Journal of Energy Storage 86 (2024) 111179

Contents lists available at ScienceDirect

Journal of Energy Storage

journal homepage: www.elsevier.com/locate/est

Review article

A review of battery energy storage systems and advanced battery

management system for different applications: Challenges
and recommendations
Shaik Nyamathulla , C. Dhanamjayulu *
School of Electrical Engineering, Vellore Institute of Technology, Vellore, Tamil Nadu, India

A R T I C L E I N F O A B S T R A C T

Keywords: Energy storage systems are designed to capture and store energy for later utilization efficiently. The growing
Energy storage system (ESS) energy crisis has increased the emphasis on energy storage research in various sectors. The performance and
Battery management systems (BMS) efficiency of Electric vehicles (EVs) have made them popular in recent decades. The EVs are the most promising
State of charge (SoC)
answers to global environmental issues and CO2 emissions. Battery management systems (BMS) are crucial to the
State of health (SoH)
functioning of EVs. An efficient BMS is crucial for enhancing battery performance, encompassing control of
Cell balancing
Electric vehicles (EVs) charging and discharging, meticulous monitoring, heat regulation, battery safety, and protection, as well as
precise estimation of the State of charge (SoC). The current understanding of EV technology, its advancements,
limitations, and effects on achieving BMS (Sustainable Development Goals) SDGs remains unexplored, despite
the existence of several studies on the topic. This article reviews various aspects of battery storage technologies,
materials, properties, and performance. This review highlights the significance of battery management systems
(BMSs) in EVs and renewable energy storage systems, with detailed insights into voltage and current monitoring,
charge-discharge estimation, protection and cell balancing, thermal regulation, and battery data handling. The
study extensively investigates traditional and sophisticated SoC estimation methods, highlighting their pros and
cons. The review underscores the critical role of advanced BMSs for successful EV adoption and addresses the
challenges that must be overcome. This comprehensive resource offers valuable insights for engineers, re­
searchers, and EV manufacturers, presenting detailed analyses, applications, challenges, and recommendations
relevant to the field.

1. Introduction of industries are discovering applications for energy storage systems

(ESS), encompassing areas like EVs, renewable energy storage, micro/
Energy storage systems (ESS) serve an important role in reducing the smart-grid implementations, and more. The latest iterations of electric
gap between the generation and utilization of energy, which benefits not vehicles (EVs) can reliably replace conventional internal combustion
only the power grid but also individual consumers. An increasing range engines (ICEs). Different fossil fuels are used by ICE-powered

Abbreviations: EV, Electric vehicle; Li-ion, Lithium-ion; V2G, Vehicle to grid; PV, Photovoltaic; PVB, Photovoltaic battery; MPPT, Maximum power point tracking;
FL, Fuzzy Logic; PbA, Lead-acid; SoC, State of charge; SoH, State of Health; SoF, State of function; SoP, State of power; SoE, State of Energy; SoS, State of Safety; BMS,
Battery management system; CAN, Controlled area network; CNN, Convolutional Neural Networks; EMI, Electromagnetic interference; ISO, International Organi­
zation for Standardization; KF, Kalman filter; ANFIS, Adaptive Neuro-Fuzzy Inference System; SPKF, Sigma-Point Kalman Filter; EIS, Electrochemical Impedance
Spectroscopy; EKF, Extended Kalman Filter; ECM, Electrochemical Models; AF, Adaptive Filter; LCO, Lithium cobalt oxide; LFP, Lithium iron phosphate; LMO,
Lithium manganese oxide; LNMC, Lithium nickel manganese cobalt oxide; LNCA, Lithium nickel cobalt aluminium oxide; LTO, Lithium titanate oxide; NAMHE, Noise
Adaptive Moving Horizon Estimation; LSTM, Long Short-Term Memory; LNO, Lithium Nickel Oxide; NN, Neural Network; AEKF, Adaptive Extended Kalman filter;
CC, Coulomb Counting; ADC, Analog-to-digital conversion; ANN, Artificial neural network; SVM, Support Vector Machine; RC, Resonant converter; PTT, Pulse Test
Technique; OCV, Open-Circuit Voltage; RES, Renewable energy sources; DNN, Deep Neural Networks; RUL, Remaining useful life; SDG, Sustainable Development
Goals; GA, Genetic Algorithm; SMO, Sliding Mode Observer; NLO, Nonlinear Observer; UKF, Unscented Kalman Filter; UPF, Unified particle Filter; RVM, Regression
Vector Machine; PSO, Particle Swarm Optimization; AFS, Artificial fish swarm; RMSE, Root Mean Square Error; MMAE, Multiple Model Adaptive Estimation.
* Corresponding author.
E-mail address: [email protected] (C. Dhanamjayulu).

