energies

Article
Sizing of Battery Energy Storage Systems for Firming PV Power
including Aging Analysis
Juan A. Tejero-Gómez and Ángel A. Bayod-Rújula *

Department of Electrical Engineering, University of Zaragoza, 50018 Zaragoza, Spain; [email protected]

* Correspondence: [email protected]

Abstract: The variability of solar radiation presents significant challenges for the integration of
solar photovoltaic (PV) energy into the electrical system. Incorporating battery storage technologies
ensures energy reliability and promotes sustainable growth. In this work, an energy analysis is
carried out to determine the installation size and the operating setpoint with optimal constant
monthly power through an iterative calculation process, considering various operating setpoints and
system parameters. A degradation model is integrated according to the curves offered by battery
manufacturers and the charge–discharge cycles are calculated using the rainflow method to guarantee
a reliable analysis of the plant. Through massive data analysis in a long-term simulation, indicators
are generated that allow for establishing a relationship between the energy unavailability of the
system and the BESS dimensions.

Keywords: large-scale battery energy storage system; utility-scale PV-plus-battery; aging; battery
degradation; firm power

1. Introduction
The global energy system is currently undergoing a transition towards a new paradigm
characterized by decarbonization, the decentralization of generation, the electrification of
the economy and a more sustainable use of resources. Thanks to technological advances and
cost reduction, the implementation of electrical energy production systems is increasingly
Citation: Tejero-Gómez, J.A.;
high. They represent a high percentage of the total power of the generation system and
Bayod-Rújula, Á.A. Sizing of Battery
contribute to an increasing extent to the production mix of electric power. The increase
Energy Storage Systems for Firming
PV Power including Aging Analysis.
in renewables as a share of energy supply in 2022 was the second largest in history. A
Energies 2024, 17, 1485. https://
total of 25% of the electricity generated worldwide was the responsibility of renewable
doi.org/10.3390/en17061485
photovoltaic and wind generation plants [1]. Even faster increases are needed to align with
the Net Zero Emissions Scenario [2]. The European Green Deal establishes the pace of
Academic Editor: Abdul-Ghani Olabi
decarbonization required in the coming years [3].
Received: 23 February 2024 Electrical generation from renewable energies is associated with the resources (wind,
Revised: 13 March 2024 photovoltaic) and is variable, intermittent and not dispatchable. Whilst the penetration of
Accepted: 17 March 2024 these generation sources is not high, they do not pose major problems for the operation and
Published: 20 March 2024 security of the electrical networks to which they are connected. But the high implementation
of renewable energy is a challenge [4]. High voltage variations and fluctuations or strong
ramps may occur in the generation curves. This requires greater power capacity available
in the electrical network and requires other flexible systems, both for generation and the
Copyright: © 2024 by the authors.
response to demand.
Licensee MDPI, Basel, Switzerland.
In the planning process for new plants, power companies set limitations on the amount
This article is an open access article
of power that can be connected. Relevant limitations are usually placed on the maximum
distributed under the terms and
power that can be injected into the network to limit these voltage fluctuations. The studies
conditions of the Creative Commons
Attribution (CC BY) license (https://
are usually conducted based on considering the most unfavorable situation for the system,
creativecommons.org/licenses/by/
that is, moments of very high production (the nominal power of the inverters is usually
4.0/).
considered). However, such a high generation is produced in very few hours a year. This

Energies 2024, 17, 1485. https://doi.org/10.3390/en17061485 https://www.mdpi.com/journal/energies

Energies 2024, 17, 1485 2 of 30

hinders the further implementation of photovoltaic plants, or the need to reinforce lines
and substations. The operation of the electric companies’ assets is therefore very restricted
since their capacity to conduct electrical energy is limited by the maximum values, and
most of the time they are thus underused. But the maximum power is produced in very
few moments a year, so the capacity of the networks is underutilized.
Owners of photovoltaic installations are also affected due to these characteristics
of their systems. With the progressive increase in photovoltaic production during the
central hours of the day, MWh prices tend to be very low precisely at the times of greatest
production. This effect called “cannibalization” represents a reduction in its benefits, so its
economic viability is compromised. At times when there is excess generation, curtailments
are necessary, which further reduce its economic viability. These curtailments are becoming
higher in some countries. On the other hand, energy generation from photovoltaic plants is
subject to forecasting errors, which incur penalizing costs [5].
Accuracy in long-term estimates of energy production, with time horizons greater
than one year, is of utmost importance for the technical planning of future distributed
generation (DG) units. An adequate estimate allows for the correct allocation of resources,
the optimization of the infrastructure and the better integration of the DG units into the
energy system. As baseload generation from the combustion of fossil fuels is replaced
by renewable energy sources; ensuring reliable capacity firming support for the power
grid will become crucial. Firm energy determines the maximum volume of energy that a
generation unit can sell with a given level of reliability. The capacity value (or “capacity
credit”) indicates the extent to which ERVs can be relied upon just like conventional power
plants. The capacity value of solar photovoltaic (PV) is very low [6–8].
The definition of the appropriate mechanisms to achieve the complete integration of
renewable energies into the energy system is still under development. The selection of
these mechanisms will depend on a variety of factors, including the climatic conditions of
the area, policies implemented by the government, the specific geographical location and
changes in costs that are revealed through techno-economic analyses [9].
In this context, storage systems are positioned as a necessary technology to allow for
the adequate grid integration of the growing renewable power. In order to move towards
the net zero scenario, storage capacity must be significantly increased, reaching an annual
average of 80 GW during the period between 2022 and 2030 [6], and it is necessary to
exploit different options. In particular, battery energy storage systems (BESSs) can offer
such robust capacity, giving the system management capabilities for the generated PV
energy. The storage industry is projected to grow to hundreds of times its current size in
the coming decades. The dataset [10] points to a considerable reduction in the prices of
lithium-ion storage systems in utility applications over the last decade. The average cost
has decreased from $1659/kWh in 2010 to $285/kWh in 2021.
Security of supply is a key factor in the electrical system. Installing batteries in parallel
reduces overgeneration and the cutting and sale of energy during peak production hours.
BESSs can offer firm capacity, providing the system with management capabilities for
the generated photovoltaic energy. However, for economic reasons, it has been rare until
now to discuss the operation of large-scale PV plants operating as baseload generation
plants. From the point of view of hosting capacity, the power to be taken into account
for the PV-BESS would no longer be the peak power of the collector field, but rather the
constant power to be delivered to the network. This increases the number of facilities that
fit into a given distribution network [11]. For example, in Spain, the total nominal power of
the distributed generators (DGs) should be less than 50% of the nominal capacity of the
Medium Voltage/Low Voltage (MV/LV) transformer at the interconnection point [12].
Another benefit of capacity firming is that the system and all its components operate
close to their nominal power ratings for a longer time. Furthermore, the cost of converters
in large-scale PV plants will be lower due to the power output smoothing that is supplied
to the grid.
Energies 2024, 17, 1485 3 of 30

Unlike conventional substations, a substation for a large-scale PV plant has an uneven

utilization coefficient due to the characteristics of solar radiation. BESSs can shift the peak of
daytime photovoltaic production to the night. Furthermore, the capacity of the transformer
of the substation connected to the photovoltaic plant can be decreased by directing the
generation peaks of the solar panels towards the BESS [9] for subsequent supply.
Due to all of the above, the concept of constant photovoltaic power generation (PV-
CPG) arises to overcome these problems. The problems of voltage variations due to variable
currents, ramps, the need for backup sources, the cannibalization effect, etc., are reduced.
References [13–16] accurately determine the optimal capacity of battery energy storage
systems (BESSs) when integrated with renewable energy sources, taking into account
the effects of battery degradation. Although the main thrust of these studies bears some
resemblance to the present manuscript, they are not directly comparable. They focus on
economic analysis. Our manuscript aims to completely abstract from economic analysis
and focus exclusively on energy aspects. Furthermore, this model lacks comparability in
terms of degradation, since the system does not provide a 24 h baseload.
In [14], iterative methods are used, similar to the approach presented in this manuscript.
However, it is important to note that the proposed degradation model is rather simplified
and is based solely on another author’s model.
Study [15] is consistent with our manuscript but focuses specifically on extremely fast
charging stations.
The proposed system is designed to provide firm power or backup capacity to the
electricity system. This steady flow of power from the generating plant is guaranteed with
a high level of reliability. Therefore, its massive integration into the power grid could lead
to a reduction in the need for frequency regulation technologies, as a high level of available
power is guaranteed with reduced uncertainty. Some examples of firming power in the
literature are the following.
In [8], the goal is to achieve constant power output from 24 h to a few hours during
the day. In [17], a battery management and control algorithm mitigates fluctuations in
intermittent PV generation for hourly periods. In [18], the required energy storage capacity
is estimated in order to guarantee the constant and reliable delivery of energy by wind and
photovoltaic means in Japan’s electrical system.
A 50 MW AC PV system with 60 MW/240 MWh battery storage modeled in California
can provide >98% capacity factor over a 7–10 p.m. target period with a lower lifetime
cost of operation than a conventional combustion turbine natural gas power plant [19].
However, this scenario is conservative, and it is not proposed to establish a 24 h baseload
as proposed in the present study. In a similar way, in [20], the storage is dimensioned for a
large PV plant with storage in baseload operation.
In [5], PV production is regulated with BESSs to maintain a constant output level,
which significantly contributes to power grid stability, but only in a certain time interval
during the day. In [21], an analysis of the optimal relationship between the optimal size of
the required battery and the peak power of the generation system is presented, as well as
the optimal operating setpoint of a PV + STG system to provide firm capacity all year.
Given that the solar resource is notably lower in the winter months, a similar analysis
was presented in [22] but for constant operating setpoints on a monthly basis.
All these cases mentioned are conservative and do not propose to establish a 24 h
baseload. They also do not take into account degradation phenomena in the BESS, as
proposed in this study.
The present study is conceived within the framework of a planning phase for the
electric grid and focuses on the technical optimization between a PV plant and the BESS,
where the main requirement is the supply of a 24 h baseload generation of the system,
defined as the production of constant and continuous energy 24 h a day. This means that
the generation does not fluctuate and remains stable at all times, meeting the definition of
baseload generation. In addition, a proven battery degradation model is integrated, and
 where the main requirement is the supply of a 24 h baseload generation of the system,
defined as the production of constant and continuous energy 24 h a day. This means that
Energies 2024, 17, 1485 4 of 30
the generation does not fluctuate and remains stable at all times, meeting the definition of
baseload generation. In addition, a proven battery degradation model is integrated, and
charge–discharge cycles are accurately calculated using the rainflow-counting algorithm
charge–discharge
to ensure reliablecycles
plant are accurately calculated using the rainflow-counting algorithm to
analysis.
ensure reliable plant analysis.
The proposed study identifies the optimal dimensions of a PV plant hybridized with
The proposed
a battery studysystem
energy storage identifies
thatthe optimal
supplies dimensions
a constant power of setpoint
a PV plant hybridized
during monthly
with a battery energy storage system that supplies a constant power
intervals. Through massive data analysis in a long-term simulation, indicators are setpoint during
gen-
monthly intervals. Through massive data analysis in a long-term simulation,
erated that establish a relationship between the energy unavailability of the system and indicators
are
thegenerated that establish
BESS dimensions. a relationship
Through between
an iterative the energy
calculation unavailability
process, consideringof various
the system
op-
and the BESS dimensions. Through an iterative calculation process, considering
erating instructions and system parameters, a comparative analysis is carried out with various
operating
the purpose instructions and system
of determining parameters,
the optimal a comparative
installation analysis
size for each is carried out with
scenario.
the purpose of determining the optimal installation size for each scenario.
The aging effects of lithium-ion battery storage systems have been considered, ac-
The aging
cording to the effects
curvesofoffered
lithium-ion batterymanufacturers.
by battery storage systemsThehave been considered,
replacement of the accord-
battery
ing to the curves offered by battery manufacturers. The replacement of the battery bank at
bank at the end of its useful life has also been taken into account.
the end of its useful life has also been taken into account.
This manuscript proposes an innovative and risky study that may not be economi-
This manuscript proposes an innovative and risky study that may not be economically
cally feasible at this time. It therefore focuses on calculating energy balances and is con-
feasible at this time. It therefore focuses on calculating energy balances and is conceived
ceived within the framework of a planning phase for the electric grid and focuses on the
within the framework of a planning phase for the electric grid and focuses on the technical
technical optimization between a PV plant and the BESS, where the main requirement is
optimization between a PV plant and the BESS, where the main requirement is the supply
the supply of a 24 h baseload generation of the system, defined as the production of con-
of a 24 h baseload generation of the system, defined as the production of constant and
stant and continuous energy 24 h a day.
continuous energy 24 h a day.
The remainder of this paper is organized as follows: Section 2 presents the proposed
The remainder of this paper is organized as follows: Section 2 presents the proposed
model detailing its equations, phases, parameters and indicators. In addition, the mod-
model detailing its equations, phases, parameters and indicators. In addition, the modeling
eling of the BESS degradation is included. The results of the simulation are shown in
of the BESS degradation is included. The results of the simulation are shown in Section 3.
Section
In Section3. 4,
Inconclusions
Section 4, conclusions
are drawn.are drawn.

