprenkle V,
Baxter R. 2020 Grid Energy Storage Technology Cost and
Performance Assessment. Washington, D.C: 2020.