https://doi.org/10.1016/j.est.2024.111179
Received 7 September 2023; Received in revised form 5 February 2024; Accepted 28 February 2024
Available online 7 March 2024
2352-152X/© 2024 Elsevier Ltd. All rights reserved.
S. Nyamathulla and C. Dhanamjayulu Journal of Energy Storage 86 (2024) 111179

Fig. 1. A detailed analysis of the number of reviewed published articles on BMS.

(Data sources: Scopus; keywords: Li-ion battery for EV and battery management system; access date: 25 November 2022). https://www.sciencedirect.com/search?
qs=Li-ion%20batteries%20for%20EV.

transportation (cars, trucks, aircraft, etc.). Carbon dioxide (CO2), sulfur energy storage systems now in use, such as lithium-ion batteries, lead
dioxide (SO2), carbon monoxide (CO), and nitrogen oxide (NO) emis­ acid batteries, nickel-cadmium batteries, sodium-sulfur batteries, and
sions have two primary causes: internal combustion engines (ICE) and zebra batteries. According to Baker [1], there are several different types
industries. These gases cause air pollution, which adds to the greenhouse of electrochemical energy storage devices. The lithium-ion battery
effect. Increasing carbon emissions are the principal cause of global performance data supplied by Hou et al. [2] will also be analysed. Nitta
warming and are now one of the most significant concerns for scientists et al. [2] presented a thorough review of the history, current state of the
and academics. However, there exists a requirement for extensive art, and prospects of research into anode and cathode materials for
research on a broad spectrum of concerns, which encompass, among lithium batteries. Nitta et al. presented several methods to improve the
other things, the selection of appropriate battery energy storage solu­ efficiency of Li-ion batteries in their study. These include scaling down
tions, the development of rapid charging methodologies, the enhance­ the size of the active material, combining many materials into one,
ment of power electronic devices, the optimization of conversion doping and functionalizing the material, fine-tuning the particle shape,
capabilities, and the integration of hybridizing algorithms or coating or encasing the material, and changing the electrolyte.
methodologies. Lithium batteries are becoming increasingly important in the elec­
This article provides an overview of the many electrochemical trical energy storage industry as a result of their high specific energy and

Fig. 2. A comprehensive examination of battery management systems research.

2
S. Nyamathulla and C. Dhanamjayulu Journal of Energy Storage 86 (2024) 111179

Fig. 3. Categorization of energy storage systems.