2.2. Case
CaseStudy
Study
2.1.Problem
2.1. ProblemFormulation
Formulation
The
Theaim
aimisisto
totransform
transformaaphotovoltaic
photovoltaicproduction
productionprofile,
profile,which
whichhashasaacharacteristic
characteristic
bell-shaped
bell-shapedshape,
shape,into
intoaacontinuous
continuousandandconstant
constantsupply
supplyof ofpower
powerthat
thatremains
remainsuniform
uniform
throughout
throughout2424h ha day with
a day high
with guarantees
high of availability.
guarantees A typical
of availability. daily daily
A typical profileprofile
comparing
com-
both curves
paring bothiscurves
shownisin Figure
shown in1.Figure 1.

Figure 1. A typical profile for PV-BESS energy for a typical winter day (3 January).
Figure 1. A typical profile for PV-BESS energy for a typical winter day (3 January).

2.2. Proposed Model

In coherence with our previous research [21,22], the present study continues with
the analysis of the behavior of a PV-BESS hybrid plant. The quantification of the BESS
degradation and the appropriate assignment of the constant power are carried out for a
 degradation and the appropriate assignment of the constant power are carried out for a
correct evaluation of the system. The temporal frequency and the geographical point are
fundamental components when defining the methodology of the study. Time frequency
refers to the regularity with which data are collected and analyzed over time. The evalu-
Energies 2024, 17, 1485
ation of large-scale PV plants, integrated into the electrical grid, can be achieved effec-
5 of 30
tively through the use of hourly data [23]. The short-term variability inherent to photo-
voltaic generation is mitigated by the periodicity of solar radiation over a broader set of
time. The Sun exhibits cyclical patterns with a periodicity of around 11 years. For this
correct evaluation of the system. The temporal frequency and the geographical point are
reason, it is proposed to carry out a simulation that covers a period of 11 years starting
fundamental components when defining the methodology of the study. Time frequency
from 2006. On the other hand, the geographical point refers to the specific location where
refers to the regularity with which data are collected and analyzed over time. The evaluation
the data are collected. In the present case, the School of Engineering and Architecture of
of large-scale PV plants, integrated into the electrical grid, can be achieved effectively
the University
through the useof ofZaragoza,
hourly data located
[23]. onTheMaría de Luna
short-term Street in Zaragoza,
variability inherent toSpain, has been
photovoltaic
generation is mitigated by the periodicity of solar radiation over a broader set of time. to
selected as the geographical point of reference. Since plant PV production is linked Thethe
amount of solar irradiation, selecting a specific site involves collecting
Sun exhibits cyclical patterns with a periodicity of around 11 years. For this reason, it is data for repre-
sentativetoconditions
proposed carry out aunique to that
simulation thatregion.
coversHowever,
a period ofit 11
is possible to extrapolate
years starting from 2006. theOn re-
the other hand, the geographical point refers to the specific location where the data areis
sults to any region by modifying the irradiance data. Therefore, although the study
carried out
collected. in apresent
In the specificcase,
location, its findings
the School may have
of Engineering and much broader of
Architecture relevance and ap-
the University
plicability.
of Zaragoza, located on María de Luna Street in Zaragoza, Spain, has been selected as the
A simplified
geographical point of scheme of the
reference. proposed
Since plant PV PV-BESS system
production is illustrated
is linked in Figure
to the amount of 2. The
solar
photovoltaic
irradiation, plant transfers
selecting a specificthe
sitegenerated energy todata
involves collecting the for
electrical system through
representative conditions the
inverter group. For its part, the energy storage system (BESS) has
unique to that region. However, it is possible to extrapolate the results to any region by the capacity to inject
energy into
modifying thethe system atdata.
irradiance times when thealthough
Therefore, photovoltaic plant iscannot
the study carriedsatisfy
out inthe demand
a specific
or, alternatively, to store the surplus energy generated
location, its findings may have much broader relevance and applicability. by the photovoltaic system. It is
imperative that, scheme
A simplified at the point
of theofproposed
commonPV-BESS couplingsystem
(PCC),isaillustrated
constant powerin Figure setpoint
2. Theis
maintained to
photovoltaic guarantee
plant transfersthethe
reliability
generated of the system.
energy to the electrical system through the
inverterBygroup.
analyzing extensive
For its part, thephotovoltaic
energy storage production datasets,
system (BESS) hasthere is a tendency
the capacity to injectto
neutralize
energy into fluctuations
the system atover timesshort
when timethehorizons due toplant
photovoltaic incidents
cannot orsatisfy
cloud cover. The al-
the demand
gorithm
or, developed
alternatively, in MATLAB
to store the surplus R2021a
energy classifies photovoltaic
generated productionsystem.
by the photovoltaic according It isto
months, thus
imperative that,identifying
at the point inherent
of commonseasonal patterns.
coupling In this
(PCC), way, thepower
a constant systemsetpoint
power set-is
point is adjusted
maintained monthly,
to guarantee the based on the
reliability estimated
of the system.photovoltaic production.

Figure2.2.Scheme
Figure Schemeofofthe
thePV-BESS
PV-BESSsystem.
system.

By analyzing
The efficiencyextensive photovoltaic
of a PV-BESS production
is influenced by the datasets,
power there is a tendency
ratio between both to neu-
compo-
tralize
nents.fluctuations over short
Since the charging time horizons
operation of the due
BESStoisincidents
limited or
to cloud cover.generated
the power The algorithm
by the
developed
PV plant, in MATLAB
any power ofR2021a classifies
the BESS photovoltaic
that exceeds the powerproduction according togeneration
of the photovoltaic months,
thus
may be inefficient. Therefore, to avoid the underutilization of the BESS, a restrictionisis
identifying inherent seasonal patterns. In this way, the system power setpoint
adjusted monthly, based on the estimated photovoltaic production.
The efficiency of a PV-BESS is influenced by the power ratio between both components.
Since the charging operation of the BESS is limited to the power generated by the PV plant,
any power of the BESS that exceeds the power of the photovoltaic generation may be
inefficient. Therefore, to avoid the underutilization of the BESS, a restriction is established
that the power of the BESS must be less than the peak power of the photovoltaic generation
(PBESS = 1 MW).
In the context of a preliminary theoretical analysis, the sizing of the BESS is addressed
under a hypothesis of a daily balance between energy production and its supply. According
to this model, the BESS is charged during peak daytime generation hours and discharged
during the remaining hours to ensure the supply of a constant baseload. For the specific
case of the location of this study, according to the PVGIS website [24], it is determined that
Energies 2024, 17, 1485 6 of 30

the average daily production of a PV plant is 4.36 kWh/kW. Therefore, for a 1 MW PV

plant, the theoretical capacity of the BESS required to displace energy production at any
time of the day would be up to 4.36 MWh. Therefore, the approximate C-rate of the BESS
is 0.23.

2.3. Model Equations

In this paper, part of the model previously proposed in [21,22] is used. Following that,
the equations that characterize the PV-BESS model are briefly described.
For the PV plant, 2041 Tallmax TSM-DE15V(II) 490W monocrystalline photovoltaic
modules are selected, developed in China by Trinasolar for utility-scale PV applications [25].
Using Equations (1)–(3), the photovoltaic production for each hour is obtained. PV
output was simulated for fixed systems facing south at 35◦ tilt.
For an accurate assessment of PV production in a specific area, it is crucial to have
long periods of irradiance data. This is because solar irradiance is subject to significant
variations resulting from changes in meteorological and seasonal conditions. The long-term
solar radiation data are analyzed for a period of 11 years, taking the hourly irradiance from
the PVGIS database [24]. The temperature reached by the PV cells is defined in Equation (1)
when the module is exposed to operating conditions. This temperature is influenced by the
ambient temperature and irradiance.

NOCT − 20
TC = TA + GM · (1)
800
where
• TC is the module temperature [◦ C].
• TA is the ambient temperature [◦ C].
• GM is the global irradiance incident [W·m−2 ].
• NOCT is the Nominal Operating Cell Temperate [◦ C].
Equation (2) models how the power generated by a photovoltaic panel varies as a
function of cell temperature and incident irradiance.

(1 + γ·(TC − 25))
PMOD = PN · GM · (2)
GSTC

where
• PMOD is the module PV power [W].
• γ is the temperature coefficient of PN [%·◦ C−1 ].
• PN is the nominal power of the photovoltaic module [W].
• GSTC is the irradiance under standard measurement conditions [W·m2 ].
• GM is the irradiance [W·m2 ].
Finally, using Equations (1) and (2), the hourly power production of the PV plant is
calculated using Equation (3).

EPV = NP · PMOD ·CLOSS · TS (3)

where
• EPV is the hourly PV energy production.
• NP is the number of PV modules of the plant.
• CLOSS are the losses associated with the PV plant (p.u.).
• TS is the time slot of 1 h.
The accumulation of losses is quantified with the CLOSS coefficient of 13.4% [26–28].
Using Equations (4)–(7), the energy stored in the BESS in each hour is calculated,
taking into account the constraints and its equivalent model.
Energies 2024, 17, 1485 7 of 30

To adequately represent the energy balance of the BESS, the power restrictions specified
in Equations (4) and (5) must be taken into account. PMAX is equivalent to EB,MAX since the
maximum C-rate is 1.
The charging (PBC,t ) and discharging (PBD,t ) power of the storage system is PMAX ,
which is equivalent to its capacity, since the maximum C-rate is 1. During the charging
process of the BESS, the power constraint is determined by the difference between the
production of the PV plant and the energy supplied to the grid. In the discharge process of
the BESS, the power constraint is based on the difference between the energy supplied to
the grid and the output of the PV plant.
(
PMAX ∀ ( P PV,t − PGRID,t ) > PMAX
PBC,t = (4)
( P PV,t − PGRID,t ) ∀ ( P PV,t − PGRID,t ) ≤ PMAX

PMAX ∀ ( PGRID,t − PPV,t ) > PMAX
PBD,t = (5)
( P PV,t − PGRID,t ) ∀ ( PGRID,t − PPV,t ) ≤ PMAX
where
• PBC,t is the effective charging power of the BESS at the current time (t) in W.
• PBD,t is the effective discharge power of the BESS at the current time (t) in W.
• PMAX is the nominal power of the BESS in W.
• PPV,t is the photovoltaic power supplied at the current time (t) in W.
• PGRID,t is the power setpoint to be supplied to the electrical grid at the current time (t)
in W.
The stored energy at the current time (t) is related to the stored energy at the previous
time (t − 1) [29,30] according to Equation (6).
  
 EB,t−1 ·(1 − σ ) − ∆t·ηB,EF · PPV,t − PBC,t t ∈ charging
EB,t =  ηB,I NV (6)
 EB,t−1 ·(1 − σ ) − ∆t· PBD,t
ηB,I NV − PPV,t t ∈ discharging

where
• EB,t−1 is the capacity of the BESS at the previous time (t − 1) in Wh.
• η B,EF is the round-trip efficiency in p.u.
The efficiency (η B,INV ) of Equation (6) varies in the proposed model depending on
the power supplied. The efficiency of the conversion in the inverter has been taken from
the Battery Energy Storage inverter Ingecon Sun Storage datasheet 100TL [31]. Therefore,
Table 1 shows efficiencies for significant operating points with respect to the nominal power
of the inverter that have been taken.

Table 1. Main parameters of utility-scale inverters and BESS.

Inverter BESS BESS

(Ingecon 100TL) (TS-I HV 100)
Parameter Value Parameter Value
Euro-Efficiency 98.1% Max C-rate 1C
Efficiency (0.2 · PN ) 98.5% Energy (EB,MAX ) 96 kWh
Efficiency (0.4 · PN ) 98.5% Round-trip efficiency Up to 98%
Efficiency (0.6 · PN ) 98.5% Self-consumption 5W
Efficiency (0.8 · PN ) 98.2% DoD 100%
Efficiency (PN ) 98%

Self-discharge is measured as the charge lost over time and limits the energy available
in the battery. The internal chemical reactions that contribute to self-discharge are mainly
quantified when the BESS is not used for long periods of time. Due to the daily use of the
BESS, the effect of self-discharge can be considered insignificant since the manufacturer of
Energies 2024, 17, 1485 8 of 30

the BESS identifies 5 W of losses [32]. Therefore, self-discharge phenomena are ruled out in
the proposed model.
Regarding Equation (6), some operational limits related to the storage capacity of
the BESS are established, which are identified from Equation (7). Although the BESS
manufacturer indicates a 100% depth of discharge [32], for safety, a depth of discharge of
90% is chosen (DoD = 0.9).

 EB,MI N ∀ EB,t < EB,MAX ·(1 − DoD )
EB,h = EB,t ∀ EB,MAX ·(1 − DoD ) < EB,t < EB,MAX (7)
EB,MAX ∀ EB,t > EB,MAX


where
• EB,h is the energy stored in the BESS for each hour.
• EB represents the energy exchanged in 1 h with the BESS in Wh. Positive if it is
charging and negative if it is discharging.
• EB,MAX is the upper capacity limit of the BESS in Wh.
• EB,MIN is the lower capacity limit of the BESS in Wh.
• EB,t is the hourly charging or discharging energy of the BESS at the current time (t)
in Wh.

2.4. Model Phases

In the simulation process, a time step (∆t) of 1 h has been adopted. Consequently, all
equations described in this section are evaluated at hourly intervals.
PV forecasting can be categorized into physical models and data-driven models.
Physical models use numerical weather prediction, which shows good performance for
forecast horizons from several hours to six days [11]. To analyze long-term data, data-
driven models are often used. Although the proposed algorithm is not based on machine
learning, its approach is similar to data-driven models. It adopts a strategy that is based
Energies 2024, 17, x FOR PEER REVIEW 9 of 31
on the processing and analysis of historical data accumulated over an extended period.
This approach is characterized by unraveling meaningful patterns and relationships within
a dataset.
The algorithm
adjusted, in MATLAB
thus repeating the mainis simulation.
divided intoAs
two subsections
a result, to of
a matrix solve the problem,
indicators as
is gener-
shown in Figure
ated based 3.
on parameters, which is used to graphically represent the results obtained.

Figure 3. Model phases.

Figure 3. Model phases.

2.5. System Parameters and Indicators

The specific parameters of the proposed BESS are provided in Table 2 This table lists
the characteristics of the BESS, including the rated capacity, voltage, efficiency and ex-
pected lifetime according to data provided by the manufacturer [31].
One of the most important characteristics of any electrical power system is its rated
power, that is, how many kW it can produce on a continuous, full-power basis. If the
Energies 2024, 17, 1485 9 of 30

Left column blocks: As a preliminary step to the main simulation, a complete simu-
lation of the time series is carried out using irradiance and hourly temperature as input
variables. The hourly production of photovoltaic (PV) energy is determined for each hour
of the simulation in order to calculate the average production of PV energy for each month
of the year.
Center column blocks: The main simulation begins, in which, for each hour, the
calculation of photovoltaic energy production is carried out based on the meteorological
data provided by PVGIS. A constant power setpoint is established proportional to the
average PV production of each month, obtained in the preliminary simulation. The SoC
is calculated in the BESS according to its mathematical model. Charge and discharge
cycles are counted by applying the rainflow-counting algorithm. Finally, the degree of
degradation resulting from this cycling is estimated to adjust the capacity of the BESS.
This process is repeated iteratively until the simulation is completed. At the end of the
main simulation, the energy unavailability indicators are recorded. To identify significant
operating points, the key parameters of the power setpoint and the size of the BESS are
adjusted, thus repeating the main simulation. As a result, a matrix of indicators is generated
based on parameters, which is used to graphically represent the results obtained.

2.5. System Parameters and Indicators

The specific parameters of the proposed BESS are provided in Table 2 This table lists
the characteristics of the BESS, including the rated capacity, voltage, efficiency and expected
lifetime according to data provided by the manufacturer [31].

Table 2. Parameters of the proposed BESS [32].

Parameter Value
Nominal capacity 0–6 MWh
Rated Cycle-Life 1 6000 at 100% DoD and 1 C
Rated Cycle-Life 2 8000 at 100% DoD and 0.5 C
Operating voltage 845–1096 V
Cell Type Lithium
Coulomb efficiency 98%
BESS inverter efficiency 96%

One of the most important characteristics of any electrical power system is its rated
power, that is, how many kW it can produce on a continuous, full-power basis. If the
system has a generator, the rated power is determined by the rated power of the generator.
If the generator were to deliver its rated power for a full year, then the energy supplied
would be the product of the rated power times 8760 h/year. Since power systems do not
run at full power all year, they output less than the maximum amount. The capacity factor
(CF) is a convenient, dimensionless quantity between 0 and 1 that connects the rated power
to the energy delivered [33].
In Equation (8), the term CF represents the capacity factor, which is a dimensionless
quantity between 0 and 1. EGRID represents the energy supplied annually by the plant,
while PR is the rated power of the plant.

EGRID
CF = (8)
PR ·8760

By the year of commissioning, the global weighted average capacity factor for new
utility-scale solar PV reached 17.2% in 2021 [34].
The S2P parameter defines the storage capacity of the system according to Equation (9) [21].
CMAX represents the storage capacity of the BESS in MWh, and PP is the peak power of the
Energies 2024, 17, 1485 10 of 30

photovoltaic installation in MW. S2P is a parameter equivalent to the number of hours of

autonomy of a photovoltaic installation and is expressed in hours.

C MAX
S2P = (9)
PP

BESSs are designed to be versatile and easy to expand. The technical specifications for
large batteries [32] detail the capacity of each storage module. Specifically, the TS-IHV100E
storage module by Tesvolt Lutherstadt Wittenberg (Germany) has a capacity of 80 kWh
and allows four modules to be connected in parallel in a cabinet (total of 320 kWh).
Assuming that an S2P of 0.8 is to be selected and that the peak power of the PV plant
is 1 MWp, the storage capacity of the BESS must be 0.8 MWh. This would require two full
cabinets and an additional cabinet with two modules.
In Equation (10), the concept of the Constant Power Operating Factor is introduced [21].
PGRID is the constant power that must be supplied to the electrical grid. This equation relates
the power of the photovoltaic installation and the energy that must be provided to the
grid. Therefore, the CPOF parameter determines the equivalent operational performance of
the system.
P
CPOF = GRID (10)
PP
To ensure a constant and reliable energy supply, the system operation factor (CPOF )
must be lower than that associated with the average annual production of photovoltaic
energy. If this requirement is not met, the photovoltaic installation will incur an annual
energy deficit that the PV-BESS system cannot compensate. The CPOF must be between 0
and 0.172 as mentioned for the average capacity factor previously. In Equation (11), the
Annual Energy Deviation (AED) is defined as a relative accumulator of unavailable energy
throughout the entire simulation period, where the numerator represents the unavailable
energy, and the denominator represents the energy supplied throughout the entire simula-
tion period [26]. EPV represents the PV-BESS production for each hour, EGRID represents the
energy supplied each hour and h represents the number of hourly periods of the simulation.
To obtain the average annual unavailability, the Annual Energy Deficit (AED) indicator will
be used, which will take the average data of all simulation years.
In [35], in order to determine the firm capacity of a distributed generation, a specific
method is used that aims to obtain the minimum expected return on an investment, taking
into account a pre-established confidence level. In accordance with the criteria established
in the said study, a confidence level of 95% has been determined. Therefore, this study
aims to guarantee that 95% confidence level. This means that the AED indicator should not
exceed 5%.
According to data provided by the U.S. Energy Information Administration (EIA), in
the period between 2013 and 2022, the capacity factor of nuclear power plants in the United
States ranged between 90.8% and 93.4% [6]. The guarantee of a 95% confidence level in
the proposed PV-BESS equates its availability to that of a plant with the capacity to supply
firm power in the electrical system.

∑hh=
=1
96360
(E PV − EGRID ) h=96360
AED = V ( ∑ h=1 (E PV − EGRID ) < 0) (11)
∑hh=
=1
96360
EGRID

The fundamental parameters of the BESS are evaluated primarily by Equations (12)–(14),
presented in [36]. It is common to use the SoC and SoH to estimate the energy performance,
while the C-rate is used to approximate power performance. The SoC represents the current
capacity in the BESS and is defined by Equation (12), where QC is the content of the electric
charge at the present state, and QMAX is the maximum electric charge storage capacity at
the present state.
Energies 2024, 17, 1485 11 of 30

QC
SoC = (12)
Q MAX
The SoH represents the state of health of the BESS and is defined by Equation (13),
where QS is the maximum electric charge storage capacity in the specification, which
indicates the fully charged battery capacity at the initial stage without degradation.

Q MAX
SoH = (13)
QS

C-rate is a parameter to describe the charging and discharge speed, which is calculated
in Equation (14) where I is the current.

I
C − rate = (14)
QS

2.6. BESS Degradation Modeling

2.6.1. Equivalent Cycle Estimation
As shown in Figure 3, it is necessary to incorporate a degradation model for the BESS
based on the accounting of charge and discharge cycles. The methodology that is usually
presented to predict the useful life of batteries is based on situations where an identical
charge–discharge cycle is repeated throughout the useful life of the battery. However, in
most applications, the cycles change dynamically over time. In applications such as energy
storage with renewable generation, daily variations in surplus generation lead to variability
in the cyclic depth of discharge [37].
To identify the operational behavior of the BESS, it is necessary to use a consistent
method that adequately accounts for charge and discharge cycles.
The rainflow-counting algorithm has been widely used in research to analyze the
fatigue and degradation of materials [38]. This method identifies the loading cycles of
a material subjected to fluctuating loads during its useful life. The algorithm allows for
evaluating the durability of a component by estimating the stress levels necessary to
produce fatigue failure.
Battery degradation is a complex phenomenon composed of several chemical pro-
cesses. However, an analogy can be established with the fatigue phenomenon in materials.
In both cases, part of the degradation process occurs as a result of the successive application
of loads.
The method was first proposed by Matsuisi and Endo in 1968 and follows the ASTM [39].
The algorithm counts the equivalent charge–discharge cycles from the time series
corresponding to the state of charge (SoC) of the BESS [40]. To facilitate the understanding
of the process, a theoretical explanation of the method has been combined with a fictitious
example based on Figure 4.
The procedure is listed below:
1. Reduce the SoC curve to a sequence of peaks and valleys by identifying inversions in
the slope of the time series. Figure 4a identifies the sequence of points from 0 to 9;
2. Set the initial state of the SoC as the starting point for counting. In Figure 4a, this is
marked as point 0;
3. Identify two consecutive inflection points and calculate the range of the three segments
affected by that inflection. Each segment Ri is identified by a line delimited by two
points. The absolute value of the difference between the extreme values is the range
of the segment. In Figure 4a, the inflection points are 1 and 2. The range of segment 1
is R1 = |P1 − P0| = 30. The range of segment 2 is R2 = |P2 − P1| = 50, and the range
of segment 3 is R3 = |P3 – P2| = 30;
4. A complete cycle is identified if the range Ri is less than the range before and after.
Ri < min(Ri – 1, Ri + 1). This condition is met in Figure 4c, where the range of segment
S4 is smaller than in S3 and S5;
Energies 2024, 17, x FOR PEER REVIEW 12 of 31

Energies 2024, 17, 1485 4. A complete cycle is identified if the range Ri is less than the range before and12after. of 30
Ri < min(Ri – 1, Ri + 1). This condition is met in Figure 4c, where the range of seg-
ment S4 is smaller than in S3 and S5;
5.
5. IfIf the
the conditions
conditions of of stepstep 44 are
are met,
met, eliminate
eliminate thethe points
points associated
associated with
with the
the range
range RiRi
and repeat the process from step 3. In Figure 4d, points 3 and
and repeat the process from step 3. In Figure 4d, points 3 and 4 are eliminated; 4 are eliminated;
6.
6. IfIfaacomplete
complete cycle
cycle is isnot
notidentified,
identified, proceed
proceed toto the
the next
next points
points and
and repeat
repeat the
the process
process
from
from stepstep 3.
3. In
In Figure
Figure 4a, 4a, segment
segment S2 S2 does
does not meet the identification of a complete
cycle
cycle from
from step
step 4.4. ItIt must
must bebe analyzed
analyzed in in the
the next
next segment
segment (S3)
(S3) in
in Figure
Figure 4b;
4b;
7. The
The cycle
cycle detection
detection process
process isis repeated
repeated until the remaining profile is completed;
8. Once
Once all all the
the complete
complete cyclescycles have
have been
been identified,
identified, the
the remaining
remaining segments
segments areare con-
con-
sidered
sidered as as half
half cycles.
cycles. Segment
SegmentS2 S2 in
in Figure
Figure 4a4a and
and segment
segment S3
S3 in
in Figure 4b are half
cycles
cycles of of the
the series.
series.

(a) (b)

(c) (d)
Figure 4. Example of rainflow-counting algorithm. (a) Half-cycle counting in segment S2; (b)
Figure 4. Example of rainflow-counting algorithm. (a) Half-cycle counting in segment S2; (b) half-
half-cycle count in segment S3; (c) counting a complete cycle in segment S4; (d) half-cycle count
cycleredoing
after count inpoints
segment S3;new
in the (c) counting a complete cycle in segment S4; (d) half-cycle count after
segment S5.
redoing points in the new segment S5.
Each semicycle or cycle is classified according to its magnitude. The parameters that
Each semicycle or cycle is classified according to its magnitude. The parameters that
identify each magnitude are the average value of the path and the discharge depth. These
identify each magnitude are the average value of the path and the discharge depth. These
parameters are essential to analyze the useful life of the batteries.
parameters are essential to analyze the useful life of the batteries.
Table 3 shows all the half cycles or full cycles associated with the count in Figure 4.
Table 3 shows all the half cycles or full cycles associated with the count in Figure 4.
The count column identifies a half cycle (0.5) or complete cycle (1). The range column
The count column identifies a half cycle (0.5) or complete cycle (1). The range column
represents the discharge
represents the depth of
discharge depth of the
the BESS.
BESS. The
The mean
mean column
column represents the average
represents the average SoC
SoC
of each half cycle or cycle. The start and end columns identify each segment in the timeline.
 of each half cycle or cycle. The start and end columns identify each segment in the time-
line.
Energies 2024, 17, 1485 13 of 30
Table 3. Identification of sections in cycles and semicycles using the rainflow method for the pro-
posed example.
Table 3. Identification of sections in cycles and semicycles using the rainflow method for the
Count
proposed example. Range (%) Mean (%) Start (h) End (h)
0.5 30 45 0 1
Count
1 Range
10(%) Mean
35(%) Start3(h) End4(h)
0.5
0.5 30
50 45
35 01 12
10.5 10
70 35
45 32 45
0.5 50 35 1 2
0.5
1 20
70
40
45 2
7 5
8
10.5 70
20 45
40 75 86
0.5
0.5 80
70 50
45 56 69
0.5 80 50 6 9
In summary, this algorithm computed in MATLAB that applies the rainflow method
to the proposed fictitious
In summary, case in
this algorithm Figure 4 in
computed has identified
MATLAB two
that cyclesthe
applies and five half
rainflow cycles.
method
toThe
thedischarge
proposeddepth and case
fictitious the average
in Figurevalue of identified
4 has each section
twoallow
cyclesforand
a BESSfive degradation
half cycles.
analysis
The to be depth
discharge carriedand outthe
from a cycling
average point
value of view.
of each section allow for a BESS degradation
Figure
analysis to be5carried
shows outthefrom
number of equivalent
a cycling cycles of the BESS as a function of the
point of view.
depth of discharge
Figure 5 shows(DoD) and theofaverage
the number valuecycles
equivalent of each
of cycle or half
the BESS as cycle, applying
a function the
of the
rainflow-counting algorithm. The number of equivalent cycles is represented
depth of discharge (DoD) and the average value of each cycle or half cycle, applying the by a color
scale, where yellow
rainflow-counting corresponds
algorithm. Thetonumber
a higher ofnumber of equivalent
equivalent cycles. Eachby
cycles is represented cell in the
a color
figurewhere
scale, represents
yellow the number of equivalent
corresponds to a highercycles
number adjusted according
of equivalent to itsEach
cycles. amplitude and
cell in the
figure represents
average the number
value, since of equivalent
each component cycles adjusted
contributes according
a different weighttoto itsthe
amplitude and
BESS degra-
average
dation. value, since each component contributes a different weight to the BESS degradation.