energy density. The literature provides a comprehensive summary of the • This review provides a comprehensive analysis of several battery
major advancements and key constraints of Li-ion batteries, together storage technologies, materials, properties, and performance.
with the existing knowledge regarding their chemical composition. The • This article provides a comprehensive explanation of the advanced
Li-ion battery is classified as a lithium battery variant that employs an techniques, algorithms, and optimization methodologies utilized in
electrode material consisting of an intercalated lithium compound. The electric vehicles (EVs).
authors Bruce et al. (2014) investigated the energy storage capabilities • This research work comprehensively investigated the categorization
of Li-ion batteries using both aqueous and non-aqueous electrolytes, as of traditional and sophisticated SoC estimation methodologies as
well as lithium-Sulfur (Li–S) batteries. The authors also compare the well as the associated advantages and drawbacks.
energy storage capacities of both battery types with those of Li-ion • Battery management systems (BMSs) are discussed in depth, as are
batteries and provide an analysis of the issues associated with cell their applications in EVs, and renewable energy storage systems are
operation and development. The authors propose that both batteries presented in this article.
exhibit enhanced energy density in comparison to Li-ion batteries and • This review covers topics ranging from voltage and current moni­
may also possess a greater potential for cost competitiveness relative to toring to the estimation of charge and discharge, protection and
Li-ion batteries. Still, it is suggested that more research into the chemical equalization to thermal management, and actuation of stored battery
processes involved with Li–O2 and Li–S cells is needed before they can data.
be sold to the public [3]. Thackeray and colleagues in 2015 presented a • This review demonstrates the difficulty of EV adoption without
comprehensive historical analysis of lithium-ion batteries, including addressing existing problems and developing superior BMSs.
their current state and advancements in lithium-air battery technology • The Engineers and researchers working on electric vehicles and
[4]. The number of reviewed published articles detailing the comparison manufacturers of EVs will benefit from the detailed discussion,
across Li-ion batteries and BMS is presented in Fig. 1. analysis, applications, challenges, and recommendations presented
The battery management system (BMS) is an essential component of in this article.
an energy storage system (ESS) and plays a crucial role in electric ve­
hicles (EVs), as seen in Fig. 2. This figure presents a taxonomy that The following sections of this article are divided into six categories:
provides an overview of the research. The Battery Management System Section 2 offers an overview of different battery energy storage tech­
(BMS) is a comprehensive framework that incorporates various pro­ nologies that have been demonstrated to differ in important perfor­
cesses and performance evaluation methods for several types of energy mance areas, such as specific power and specific energy. Section 3
storage devices (ESDs). It encompasses functions such as cell moni­ presents in depth the major components of battery management sys­
toring, power management, temperature management, charging and tems: algorithms, methodologies, approaches, controllers, and optimi­
discharging operations, health status monitoring, data acquisition, cell zation technologies. Section 4 reports on BMS applications. Section 5
protection, and lifespan estimation [5]. To ensure the effective moni­ highlights the outstanding difficulties and recommendations. Section 6
toring and operation of energy storage devices in a manner that pro­ concludes with the conclusions and future tendencies.
motes safety and well-being, it is necessary to employ a range of
techniques and control operations [6]. These measures should be 2. Energy storage systems (ESS)
designed to operate autonomously and without delay [7].
Electric vehicles (EVs) are regarded as an energy storage system An energy storage system (ESS) is a technology that captures and
(ESS) that is communicated inside a smart/micro-grid system. This stores energy for later use. The classification of energy storage encom­
system uses synchronized charging energies to offset the uneven power passes several categories. In the present scenario, Fig. 3 illustrates the
output from solar and wind sources. The integration of renewable en­ diverse energy storage categories, providing information on their tech­
ergy sources into the electrical grid may be effectively facilitated nical and economic specifications alongside their respective applications
through the utilization of vehicle-to-grid (V2G) and grid-to-vehicle [8].
(G2V) systems. This study aims to address the current limitations by
emphasising the potential of integrating electric vehicles (EVs) with 2.1. Battery energy storage systems (BESS)
photovoltaic (PV) systems. The research started with providing an
overview of energy storage systems (ESSs), battery management systems Electrochemical methods, primarily using batteries and capacitors,
(BMSs), and batteries suitable for EVs. can store electrical energy. Batteries are considered to be well-
The following are some of the contributions made by this review: established energy storage technologies that include notable

3
S. Nyamathulla and C. Dhanamjayulu Journal of Energy Storage 86 (2024) 111179

significant downsides. Cadmium harms both the environment and

human health [12].

2.1.3. Lithium-ion battery

One of the most popular EV batteries is lithium-ion. Li-ion batteries
are noted for their excellent energy density, efficiency, lifespan, and
high-temperature performance. It's still good for battery-powered EVs
[13]. The battery's biggest benefit is component recycling. Major
drawbacks are the high cost per kWh (135 USD/kWh) and the material's
unavailability. In terms of voltage, power, and energy, the LMO, LNMC,
and LNCA batteries are excellent [14]. For excellent lifetime and safety,
utilize LFP and LTO batteries. Additionally, LTO is cost-effective and
high-performance [15]. Table 1 presents a comparative analysis of
several categories of lithium-ion batteries [16].
The different positive electrode materials are explored in more detail
below [17].