Figure 5. The number of equivalent cycles according to the rainflow-counting method as a function
Figure 5. The number of equivalent cycles according to the rainflow-counting method as a function
of the depth of discharge (DoD) and average value (average) of the BESS.
of the depth of discharge (DoD) and average value (average) of the BESS.
In the simulations carried out in this study, the rainflow-counting algorithm is used
In the simulations carried out in this study, the rainflow-counting algorithm is used to
quantify the the
to quantify charge
charge andand discharge
discharge cycles
cycles in battery.
in the the battery.
In the Results section, cycling maps will be presented, similar to the example illustrated
in Figure 5, which classify and provide a concise visualization of the distribution of charge
and discharge cycles throughout the entire simulation.
Energies 2024, 17, 1485 14 of 30

Previous experimental research has shown that deep cycling causes the most signif-
icant degradation, whereas smaller cycles are less significant [37]. Thus, to achieve an
accurate quantification of BESS degradation, it is necessary to determine the number of
cycles for each level of discharge depth. Each segment represented in Figure 5 generates a
different level of degradation that will be quantified below.

2.6.2. Degradation Model of BESS

Batteries are electrochemical devices used to store and supply electrical energy. Due
to the nature of their chemical composition, they are subject to a degradation process. The
design, the electrochemical system, the temperature, the type of discharge and the storage
period are factors that affect the useful life of batteries [26].
There are different approaches to degradation modeling to quantify aging. Empirical
models are based on data obtained through tests, semi-empirical models are based on data
from aging studies and physicochemical models model aging mechanisms exclusively with
differential equations. The simulations proposed to analyze the degradation in this paper
are based on an empirical model supported by the tests carried out in [41,42].
Battery degradation is attributed to two independent components, known as cycling
aging and calendar aging. In Equation (15), the term DTOT represents the total degradation,
while DCAL represents the degradation attributable to calendar aging, and DCYC represents
the degradation related to the battery charging and discharging processes.

DTOT = DCAL + DCYC (15)

The factors that can influence each component are listed below.

Cycling Aging
To describe the degradation resulting from charge and discharge cycles, the term
cycling aging is used. The main factors that contribute to increased degradation due to
cycling aging are listed below.
Temperature (T): High temperatures cause cathode degradation and volume expan-
sion, which accelerates the growth of the solid electrolyte interface (SEI) on the anode. Both
causes are associated with a decrease in capacity and an increase in internal resistance [43].
The Arrhenius equation states that the rate of a chemical reaction increases exponentially
with temperature [41]. Using Equation (16), the coefficient that relates temperature to
degradation (KCYC-T ) is modeled. KT is a parameter determined by empirical methods [41],
TR is the reference temperature of 25 ◦ C and T is the ambient temperature.
TR
KCYC−T ( T ) = ek T ·(T −TR )· T (16)

Average state of charge (SoC): The (SoC) of a charge–discharge cycle has a slight
influence compared to other factors such as the depth of discharge (DoD). In [44], results are
presented that indicate that carrying out cycling at high SoC ranges results in a significant
loss of active lithium. This is due to the considerable expansion experienced on the surface
of the graphite particles at a high SoC during the transition between the lithiation stages.
As a consequence, the SEI layer on the surface breaks down and rebuilds itself irreversibly,
leading to the consumption of active lithium. The relationship between this degradation
coefficient (KCYC-SOC ) and the SoC is modeled by Equation (17). Kσ is a parameter that
characterizes the sensitivity of the degradation determined in this equation.

KCYC−SOC (σ ) = ekσ ·(σ−σR ) (17)

Depth of discharge (DoD): It represents the difference between the SoC levels reached
during a charge or discharge half cycle. It has been observed that as the degree of discharge
(DoD) increases, more degradation occurs. High discharge values imply a greater risk of
cracking and the formation of a new solid electrolyte interface (SEI) layer due to volume
Energies 2024, 17, 1485 15 of 30

expansion in the graphite anode, especially when crossing the phase change regions in the
anode [44]. Empirical methods, such as [35] and [41], propose degradation models as a
function of the DoD using exponential functions. In Equation (18), KCYC-DOD represents the
degradation coefficient as a function of the DoD, where KBAT is the degradation constant
provided by the manufacturer, δ is the DoD and Kδ1 is a parameter identified empirically
in [45].
KCYC− DOD (δ) = (K BAT ·δ)Kδ1 (18)
The KBAT degradation constant can be obtained from the data provided by the manu-
facturer. In Equation (19), DEoL represents the degradation per unit expected at the end of
the useful life of the batteries (approximately 0.8), while NCYC corresponds to the number
of cycles necessary to reach this degradation.

DEoL
K BAT = (19)
( N CYC ·(1 − DEoL ))Kδ 1

Number of equivalent cycles (NEC): Charging and discharging a battery does not
always start with a 100% state of charge (SoC) and end at 0% SoC, which would represent
a full cycle. The NEC for a given DoD is defined as the number of cycles equivalent
to a scenario where a full charge or discharge is performed in each cycle. The most
pronounced aging mechanism is attributed to battery cycling, which involves graphite
exfoliation, structural decomposition, transition metal dissolution and the growth of the
solid electrolyte interface (SEI) layer. As a result of the factors explained above, the cycling
aging component is obtained using Equation (20).

KCYC = KCYC−T ( T )·KCYC−SOC (σ )·KCYC− DOD (δ) (20)

The resulting degradation does not follow a linear relationship. In order to fit the
behavior of real cycling aging to the model proposed in this article, Equation (21) is used,
which applies a non-linear relationship. K1 and K2 are the constants identified in [41] to fit
the model. The rainflow-counting algorithm method used is essential for quantifying the
degradation due to cycling, since it provides information on the average value of the SoC,
the DoD and the NEC.
−K2 ·KCYC · NEC
DCYC = 1 − K1 − (1 − K1 )−KCYC · NEC (21)

Calendar Aging
The aging effects due to storage or the period of inactivity of the battery are called
calendar aging. These effects are related to the loss of capacity due to the passage of time
and are not independent of the charge and discharge cycles [42]. The estimation of the
degradation caused by the chemical decomposition of the electrolyte solution is clearly
identified. Below are the factors that intensify the degradation related to calendar aging.
Temperature (T): Elevated temperature leads to the faster growth of the solid elec-
trolyte interface (SEI) layer. This phenomenon accelerates cell degradation [46]. The KCYC-T
factor of Equation (16) is valid to model this behavior.
Time (t): Reactions due to SEI growth and binder decomposition increase over time.
Even in the absence of charge and discharge cycles, the degradation mentioned in [46]
occurs. The Kt factor of Equation (22) models this behavior by taking the constant KC1
through empirical tests in [41].
Kt = KC1 ·t (22)
State of Charge (SOC): If the battery charge level is high, there is a low potential at the
anode and a high potential at the cathode. As mentioned in [46], this low potential at the
anode accelerates the growth of the SEI, resulting in the acceleration of the degradation.
Furthermore, the oxidation of lattice oxygen is an electrochemical process that occurs
mainly at high state of charge (SOC) ranges. This process results in a greater dissolution
Energies 2024, 17, 1485 16 of 30

of the transition metal due to the oxidation of the lattice oxygen of the cathode [47]. The
KCYC-SOC factor of Equation (3) is valid for modeling this behavior. As a result of the factors
explained above, the calendar aging component is obtained using Equation (23).

KCAL = KCYC−T ( T )·KCYC−SOC (σ )·Kt (23)

As in cycling aging, for calendar aging, the degradation does not follow a linear
relationship. To adjust the behavior, Equation (24) is used, which applies a non-linear
relationship.
DCAL = (1 − K1 )−K2 ·KCAL − (1 − K1 )−KCAL (24)
Table 4 presents the parameters used to identify the proposed degradation model,
which are obtained from experimental results taken from [41,45,48].

Table 4. Degradation equation parameters.

Name Value Equation

KT 0.0693 2
TR 25 ◦ C 2
Kσ 1.039 3
σR 0.45 3
Kδ1 2.03 4 and 5
K1 0.0575 7
K2 121 7
KC1 4.14 × 10−10 8

2.6.3. Issues with Lithium-Ion Batteries

The main limitations of lithium-ion batteries are the high price, lack of lithium supply
and safety issues related to electrolytes.
The review article [49] presents the main issues discussed in the literature on lithium-
ion batteries. They are summarized below:
• Increasing temperature has a significant impact on the accuracy of methods used to
estimate the state of health, as well as increasing degradation in cells;
• It is essential that proper heat dissipation methods are implemented in batteries.
Failure to do so can result in reduced battery life, poor cycling performance and an
increased risk of explosion;
• Chemical separation processes for recycling face significant challenges on a large scale.
Ethical and environmental concerns arise in the extraction of lithium, cobalt and other
mineral resources, as well as issues related to their safe disposal;
• Despite the efforts made in the research of cathodes, anodes and electrolytes, signif-
icant challenges remain. These include cathode dissolution, the negative effect of
electrostatic interaction, dendrite formation, corrosion and passivation. These difficul-
ties can lead to significant capacity degradation, low efficiency and short circuits;
• Interface stability is a challenge because excessive overcharging or discharging, to-
gether with short circuits, causes cells to charge or discharge unevenly.

3. Results
The simulations carried out with constant power setpoints up to Section 3.4 have
been configured to maintain an intermediate value (CPOF = 0.1). In [21], simulations were
carried out using various power setpoints, and it was identified that a CPOF of 0.1 achieves
an appropriate balance between the amount of energy supplied to the grid and viable
energy unavailability, without the need to significantly oversize the BESS. If all the energy
produced by the PV plant could be supplied, the capacity factor would be equivalent to the
CPOF power setpoint. In the Introduction, statistical data have been provided indicating
that the average capacity factor in PV plants is 0.172. Therefore, it would not be feasible to
Energies 2024, 17, 1485 17 of 30

supply power setpoints that exceed this value, and a setpoint with a CPOF of 0.1 is a value
close to the intermediate value.
In Section 3.1, an initial approach is carried out to analyze the influence of each
parameter on the degradation in the BESS.
In Section 3.2, using the rainflow-counting algorithm, the model proposed in this
study is simulated in order to analyze how degradation affects the cycling distribution in
the BESS. The gradual degradation of the BESS has a significant impact on the fundamental
Energies 2024, 17, x FOR PEER REVIEW 18 of 31
degradation variables, which are the DoD and the SoC.
After analyzing the influence of each parameter on the degradation and distribution of
cycles in the simulation, starting in Section 3.3, results such as the predicted EoL and energy
depth of discharge.
unavailability The SoHThe
are presented. after 10,000 is
purpose cycles in the blue
to determine anreference
appropriate curve is 94.32%,size,
installation and
by increasing
taking the DoD
into account by 40%ininS2P
variations theand
orange
CPOcurve,
F . the SoH drops to 82.25%. These results
support the hypothesis that the DoD has a substantial impact on the degradation phe-
3.1. Influence
nomena of Degradation
mentioned according
in Section 2.6. to its Parameters
In the
The secondformulated
equations case, the yellow curve
in Section 2.6in Figurea6mathematical
provide with a DoD of 30% and to
framework 𝑆𝑜𝐶 of 85%
quantify
represents
the a simulation
degradation experiencedin which
by a BESS 𝑆𝑜𝐶 toiscycling
the due increasedandby 40% with
calendar respect to the blue
aging.
reference
In thiscurve. Theasystem
section, performs
simulation symmetrical
of the proposed modelchargesisand discharges
carried out toinanalyze
the range
theof
SoC values
influence ranging from
of calendar aging70%
and to 100%. The
cycling. the 𝑆𝑜𝐶
Onlyresults has been
obtained modified
confirm in comparison
the validity of the
model andsimulation
with the ensure thatofthethedegradation
reference blue is correctly
curve, thusestimated
allowing infor
subsequent of the 𝑆𝑜𝐶
simulations.
the influence
To achieve this, theoretical scenarios are simulated in which charges
to be analyzed. The SoH after 10,000 cycles in the blue reference curve is 94.32%, and by and discharges
are applied to
increasing the𝑆𝑜𝐶
the storage system
by 40% following
in the cyclic patterns.
yellow curve, the SoH Theonlygraphical
drops to results
92.16%.are shown
in Figures
The 6conclusion
and 7, where the state
drawn is thatofthe
health due toofcyclic
influence agingisismore
the DoD identified. On the
significant other
than that
hand, 𝑆𝑜𝐶 but
of theFigure 7 shows
both the
muststate of health
be taken intodue to calendar
account for the aging.
cycling degradation analysis.

Figure
Figure BESS
6. 6. cycling
BESS degradation
cycling curvecurve
degradation for different DoD and
for different SoC.and 𝑆𝑜𝐶 .
DoD
To estimate the degradation, it is common to propose a cycling aging model based
on the DoD and 𝑆𝑜𝐶 variables, using the rainflow-counting algorithm for cycle counting
[28,33].
To validate the proposed cycling aging model, it is compared with long-term aging
tests of lithium batteries. An industry reference dataset from the Oxford University Re-
search Archive [44,50] is used. This dataset contains a long-term battery aging test for
lithium-ion cells.
Figure 7 shows a graphical representation of the SoH or degradation as a function of
the number of battery cycles. The blue curve with a solid line corresponds to the pro-
posed model, while the curves with dashed lines represent the degradation of four types
Energies 2024,17,
Energies2024, 17,1485
x FOR PEER REVIEW 1930of 31
18 of

Figure 7. BESS cycling degradation curve of the proposed model with respect to the degradation of
Figure 7. BESS cycling degradation curve of the proposed model with respect to the degradation of
real batteries.
real batteries.

To validate
3.1.1. Cycling Aging theinfluence
experimental model, a comparison is made with the data provided
by the manufacturer
A simulation is carried of a out
commercial
in MATLAB battery.
using The Lithium model
the proposed Iron-Phosphate
in order toPowerBrick
analyze
battery datasheet [51] presents the number of cycles vs. depth
the influence of its main variables, the DoD and SoC. In this theoretical scenario, of discharge graph. Forthe the
curve with
variables DoD a DoD
and SoCof 90% and aconstant
remain C-rate ofthroughout
0.25 C, which the resembles
simulation,the average
which value
implies of the
that
proposed
periodic simulation,
charge 7200 useful
and discharge cycleslifewithcycles are identified.
identical magnitudes are performed. A constant
According
temperature of 25to ◦ Cthe manufacturer’s
is assumed, to makespecifications,
the temperature thevariable
service independent
life of the battery is con-
of aging.
sidered
The toobtained
results cover upidentify
to 80% theof the initial of
influence capacity. Therefore,
each variable on the ingradual
Figure 7, theofEnd
loss of Life
battery
capacity
according withtoeach cycle.
the manufacturer has been identified with an “x”.
Figure
In the 6proposed
shows a model,comparison of the SoH between
the intersection up to 10,000 cycles,
the limit for of
value three
80%cases
and with
the blue
different cycling.
“model” curve in Figure 7 identifies 6700 equivalent cycles, which is similar to the real
valueTheprovided
blue curveby in Figure 6 is taken as aThe
the manufacturer. reference,
data withare acompared
DoD of 30% withandaSoC of
Lithium
45%. In this system, symmetrical charges and discharges are carried
Iron-Phosphate battery, because the composition of this type of battery has been used in out in the range of
SoC values from
large-scale BESSs 30% suchto 60%. The simulation corresponding to the blue curve in Figure 6
as [32].
modifies each variable with an increase of 40% to analyze which of them has a greater
influence on the degradation.
3.1.2. Calendar Aging Influence
In the first case, the orange curve in Figure 6 represents a simulation in which the
DoD isThe equations
increased formulated
by 40% in Section
with respect to the 2.6
blueonreference
calendarcurve.
agingSymmetric
allow for the degradation
charges and
to be quantified, considering both the passage of time and
discharges are carried out in the range of SoC values ranging from 10% to 80%. Since the SoC value. Several simu-
lations
this curveare carried
shares theoutsamein order
SoC astothe validate the proposed
blue curve, it isolatesmodel, assuming
the influence thatSoC
of the there
and is no
cycling ofthe
facilitates theverification
battery and ofthe
the charge levelrelationship
significant is kept constant.
between the degradation and the
depth Figure 8, with The
of discharge. the legend
SoH after “sim”, shows
10,000 theincalendar
cycles the blueaging caused
reference by the
curve proposed
is 94.32%,
and by increasing the DoD by 40% in the orange curve, the SoH drops to 82.25%. These 8 is
model. Compare with a degradation dataset obtained in [52] whose legend in Figure
“test”.support the hypothesis that the DoD has a substantial impact on the degradation
results
Note that
phenomena the higher
mentioned the average
in Section 2.6. state of charge 𝑆𝑜𝐶 , the greater the degradation.
AfterIn 2.5
the years,
secondthe case, the yellow
calendar aging curve in Figure
simulation 6 with a aDoD
estimates SoHofof30% 88% and SoC the
when of 85%
SoC is
represents a simulation in which the SoC is increased
100%. However, the SoH is 91% when the SoC is 40%, as shown in Figure 8. by 40% with respect to the blue
reference curve. The system
To corroborate the agingperforms
calendar, symmetrical
degradation charges
data ofand discharges in
a lithium-ion the range
storage system
ofare
SoC values ranging
presented in [53].fromIt is 70% to 100%.
observed thatOnly the SoC has been
the degradation modifiedcoincides
experienced in comparison
with that
with the simulation
represented in Figureof the reference
8 after blue curve,2thus
approximately allowing
years. For anforSoC theofinfluence
40%, an SoH of theofSoC
92% is
toreported
be analyzed. The SoH after 10,000 cycles in the blue reference
in [53]. This figure coincides with the value represented in Figure 8, where curve is 94.32%, and byit is
increasing the SoC
also observed thatbyfor40%anin theof
SoC yellow
40%, curve,
there is theanSoH
SoHonly drops
of the to 92.16%.
order of 92%. For an SoC of
100%, the SoH in both cases is around 90%.
Energies 2024, 17, 1485 19 of 30

The conclusion drawn is that the influence of the DoD is more significant than that of
the SoC but both must be taken into account for the cycling degradation analysis.
To estimate the degradation, it is common to propose a cycling aging model based on
the DoD and SoC variables, using the rainflow-counting algorithm for cycle counting [28,33].
To validate the proposed cycling aging model, it is compared with long-term aging
tests of lithium batteries. An industry reference dataset from the Oxford University Re-
search Archive [44,50] is used. This dataset contains a long-term battery aging test for
lithium-ion cells.
Figure 7 shows a graphical representation of the SoH or degradation as a function of
the number of battery cycles. The blue curve with a solid line corresponds to the proposed
model, while the curves with dashed lines represent the degradation of four types of
batteries in the dataset [44,50].
It can be seen that the degradation levels in Figure 7 are comparatively similar, al-
though not identical. It is complex to identify an exact degradation pattern for each battery.
Based on this observation, it is considered that the proposed model has substantial validity
in predicting the degradation process until the End of Life or a SoH of 80%.
To validate the experimental model, a comparison is made with the data provided
by the manufacturer of a commercial battery. The Lithium Iron-Phosphate PowerBrick
battery datasheet [51] presents the number of cycles vs. depth of discharge graph. For the
curve with a DoD of 90% and a C-rate of 0.25 C, which resembles the average value of the
proposed simulation, 7200 useful life cycles are identified.
According to the manufacturer’s specifications, the service life of the battery is con-
sidered to cover up to 80% of the initial capacity. Therefore, in Figure 7, the End of Life
according to the manufacturer has been identified with an “x”.
In the proposed model, the intersection between the limit value of 80% and the blue
“model” curve in Figure 7 identifies 6700 equivalent cycles, which is similar to the real value
provided by the manufacturer. The data are compared with a Lithium Iron-Phosphate
battery, because the composition of this type of battery has been used in large-scale BESSs
such as [32].

3.1.2. Calendar Aging Influence

The equations formulated in Section 2.6 on calendar aging allow for the degradation to
be quantified, considering both the passage of time and the SoC value. Several simulations
are carried out in order to validate the proposed model, assuming that there is no cycling
of the battery and the charge level is kept constant.
Figure 8, with the legend “sim”, shows the calendar aging caused by the proposed
model. Compare with a degradation dataset obtained in [52] whose legend in Figure 8
is “test”.
Note that the higher the average state of charge SoC, the greater the degradation. After
2.5 years, the calendar aging simulation estimates a SoH of 88% when the SoC is 100%.
However, the SoH is 91% when the SoC is 40%, as shown in Figure 8.
To corroborate the aging calendar, degradation data of a lithium-ion storage system
are presented in [53]. It is observed that the degradation experienced coincides with that
represented in Figure 8 after approximately 2 years. For an SoC of 40%, an SoH of 92% is
reported in [53]. This figure coincides with the value represented in Figure 8, where it is
also observed that for an SoC of 40%, there is an SoH of the order of 92%. For an SoC of
100%, the SoH in both cases is around 90%.
Another relevant aspect is that the calendar aging is more pronounced during the first
years of the battery’s life. The simulations represented in Figure 8 coincide in showing a
higher level of degradation around the first year.
 results in a reduction in the storage capacity to 80%. In other words, the degradation is
estimated to be equivalent to 20%.
For the equations proposed in the model of this article, the projection of an SoC of
80% would reach an End of Life at 8.7 years and for an SoC of 60%, at 10.7 years.
Energies 2024, 17, 1485
Therefore, it can be concluded that the calendar life of the proposed model is con-
20 of 30
sistent and similar to the manufacturer’s specifications when average SoC values between
60 and 80% are considered.

Figure8.8. Calendar
Figure Calendar degradation
degradation curve
curve in
inthe
theproposed
proposedmodel
modelwith
withrespect
respecttotoreal
realbatteries
batteriesfor
for
different 𝑆𝑜𝐶 .
different SoC.

3.2. In
Theaddition,
Influence aofcomparison
Degradation on of the
thisCycling
modelDistribution in the BESS
with a commercial battery is carried out.
TakingThe the introduction
Lithium Iron-Phosphate
of degradationPowerBrick battery [51]
in the proposed as a reference,
model leads to athe manufacturer
significant altera-
establishes
tion of thein its specifications
BESS cycling process, a calendar life greater
since it causes than decrease
a gradual 10 years. in
Therefore, upon
its capacity. theof
One
expiration of this period,
the most significant factsthe
is End of Life is considered
the progressive degradationto have
thatbeen reached,
affects which
the deeper results
discharge
incycles,
a reduction in the them
preventing storage capacity
from to 80%.
reaching theirInmaximum
other words, theThis
level. degradation is estimated
implies that the BESS
tooperation
be equivalent to 20%.
strategies are modified due to a capacity limitation.
For
Two thesimulations
equations proposed in the model
of the proposed modelofarethiscarried
article,out
thewith
projection
the aimof of
ananalyzing
SoC of 80% its
would reach an End of Life at 8.7 years and for an SoC of 60%, at 10.7 years.
behavior until reaching the End of Life of the BESS, technically defined as 20% degrada-
tionTherefore,
in relationittocan
itsbe concluded
initial that the calendar life of the proposed model is consistent
capacity.
and similar to the manufacturer’s specifications when average SoC values between 60 and
80% areCycling
3.2.1. considered.
Analysis for Reduced BESS Size
A Influence
3.2. The relativelyofsmall storageonsize
Degradation the is chosen
Cycling to ensure in
Distribution that
thethe battery undergoes a full
BESS
charge or discharge cycle in each operating cycle since its small capacity brings it to a
The introduction of degradation in the proposed model leads to a significant alteration
state of full charge or discharge quickly. In this way, it is possible to systematically ob-
of the BESS cycling process, since it causes a gradual decrease in its capacity. One of the
serve how high discharge depths, which are around 90%, experience a substantial re-
most significant facts is the progressive degradation that affects the deeper discharge cycles,
duction.
preventing them from reaching their maximum level. This implies that the BESS operation
strategies are modified due to a capacity limitation.
Two simulations of the proposed model are carried out with the aim of analyzing its
behavior until reaching the End of Life of the BESS, technically defined as 20% degradation
in relation to its initial capacity.

3.2.1. Cycling Analysis for Reduced BESS Size

A relatively small storage size is chosen to ensure that the battery undergoes a full
charge or discharge cycle in each operating cycle since its small capacity brings it to a state
of full charge or discharge quickly. In this way, it is possible to systematically observe how
high discharge depths, which are around 90%, experience a substantial reduction.
Figure 9 shows the behavior of the BESS through the analysis based on the count
of load cycles (rainflow-counting algorithm). As observed in Section 2, the charge cycle
count is an essential metric in the characterization of the BESS. Figure 9 provides a visual
representation illustrating how each cycle occurs in the BESS.
 70%. This indicates that, over the life of the BESS, it is operating in a DoD range that be-
gins near its maximum capacity and gradually decreases over time as the battery de-
grades.
The observed situation has both positive and negative implications for the BESS. It is
Energies 2024, 17, 1485
beneficial that, by operating with a smaller DoD, the phenomena of cycling aging are
21 of 30
reduced. However, there is a progressive limitation on the capacity of the BESS. This re-
striction can lead to increases in energy unavailability in the system.

(a) (b)
Figure 9. Rainflow-counting results for CPOF = 0.1 and S2P = 1. (a) Without aging. (b) Aging until
Figure 9. Rainflow-counting results for CPOF = 0.1 and S2P = 1. (a) Without aging. (b) Aging
EoL.
until EoL.
The degradation of the BESS constitutes a significant restriction in terms of its stor-
Figure 9a shows a distribution of the charging cycles in which it is observed that the
age capacity, as observed in Figure 9, and, consequently, its effects are interesting to an-
highest concentration of these cycles is around a DoD close to 90%. Additionally, the SoC is
alyze in relation to the parameters of unavailability of reliable energy throughout the
estimated at approximately 55%. Consequently, during each operating cycle, the system
time.
experiences a variation in the SoC that ranges between 10% and 100% of its total capacity.
Specifically, the BESS has been subjected to 3665 operating cycles in this operating range.
The rest of the cycles, compared to the magnitude mentioned, are numerically smaller and
can be considered insignificant. This situation arises because the BESS has a relatively
limited nominal capacity. Consequently, during any energy transfer process, whether
charging or discharging, the BESS tends to reach its maximum capacity or be completely
depleted. Figure 9b incorporates the aging until its End of Life, which occurs at 44,352 h.
Only during the first operating cycles of the BESS is it possible to reach maximum depths
of discharge of 90%. Due to the manifestation of degradation phenomena inherent to the
use, the ability to reach such a DoD decreases. This development reflects the importance of
considering aging phenomena in a BESS.
In Figure 9b, the aging phenomena are taken into account until its End of Life, which
occurs at 44,352 h. It is only during the first operating cycles of the BESS that it is possible to
reach maximum depths of discharge of 90%. Due to the manifestation of degradation phe-
nomena inherent to the use, the capacity to reach such a DoD decreases. This development
reflects the importance of considering aging phenomena in a BESS.
In Figure 9b, it can be seen that the dominant distribution of operating cycles has
changed in relation to Figure 9a. Most of these cycles are now concentrated around 80%
DoD and 50% SoC. It is noted that most DoDs start around 90% and decrease to around
70%. This indicates that, over the life of the BESS, it is operating in a DoD range that begins
near its maximum capacity and gradually decreases over time as the battery degrades.
The observed situation has both positive and negative implications for the BESS. It
is beneficial that, by operating with a smaller DoD, the phenomena of cycling aging are
reduced. However, there is a progressive limitation on the capacity of the BESS. This
restriction can lead to increases in energy unavailability in the system.
The degradation of the BESS constitutes a significant restriction in terms of its storage
capacity, as observed in Figure 9, and, consequently, its effects are interesting to analyze in
relation to the parameters of unavailability of reliable energy throughout the time.
To obtain a perspective of the evolution of the energy levels in the BESS in this
simulation, Figure 10a is presented. The 3D graph illustrates the annual evolution of the
resulting SoC during the operation of the BESS. It shows the day of the year on the “x” axis
(day of the year), the sample number of each day on the “y” axis, which are the hours of
the day, and the SoC on the “z” axis. In a small capacity battery, the SoC is fully charged
and discharged most of the time. For days without PV production, the SoC remains fully
discharged, as seen on day 8 of Figure 10b.
 To obtain a perspective of the evolution of the energy levels in the BESS in this sim-
ulation, Figure 10a is presented. The 3D graph illustrates the annual evolution of the re-
sulting SoC during the operation of the BESS. It shows the day of the year on the “x” axis
(day of the year), the sample number of each day on the “y” axis, which are the hours of
Energies 2024, 17, 1485
the day, and the SoC on the “z” axis. In a small capacity battery, the SoC is fully charged
22 of 30
and discharged most of the time. For days without PV production, the SoC remains fully
discharged, as seen on day 8 of Figure 10b.

(a) (b)
Figure 10. The evolution of the SoC in the first year of the simulation for S2P = 1. (a) Full year; (b)
Figure 10. The evolution of the SoC in the first year of the simulation for S2P = 1. (a) Full year;
zoom in the first 12 days.
(b) zoom in the first 12 days.
3.2.2.Cycling
3.2.2. Cycling Analysis
Analysis for
for Oversize
Oversize BESS
BESS Size
Size
Energies 2024, 17, x FOR PEER REVIEW Figure 11 shows the distributions of the load cycles associated with an oversized 23 of 31
Figure 11 shows the distributions of the load cycles associated with an oversized BESS
BESS (S2P = 5). This situation contrasts markedly with
(S2P = 5). This situation contrasts markedly with Figure 9. Figure 9.
In a hypothetical context where BESS degradation was not a factor to consider, it is
observed in Figure 11a that the majority of the cycles are concentrated around a relatively
low DoD, approximately 30%. Additionally, the SoC remains at a high level, around 85%.
Consequently, in each operating cycle, the system goes through fluctuations that range
between 70% and 100% of its total capacity. These data suggest that, under these theo-
retical conditions, the BESS operates mostly in a higher range of its capacity, avoiding
deep discharges and maintaining a high charge level. The rest of the cycles, compared to
the magnitude mentioned, are numerically smaller and can be considered insignificant.
Figure 11b shows the aging phenomena until their End of Life, which occurs at
69,344 h. The dominant distribution of operating cycles has changed in relation to Figure
11a. Most of these cycles are concentrated for a DoD of 30% and a SoC ranging between
60% and 80%.
(a) (b)
Although degradation phenomena reduce the storage capacity, the degradation
Figure to
tends 11.be
Rainflow-counting the 𝑆𝑜𝐶
mitigated sinceresults for CPO F = 0.1 and
is reduced. S2Psuggests
This = 5. (a) Without aging.
that over the(b) Aginglife
useful until
of
Figure 11. Rainflow-counting results for CPOF = 0.1 and S2P = 5. (a) Without aging. (b) Aging
the BESS, it is operating at a DoD of 30% and 𝑆𝑜𝐶 around 70%.
EoL.
until EoL.
Figure 12 shows the evolution of the energy levels in the BESS for this specific case.
In a hypothetical context where BESS degradation was not a factor to consider, it is
Most of the time, the SoC is at levels close to 100%, and only occasionally after several
observed in Figure 11a that the majority of the cycles are concentrated around a relatively
days of low production does the SoC reach its minimum value, as can be seen in the
low DoD, approximately 30%. Additionally, the SoC remains at a high level, around
zoom to Figure 12b.
85%. Consequently, in each operating cycle, the system goes through fluctuations that
range between 70% and 100% of its total capacity. These data suggest that, under these
theoretical conditions, the BESS operates mostly in a higher range of its capacity, avoiding
deep discharges and maintaining a high charge level. The rest of the cycles, compared to
the magnitude mentioned, are numerically smaller and can be considered insignificant.
Figure 11b shows the aging phenomena until their End of Life, which occurs at 69,344 h.
The dominant distribution of operating cycles has changed in relation to Figure 11a. Most
of these cycles are concentrated for a DoD of 30% and a SoC ranging between 60% and 80%.
Energies 2024, 17, 1485 23 of 30
(a) (b)
Figure 11. Rainflow-counting results for CPOF = 0.1 and S2P = 5. (a) Without aging. (b) Aging until
EoL. Although degradation phenomena reduce the storage capacity, the degradation tends
to be mitigated since the SoC is reduced. This suggests that over the useful life of the BESS,
Figure 12at
it is operating shows
a DoD the
ofevolution
30% and SoCof the energy
around levels in the BESS for this specific case.
70%.
MostFigure
of the 12
time,
showsthe the
SoCevolution
is at levels close
of the to 100%,
energy and
levels inonly occasionally
the BESS after several
for this specific case.
days of
Most of the
lowtime,
production
the SoC isdoes the SoC
at levels closereach its minimum
to 100%, value, as can
and only occasionally beseveral
after seen indays
the
zoom
of lowto Figure 12b.does the SoC reach its minimum value, as can be seen in the zoom to
production
Figure 12b.

(a) (b)
Figure 12. The evolution of the SoC in the first year of the simulation for S2P = 5. (a) Full year; (b)
Figure 12. The evolution of the SoC in the first year of the simulation for S2P = 5. (a) Full year;
zoom in the first 12 days.
(b) zoom in the first 12 days.
3.3. The
3.3. The Influence of the
Influence of the Size
Size of
of the
the BESS
BESS on on its
its End
End of
of Life
Life
In the
In the previous
previous section,
section, it
it was
was observed
observed thatthat the
the choice
choice ofof BESS
BESS sizesize has
has aa significant
significant
effect on degradation. In order to determine the life of the BESS,
effect on degradation. In order to determine the life of the BESS, the number the number of days until
of days
the BESS reaches the End of Life for each value of the S2P indicator is
until the BESS reaches the End of Life for each value of the S2P indicator is presented inpresented in Figure
13a. This
Figure 13a.graphical representation
This graphical representation provides
providesinformation
information onon howhowdifferent
differentBESS
BESS sizes
sizes
affect system longevity.
affect longevity.
In Figure 13a, it can be seen that when selecting an S2P of up to 1.8, the degradation
behavior is uniform, reaching the EoL approximately
approximately after 1875 1875 days
days ofof operation.
operation. This
specific feature
specific featureisisattributed
attributedtotothe thelimited
limitedstorage
storagecapacity
capacity of of
thethe BESS.
BESS. TheThe storage
storage sys-
system
tem remains
remains inoperative
inoperative at its capacity
at its capacity extremes,
extremes, either
either at at maximum
maximum capacitycapacity
due to aduehightoPVa
high PV generation
generation or at its minimum
or at its minimum capacity capacity
when there when there
is no PVisproduction,
no PV production,
and the andBESSthe
is
immediately discharged. It is important to highlight that these operating conditions are
not valid in this study from the point of view of the energy unavailability of the system
because the AED indicator exceeds 5%.
By increasing the storage capacity, and in particular when an S2P value of 1.8 is
exceeded, battery charging and discharging processes operate with a lower depth of
discharge (DoD). This implies that, in each cycle, the BESS does not use its full capacity,
which has a direct effect on reducing the degradation rate of the cells. Consequently, the
durability of the system benefits from the point of view of the degradation due to cycling.
This relationship is clearly manifested in Figure 13a; as the S2P increases beyond 1.8,
the time to reach the EoL extends linearly.
 When a larger storage is selected, a decrease is observed in the total number of cycles
that the BESS performs until it reaches its EoL. As seen in Figure 13b, for a system with a
more restricted capacity, specifically with an S2P of 1.8, the rainflow method, used to
count charge and discharge cycles in storage systems, indicates that the BESS can support
up to 1400 cycles equivalents up to their EoL. For a system with a larger storage capacity,
Energies 2024, 17, 1485 24 of 30
with an S2P of 4.5, it presents an equivalent cycle reduction, supporting only 910 cycles
until its EoL.

(a) (b)
Figure 13. (a) Days to EoL as a function of S2P; (b) cycles to EoL as a function of S2P.
Figure 13. (a) Days to EoL as a function of S2P; (b) cycles to EoL as a function of S2P.
As mentioned in Section 2, the degradation of the BESS is due to calendar and cy-
Starting from an S2P of 4.5, an inflection point is observed in Figure 13a. Beyond
cling effects. In order to determine the specific influence of each of these effects in the
this value, the increase in time before reaching the EoL tends to plateau, regardless of
analysis proposed in this section, Figure 14 shows the degradation curves corresponding
how much S2P continues to increase. Having an excessively high storage capacity, the
to each of them as a function of the S2P parameter.
BESS is frequently subjected to slight charge and discharge cycles that produce low cycling
Figure 14 shows that calendar degradation increases with S2P. As the BESS capacity
degradation. Instead, the system mostly remains in a fully loaded state. This scenario
with such aitlarge
increases, is observed
S2P sizethat the SoC
is also level remains
not realistic, since high for longer
it implies periods of
the excessive time. Thisof
oversizing
behavior leads to further degradation per calendar. This correlation is corroborated with
the BESS.
the calendar degradation
When a larger storagestudy presented
is selected, in Figure
a decrease 8. In thisin
is observed study, it is number
the total observedofthat a
cycles
higher state of charge (SoC = 100%) generates considerably greater degradation com-
that the BESS performs until it reaches its EoL. As seen in Figure 13b, for a system with a
paredrestricted
more to a lower state of specifically
capacity, charge (SoCwith
= 40%). Theofrelationship
an S2P between
1.8, the rainflow the BESS
method, usedcapacity,
to count
charge and discharge cycles in storage systems, indicates that the BESS can support up to
1400 cycles equivalents up to their EoL. For a system with a larger storage capacity, with
an S2P of 4.5, it presents an equivalent cycle reduction, supporting only 910 cycles until
its EoL.
As mentioned in Section 2, the degradation of the BESS is due to calendar and cycling
effects. In order to determine the specific influence of each of these effects in the analysis
proposed in this section, Figure 14 shows the degradation curves corresponding to each of
them as a function of the S2P parameter.
Figure 14 shows that calendar degradation increases with S2P. As the BESS capacity
increases, it is observed that the SoC level remains high for longer periods of time. This
behavior leads to further degradation per calendar. This correlation is corroborated with
the calendar degradation study presented in Figure 8. In this study, it is observed that a
higher state of charge (SoC = 100%) generates considerably greater degradation compared
to a lower state of charge (SoC = 40%). The relationship between the BESS capacity, the
duration of high SoC levels and the consequent degradation due to scheduling is coherently
reflected in the results obtained, highlighting the direct influence that system capacity has
on these effects.
In terms of cycling degradation, Figure 14 shows that an increasing S2P results in less
degradation. This phenomenon is explained by the fact that a greater storage capacity
implies charge and discharge cycles with a lower discharge depth. As previously discussed
in Figure 6, the depth of discharge is the main factor in cycling degradation, and the lower
it is, the lower the cycling degradation of the BESS. Therefore, it is confirmed that the
behavior observed in this case coincides with the theoretical model previously analyzed in
Figure 6.
 firmed that the behavior observed in this case coincides with the theoretical model pre-
viously analyzed in Figure 6.
Two sizes of BESSs are identified in Figure 14. A reduced storage capacity (S2P = 1.8)
presents a degradation due to calendar aging of 13.1% and due to cycling of 7%. On the
other hand, for a BESS with high storage capacity (S2P = 4.5), the degradation by calendar
Energies 2024, 17, 1485 25 of 30
is significantly higher than that by cycling. By calendar aging it increases up to 17%,
while by cycling it is reduced to 2.6%.

Figure14.
Figure 14.Calendar
Calendarand
andcycling
cyclingdegradation
degradationuntil
untilthe
theEoL
EoLasasa afunction
functionofofS2P.
S2P.

3.4. Two
Selection
sizesofofBESS
BESSs to are
Guarantee High
identified inAvailability
Figure 14. of A Firm Power
reduced storage capacity (S2P = 1.8)
presents a degradation due to calendar aging of 13.1%
The AED indicator, used in studies [21,22], has made it possible and due to cycling of 7%. On
to quantify the
energy
other hand, for a BESS with high storage capacity (S2P
unavailability in a system similar to the one proposed in this work. = 4.5), the degradation by calendar
is significantly
The curves higher than that
in Figure 15 by
showcycling.
that By calendar
there agingof
is a range it increases up to 17%,
suitable storage while
sizes that
by cycling itthe
guarantee is reduced to 2.6%.
minimization of energy unavailability. The restriction is imposed that the
AED indicator should not exceed 5%.
3.4. Selection of BESS to Guarantee High Availability of Firm Power
Energies 2024, 17, x FOR PEER REVIEW
In order to examine the effect of degradation on the AED indicator, several scenarios 26 of 31
have The AED
been indicator,
presented in used
Figure in 15.
studies [21,22],
The “No hascurve,
Deg.” made colored
it possible to quantify
in red, energy
simulates a hy-
unavailability in a system
pothetical scenario in whichsimilar
theto the one
BESS proposed
operates in this
without anywork.degradation throughout its
Thelife
useful curves in Figure
until the End of 15 show
Life. that there
The “Deg. EoL”iscurve,
a range of suitable
represented storage
in blue, sizesa that
reflects more
tion phenomena
guarantee the in the
minimization BESS. of This
energycontrast underscores
unavailability. The the importance
restriction
realistic scenario where the system experiences degradation until the End of Life. is of incorporating
imposed that the
degradation
AED indicator factors
should in the
not design
exceed and
5%. sizing of BESSs.
The results obtained demonstrate that, to ensure an energy unavailability of less
than 5%, a BESS with a capacity of S2P = 1.5 would be adequate in the theoretical scenario
without degradation. However, when considering a realistic system of degradation in
storage systems, it is necessary to increase the capacity to a value of S2P = 1.8 in order to
reach the same threshold of energy unavailability. Consequently, it is significant that it is
necessary to increase the storage capacity by 20% when taking into account the degrada-

Figure 15. AED until EoL with and without degradation, and AED until end of simulation based on
Figure 15. AED until EoL with and without degradation, and AED until end of simulation based
S2P.
on S2P.

It is commonly accepted to replace storage systems when they reach their EoL and
the degradation reaches 20%. However, as the primary purpose of the PV-BESS is to en-
sure a constant power supply with low unavailability (AED < 5%), it is advisable to con-
sider the possibility of slightly oversizing the BESS by extending its useful life. This
oversizing, as illustrated in Figure 13, involves increasing the S2P parameter and extends
Energies 2024, 17, 1485 26 of 30

In order to examine the effect of degradation on the AED indicator, several scenarios
have been presented in Figure 15. The “No Deg.” curve, colored in red, simulates a
hypothetical scenario in which the BESS operates without any degradation throughout its
useful life until the End of Life. The “Deg. EoL” curve, represented in blue, reflects a more
realistic scenario where the system experiences degradation until the End of Life.
The results obtained demonstrate that, to ensure an energy unavailability of less than
5%, a BESS with a capacity of S2P = 1.5 would be adequate in the theoretical scenario with-
out degradation. However, when considering a realistic system of degradation in storage
systems, it is necessary to increase the capacity to a value of S2P = 1.8 in order to reach the
same threshold of energy unavailability. Consequently, it is significant that it is necessary to
increase the storage capacity by 20% when taking into account the degradation phenomena
in the BESS. This contrast underscores the importance of incorporating degradation factors
in the design and sizing of BESSs.
It is commonly accepted to replace storage systems when they reach their EoL and the
degradation reaches 20%. However, as the primary purpose of the PV-BESS is to ensure a
constant power supply with low unavailability (AED < 5%), it is advisable to consider the
possibility of slightly oversizing the BESS by extending its useful life. This oversizing, as
illustrated in Figure 13, involves increasing the S2P parameter and extends the time period
until the EoL is reached. At the same time, this oversizing compensates for the gradual
decrease in capacity due to degradation.
In Figure 15, the blue curve, labeled “Deg. EoL”, shows how the AED rate behaves
in a simulation until reaching the end of the BESS useful life of 44,352 h. On the other
hand, the curve in yellow, called “Deg. Sim.”, represents the AED index for the complete
simulation, which covers a total of 96,432 h.
Selecting a value of S2P = 1.8 is considered appropriate to ensure that unavailabilities
are less than 5% until the BESS EoL is reached. However, as seen in Figure 15, oversizing the
BESS with a value of S2P = 2 ensures that the AED index remains below 5% throughout the
simulation period, thus prolonging the useful life of batteries beyond the EoL. Specifically,
this strategy extends the useful life from 44,352 h to a total of 96,432 h.
Additionally, Figure 15 shows a saturation zone beyond which increases in storage
capacity do not lead to substantial reductions in unavailability. Therefore, the choice of the
S2P parameter must be made before reaching the saturation point to optimize the selected
storage size.
For the particular case proposed in Figure 15, the choice of S2P must be carried out in
a way that guarantees that the unavailability (AED) remains below 5% at the end of the
saturation elbow.
At the bottom of the zoom box in Figure 15, the second derivatives of the curves
are plotted to highlight possible significant changes in the trend. The second derivative
provides details about the concavity of the function. Therefore, the S2P value is chosen at
which the maximum of each of these derivatives is reached, since this indicates a change in
the trend.
The choice of S2P is optimized to ensure that unavailability remains below 5%. For
a simulation up to the EoL, a value of S2P = 1.9 is required. However, if you are looking
to prolong the life of the batteries, you could consider selecting a value of S2P = 2.1, as
long as no failures occur in the batteries, since beyond the EoL, the manufacturer does not
guarantee its correct functioning.
This strategy of prolonging the useful life of the batteries is presented as an innova-
tive and risky choice, although valid, in terms of achieving reduced unavailability while
optimizing the size of the BESS.

3.5. The Influence of the Power Setpoint on the Selection of the BESS with Degradation
In order to analyze the behavior of the AED indicator and determine the appropriate
size of the BESS for different power setpoints, a comparison of three different scenarios is
carried out. The first is a case of reduced power supply (CPOF = 0.04) to guarantee high
 curve represents the theoretical situation without degradation effects, while the second
incorporates the degradation effects proposed in this study. For ease of identification, the
curve reflecting the degradation is labeled with the letter “D” in the legend of Figure 16.
In order to guarantee that the PV-BESS complies with energy unavailability of less
Energies 2024, 17, 1485 than 5% for a CPOF of 0.04, it can be seen in Figure 16 that the case without degradation 27 of 30
requires an S2P = 0.5. However, when considering degradation phenomena, it is neces-
sary to increase S2P to 0.62, which implies an increase in storage of 24%. For a CPOF of
energy
0.1, it availability
is necessaryfortothe system.S2P
increase Thefrom
second
1.5 istoa 1.8
casewhen
with including
a high power supply close
degradation to
effects,
the maximum
which capacity
implies factor that
a 20% increase inaS2P.
PV plant
For acan
CPO provide (CPO
F of 0.16, = 0.16). The
it isF necessary tothird caseS2P
increase is
the one proposed in the previous subsection (CPO = 0.1) as a reference to
from 3.5 to 3.85 by including the degradation of the BESS, which represents an increase of
F compare results.
10%.In Figure 16, two different curves are presented for each scenario analyzed. One
curve Consequently,
represents the theoretical situation
it is significant without
to highlight thatdegradation
as the power effects, while
setpoint the second
supplied by the
incorporates the degradation effects proposed in this study. For ease of identification,
system increases, the influence of degradation decreases. This implies that it is necessary the
curve reflecting
to oversize thethe
BESSdegradation
less to meetis the
labeled with the
proposed letteravailability
energy “D” in the objective.
legend of Figure 16.

Figure
Figure 16.16.AED
AEDasasa afunction
functionofofS2P
S2Pfor
for variousCPO
various CPOwith
F with and without aging.
and without aging.
F

In order to guarantee that the PV-BESS complies with energy unavailability of less
than 5% for a CPOF of 0.04, it can be seen in Figure 16 that the case without degradation
requires an S2P = 0.5. However, when considering degradation phenomena, it is necessary
to increase S2P to 0.62, which implies an increase in storage of 24%. For a CPOF of 0.1, it is
necessary to increase S2P from 1.5 to 1.8 when including degradation effects, which implies
a 20% increase in S2P. For a CPOF of 0.16, it is necessary to increase S2P from 3.5 to 3.85 by
including the degradation of the BESS, which represents an increase of 10%.
Consequently, it is significant to highlight that as the power setpoint supplied by the
system increases, the influence of degradation decreases. This implies that it is necessary to
oversize the BESS less to meet the proposed energy availability objective.

4. Conclusions
The high cost of storage systems continues to be a gap in the academic literature when
investigating PV-BESS hybrid plants operating with a constant power supply. However,
the falling price of storage systems could make this configuration viable.
This study focuses on the energy aspect, using a realistic model that includes the
detailed degradation effects of storage. Although this research focused on a specific
case, the results can be extrapolated to any scenario with storage systems combined
with photovoltaics.
Simulations have assessed the degradation of the BESS and its significant impact on
storage capacity. This degradation has a direct impact on cycling operations. Comparing
the proposed model with and without the degradation, it is found that it is necessary to
increase the storage capacity by 20% to keep the energy unavailability indicator at the
same level.
Energies 2024, 17, 1485 28 of 30

The analysis of the AED indicator in relation to S2P identifies the presence of a
saturation bend, where increases in S2P do not significantly reduce the AED indicator.
Therefore, the saturation bend together with the 5% limit on AED delimits a region in
which to select the storage size, avoiding unnecessary oversizing.
It has also been observed that it is beneficial to oversize the BESS slightly. This not
only prolongs its lifetime but also compensates for the reduction in capacity, which keeps
the AED indicator at a low level. Specifically, for an intermediate power setpoint, it is
concluded that selecting an S2P value of 2.1 instead of 1.9 extends battery life by almost
twice as long.
The proposed model is adaptable and can be applied to other power generation
technologies. Since the implemented algorithm performs hourly operations in a vectorial
manner, it would be feasible to include an additional column vector in addition to the
current PV production. Therefore, as a future line of work, all the simulations proposed
in this article could be carried out using a combination of power generation technologies.
Wind generation could be a suitable option to combine with PV generation due to their
different generation patterns. Wind tends to produce more energy at night and on cloudy
days, while solar tends to produce more energy in the middle of the day and under
clear skies.

Author Contributions: Conceptualization, J.A.T.-G. and Á.A.B.-R.; methodology, J.A.T.-G. and

Á.A.B.-R.; software, J.A.T.-G.; validation, J.A.T.-G. and Á.A.B.-R.; formal analysis, J.A.T.-G. and
Á.A.B.-R.; investigation, J.A.T.-G. and Á.A.B.-R.; resources, J.A.T.-G.; data curation, J.A.T.-G.; writing—
original draft preparation, J.A.T.-G. and Á.A.B.-R.; writing—review and editing, Á.A.B.-R.; visualiza-
tion, J.A.T.-G.; supervision, Á.A.B.-R.; project administration, Á.A.B.-R. All authors have read and
agreed to the published version of the manuscript.
Funding: This research received no external funding.
Data Availability Statement: Data are contained within the article.
Conflicts of Interest: The authors declare no conflicts of interest. The funders had no role in the design
of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript; or
in the decision to publish the results.

Abbreviations

PV Photovoltaic
DG Distributed generation
BESS Battery energy storage systems
MV/LV Medium Voltage/Low Voltage
PV-CPG Photovoltaic Constant Power Generation
PCC Point of common coupling
CF Capacity factor
CPOF Constant Power Operation Factor
AED Annual Energy Deviation
SoC State of charge
SEI Solid electrolyte interface
DoD Depth of discharge
NEC Number of equivalent cycles

References
1. World Energy Outlook 2023—Analysis. Available online: https://www.iea.org/reports/world-energy-outlook-2023 (accessed
on 1 February 2024).
2. Net Zero by 2050—Analysis. Available online: https://www.iea.org/reports/net-zero-by-2050 (accessed on 8 February 2024).
3. Communication from the commission to the European parliament, the council, the European economic and social committee and
the committee of the regions. A Blueprint to Safeguard Europe’s Water Resources. Available online: https://www.eea.europa.eu/
policy-documents/a-blueprint-to-safeguard-europes (accessed on 1 February 2024).
Energies 2024, 17, 1485 29 of 30

4. Bayod-Rújula, A.A. Future Development of the Electricity Systems with Distributed Generation. Energy 2009, 34, 377–383.
[CrossRef]
5. Xia, Q.; Debnath, S.; Saeedifard, M.; Marthi, P.R.V.; Arifujjaman, M. Energy Storage Sizing and Operation of an Integrated
Utility-Scale PV+ESS Power Plant. In Proceedings of the 2020 IEEE Power & Energy Society Innovative Smart Grid Technologies
Conference (ISGT), Washington, DC, USA, 17–20 February 2020; pp. 1–5.
6. U.S. Energy Information Administration. Form EIA-923, Power Plant Operations Report. Capacity Factors for Utility Scale 1215
Generators Primarily Using Non-Fossil Fuels. Available online: https://www.iea.org/reports/grid-scale-storage (accessed on 16
October 2023).
7. Bird, L.; Cochran, J.; Wang, X. Wind and Solar Energy Curtailment: Experience and Practices in the United States; Report No.
NREL/TP--6A20-60983; National Renewable Energy Lab.(NREL): Golden, CO, USA, 2014; p. 1126842.
8. Abdelrazek, S.; Kamalasadan, S. A Weather-Based Optimal Storage Management Algorithm for PV Capacity Firming. IEEE Trans.
Ind. Appl. 2016, 52, 5175–5184. [CrossRef]
9. Yoo, Y.; Jang, G.; Jung, S. A Study on Sizing of Substation for PV With Optimized Operation of BESS. IEEE Access 2020, 8,
214577–214585. [CrossRef]
10. Schmid, O.; Staffell, I. Experience Curve Data for Energy Storage Technologies. In Monetizing Energy Storage—A Toolkit to Assess
Future Cost and Value; Oxford University Press: Oxford, UK, 2023.
11. Keerthisinghe, C.; Mickelson, E.; Kirschen, D.S.; Shih, N.; Gibson, S. Improved PV Forecasts for Capacity Firming. IEEE Access
2020, 8, 152173–152182. [CrossRef]
12. Ismael, S.; Abdel Aleem, S.; Abdelaziz, A.; Zobaa, A. State-of-the-Art of Hosting Capacity in Modern Power Systems with
Distributed Generation. Renew. Energy 2019, 130, 1002–1020. [CrossRef]
13. Shin, H.; Hur, J. Optimal Energy Storage Sizing with Battery Augmentation for Renewable-Plus-Storage Power Plants. IEEE
Access 2020, 8, 187730–187743. [CrossRef]
14. Shin, H.; Roh, J.H. Framework for Sizing of Energy Storage System Supplementing Photovoltaic Generation in Consideration of
Battery Degradation. IEEE Access 2020, 8, 60246–60258. [CrossRef]
15. Rehman, W.; Bo, R.; Mehdipourpicha, H.; Kimball, J.W. Sizing battery energy storage and PV system in an extreme fast charging
station considering uncertainties and battery degradation. Appl. Energy 2022, 313, 118745. [CrossRef]
16. Förstl, M.; Azuatalam, D.; Chapman, A.; Verbič, G.; Jossen, A.; Hesse, H. Assessment of residential battery storage systems and
operation strategies considering battery aging. Int. J. Energy Res. 2019, 44, 718–731. [CrossRef]
17. Abdelrazek, S.; Kamalasadan, S. An Optimal Storage Management Algorithm for PV Capacity Firming Based on Weather Patterns.
In Proceedings of the IEEE Industry Applications Society Annual Meeting, Addison, TX, USA, 18–22 October 2015; pp. 1–8.
18. Fujioka, H.; Yamamoto, H.; Okano, K. Effectiveness Verification of Energy Storage Systems toward Stable Electricity Supply
with 100% Variable Renewable Energy. In Proceedings of the International Conference on Power and Energy Systems (ICPES),
Colombo, Sri Lanka, 21–22 December 2018; pp. 60–64.
19. Roy, S.; Sinha, P.; Ismat Shah, S. Assessing the Techno-Economics and Environmental Attributes of Utility-Scale PV with Battery
Energy Storage Systems (PVS) Compared to Conventional Gas Peakers for Providing Firm Capacity in California. Energies 2020,
13, 488. [CrossRef]
20. Doukas, D.I.; Papastergiou, K.; Bakas, P.; Marinopoulos, A. Energy Storage Sizing for Large Scale PV Power Plants Base-Load
Operation—Comparative Study & Results. In Proceedings of the 2012 38th IEEE Photovoltaic Specialists Conference, Austin, TX,
USA, 3–8 June 2012; pp. 000570–000574.
21. Tejero-Gómez, J.A.; Bayod-Rújula, Á.A. Analysis of Photovoltaic Plants with Battery Energy Storage Systems (PV-BESS) for
Monthly Constant Power Operation. Energies 2023, 16, 4909. [CrossRef]
22. Bayod-Rújula, Á.A.; Tejero-Gómez, J.A. Analysis of the Hybridization of PV Plants with a BESS for Annual Constant Power
Operation. Energies 2022, 15, 9063. [CrossRef]
23. Perez, R.; Taylor, M.; Hoff, T.; Ross, J.P. Reaching Consensus in the Definition of Photovoltaics Capacity Credit in the USA: A
Practical Application of Satellite-Derived Solar Resource Data. IEEE J. Sel. Top. Appl. Earth Obs. Remote Sens. 2008, 1, 28–33.
[CrossRef]
24. Photovoltaic Geographical Information System (PVGIS)—European Commission. Available online: https://joint-research-centre.
ec.europa.eu/photovoltaic-geographical-information-system-pvgis_en (accessed on 3 January 2024).
25. Photovoltaic Module Datasheet. Model Tallmax TSM-DE15V(II) 490W Trinasolar. Available online: https://static.trinasolar.
com/sites/default/files/DT-M-0007%20E%20Datasheet_TallmaxM_DE15V(II)_NA_EN_2022_1195A_web.pdf (accessed on 12
December 2023).
26. Photovoltaic Module Soiling Map. Available online: https://www.nrel.gov/pv/soiling.html (accessed on 12 December 2023).
27. Walker, A.; Desai, J. Understanding Solar Photovoltaic System Performance: An Assessment of 75 Federal Photovoltaic Systems; Report
No. NREL/TP-5C00-80881; USDOE Office of Energy Efficiency and Renewable Energy (EERE): Washington, DC, USA, 2021;
p. 1838130.
28. PV Inverter Datasheet. Model INGECON SUN 350TL M12 Ingeteam. Available online: https://www.ingeteam.com/es-es/
sectores/energia-fotovoltaica/p15_24_727/330-350tl-m12.aspx (accessed on 16 January 2024).
29. Yang, H.; Zhang, Y.; Ma, Y.; Zhou, M.; Yang, X. Reliability Evaluation of Power Systems in the Presence of Energy Storage System
as Demand Management Resource. Int. J. Electr. Power Energy Syst. 2019, 110, 1–10. [CrossRef]
Energies 2024, 17, 1485 30 of 30

30. Almazrouei, S.; Hamid, A.; Mehiri, A. Energy Management for Large-Scale Grid Connected PV—Batteries System. In Proceedings
of the International Renewable and Sustainable Energy Conference (IRSEC), Tangier, Morocco, 4–7 December 2017; pp. 1–5.
31. Large-Scale Battery Energy Storage Inverter Datasheet. Model Ingecon Sun Storage 100TL Ingeteam. Available online:
https://www.ingeteam.com/es-es/sectores/energia-fotovoltaica/p15_24_624_450/ingecon-sun-100tl.aspx (accessed on 10
January 2024).
32. Battery Energy Storage System Datasheet. Model TS-I HV 100 E TESVOLT. Available online: https://www.tesvolt.com/en/
products/e-serie/ts-i-hv-100-e.html (accessed on 3 January 2024).
33. Masters, G. Renewable and Efficient Electric Power Systems; John Wiley & Sons, Ltd.: Hoboken, NJ, USA; Stanford University:
Stanford, CA, USA, 2004; ISBN 978-0-471-66882-4.
34. IRENA. Renewable Power Generation Costs in 2021; International Renewable Energy Agency: Abu Dhabi, United Arab Emirates, 2022.
35. Fiorotti, R.; Fardin, J.F.; Encarnação, L.F.; Donadel, C.B. A Methodology to Determine the Firm Capacity of Distributed Generation
Units. In Proceedings of the 2015 IEEE PES Innovative Smart Grid Technologies Latin America (ISGT LATAM), Montevideo,
Uruguay, 5–7 October 2015; pp. 461–466.
36. Zhao, C.; Andersen, P.B.; Træholt, C.; Hashemi, S. Grid-Connected Battery Energy Storage System: A Review on Application and
Integration. Renew. Sustain. Energy Rev. 2023, 182, 113400. [CrossRef]
37. Muenzel, V.; De Hoog, J.; Brazil, M.; Vishwanath, A.; Kalyanaraman, S. A Multi-Factor Battery Cycle Life Prediction Methodology
for Optimal Battery Management. In Proceedings of the 2015 ACM Sixth International Conference on Future Energy Systems,
Bangalore, India, 14–17 July 2015; ACM: New York, NY, USA; pp. 57–66.
38. Xu, B.; Zhao, J.; Zheng, T.; Litvinov, E.; Kirschen, D.S. Factoring the Cycle Aging Cost of Batteries Participating in Electricity
Markets. In Proceedings of the 2018 IEEE Power & Energy Society General Meeting (PESGM), Portland, OR, USA, 5–10
August 2018. [CrossRef]
39. ASTM E1049-85; Standard Practices for Cycle Counting in Fatigue Analysis. ASTM International: West Conshohocken, PA,
USA, 2017.
40. Hossain, M.A.; Pota, H.; Squartini, S.; Zaman, F.; Guerrero, J. Energy Scheduling of Community Microgrid with Battery Cost
Using Particle Swarm Optimisation. Appl. Energy 2019, 254, 113723. [CrossRef]
41. Alhaider, F.; Klein, T.; Gerhard, S. A Simplified Methodology to Quickly Build a Degradation Prediction Model for Lithium-Ion
Batteries. In Proceedings of the NEIS 2018 Conference on Sustainable Energy Supply and Energy Storage Systems, Hamburg,
Germany, 20–21 September 2018; pp. 1–8.
42. Collath, N.; Tepe, B.; Englberger, S.; Jossen, A.; Hesse, H. Aging Aware Operation of Lithium-Ion Battery Energy Storage Systems:
A Review. J. Energy Storage 2022, 55, 105634. [CrossRef]
43. Waldmann, T.; Wilka, M.; Kasper, M.; Fleischhammer, M.; Wohlfahrt-Mehrens, M. Temperature Dependent Ageing Mechanisms
in Lithium-Ion Batteries—A Post-Mortem Study. J. Power Sources 2014, 262, 129–135. [CrossRef]
44. Gantenbein, S.; Schönleber, M.; Weiss, M.; Ivers-Tiffée, E. Capacity Fade in Lithium-Ion Batteries and Cyclic Aging over Various
State-of-Charge Ranges. Sustainability 2019, 11, 6697. [CrossRef]
45. Fioriti, D.; Pellegrino, L.; Lutzemberger, G.; Micolano, E.; Poli, D. Optimal Sizing of Residential Battery Systems with Multi-Year
Dynamics and a Novel Rainflow-Based Model of Storage Degradation: An Extensive Italian Case Study. Electr. Power Syst. Res.
2022, 203, 107675. [CrossRef]
46. Vetter, J.; Novák, P.; Wagner, M.R.; Veit, C.; Möller, K.-C.; Besenhard, J.O.; Winter, M.; Wohlfahrt-Mehrens, M.; Vogler, C.;
Hammouche, A. Ageing Mechanisms in Lithium-Ion Batteries. J. Power Sources 2005, 147, 269–281. [CrossRef]
47. Edge, J.; O’Kane, S.; Prosser, R.; Kirkaldy, N.; Patel, A.; Hales, A.; Ghosh, A.; Ai, W.; Chen, J.; Jiang, J.; et al. Lithium Ion Battery
Degradation: What You Need to Know. Phys. Chem. Chem. Phys. 2021, 23, 8200–8221. [CrossRef] [PubMed]
48. Liu, K.; Ashwin, T.R.; Hu, X.; Lucu, M.; Widanage, W.D. An Evaluation Study of Different Modelling Techniques for Calendar
Ageing Prediction of Lithium-Ion Batteries. Renew. Sustain. Energy Rev. 2020, 131, 110017. [CrossRef]
49. Sayem, M.A.; Hannan, M.A.; Hossain Lipu, M.S.; Mannan, M.; Ker, P.J.; Hossain, M.J. State of the art of lithium-ion battery
material potentials: An analytical evaluations, issues and future research directions. J. Clean. Prod. 2023, 394, 136246.
50. Birkl, C. Oxford Battery Degradation Dataset 1; University of Oxford: Oxford, UK, 2017.
51. Datasheet Battery Lithium-Ion Battery 24V—50Ah—1.28kWh—PowerBrick+. LiFePO4 Battery. Available online: https:
//www.powertechsystems.eu/home/products/24v-lithium-battery-pack-powerbrick/50ah-24v-lithium-ion-battery-pack-1-
28kwh-powerbrick-lithium/ (accessed on 26 July 2023).
52. Naumann, M.; Schimpe, M.; Keil, P.; Hesse, H.; Jossen, A. Analysis and Modeling of Calendar Aging of a Commercial
LiFePO4/Graphite Cell. J. Energy Storage 2018, 17, 153–169. [CrossRef]
53. Wildfeuer, L.; Karger, A.; Aygül, D.; Wassiliadis, N.; Jossen, A.; Lienkamp, M. Experimental Degradation Study of a Commercial
Lithium-Ion Battery. J. Power Sources 2023, 560, 232498. [CrossRef]

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual
author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to
people or property resulting from any ideas, methods, instructions or products referred to in the content.