• Lithium cobalt oxide—LiCoO2

In 1991, Sony introduced lithium cobalt oxide (LCO) to the market,

Fig. 4. The specific and volumetric energy density of various battery types. employing cobalt oxide as the cathode material, which was widely uti­
lized in lithium-ion battery technology at the time. It also exhibits a
moderate lifespan, lasting for a reasonable duration before requiring
characteristics such as high energy densities and elevated voltages [9]. A
replacement. Furthermore, it demonstrates significant safety features,
comprehensive examination has been conducted on several electrode
making it suitable for use in various electronic devices such as cameras,
materials and electrolytes to enhance the economic viability, energy
laptops, and tablets [11].
density, power density, cycle life, and safety attributes of batteries.
Fig. 4 shows the specific and volumetric energy densities of various
• Lithium manganese oxide—LiMn2O4
battery types of the battery energy storage systems [10].
Fig. 5 shows the classification of various Li-ion battery materials.
The LMO battery technology was created in the Bellcore lab in 1994.
This section provides a comprehensive examination and evaluation of
The internal resistance of LMO is decreased, and the charge/discharge
the diverse attributes, qualities, and essential constituents of battery
current flow is increased thanks to its 3D spinel design. When compared
storage in the context of electric vehicle (EV) applications [10].
to cobalt-based batteries, LMO has a capacity that is around 33 % lower.
LMO is being used in production right now in the Nissan Leaf EV [12].
2.1.1. Lead–acid (Pb–acid)
Lead-acid batteries are still widely utilized despite being an ancient
battery technology. The specific energy of a fully charged lead-acid
battery ranges from 20 to 40 Wh/kg. The inclusion of lead and acid in
Table 1
a battery means that it is not a sustainable technology. While it has a few Properties of different Li-ion batteries [14–18].
downsides, it's inexpensive to produce (about 100 USD/kWh), so it's a
Battery Voltage Specific Charge Discharge Lifespan
good fit for low-powered, small-scale vehicles [11]. type (V) energy (Wh/ (c) (c) (hrs)
kg)
2.1.2. Nickel–cadmium (NiCd) battery
LTO 2.3–2.6 75–85 1 10 3000–7000
The high energy density of nickel-cadmium (NC) batteries was LNO 3.6–3.8 160–200 0.7–1 1 >300
widely used in the 1990s. NC battery technology is used in fields like LFP 3.2–3.4 90–120 1 1 1000–2100
telecommunications and portable services to improve things like power LMO 3.7–3.9 100–145 0.7–1 1 300–750
quality and energy reserves. When compared to NiMH batteries, NC LNMC 3.8–4.1 150–210 0.7–1 1 1000–2100
LNCA 3.6–3.7 200–250 0.7 1 ~500
batteries have a far longer lifespan at 1500 cycles. Toxic metals like LCO 3.7–3.8 160–200 0.7–1 1 600–1000
cadmium are used in the production of NC, which is one of the material's

Fig. 5. Classification of various Li-ion battery materials.

4
S. Nyamathulla and C. Dhanamjayulu Journal of Energy Storage 86 (2024) 111179

Table 2
Comparison of different batteries for EV applications.
Parameter LiMn2O4 battery Pb–acid battery LiFePO4 battery Ni-MH battery LiCoO2 battery Ni-Cd battery

Nominal cell voltage 3.8 V 2V 3.5 V 1.5 V 3.6 V 1.25 V

Power density 1850 180 1850 150 1850 250–1000
Self –discharge <10 % <5 % <10 % >30 % <10 % >30 %
Peak load current >30C 5C >30C 5C >3C 20C
Weight Light Heavy Light Moderate Light Heavy
Discharge cut-off voltage 2.5–3.0 V 1.75 V 2.4–3.0 V 1.0 V 2.8 V 1.0 V
Specific Energy density 100–145 Wh/Kg 35–55 Wh/Kg 95–125 Wh/Kg 65–125 Wh/Kg 145–185 Wh/Kg 50–85 Wh/Kg
Life cycle (80 % discharge) 500–1000 250–350 1000–2000 200–300 500–1000 1000
Charging time <1 h 8–16 h <1 h 2–4 h 2–4 h 1h
Cut off charge voltage 3.6 V 2.40 V 4.20 V 3.60 V 4.20 V 3.6 V
Memory No No No Little No Yes
Overcharge tolerance Very low High Very low Moderate Very low Low
Eco-friendly Yes No Yes Yes Yes No

• Lithium iron phosphate—LiFePO4 advancement of electric vehicles [13].

The findings of the investigation indicated that phosphate exhibits • Lithium titanate—Li4Ti5O12
superior performance compared to LCO or LMO batteries in conditions
of elevated temperatures and when subjected to overcharging. Phos­ The use of LMO and LNCA as cathode materials and titanate as the
phates have favourable thermal stability, functioning effectively anode material establishes the spinel architecture of lithium titanate
throughout a temperature range spanning from − 30 ◦ C to 60 ◦ C. (LTO). The LTO technology has exceptional performance capabilities
and boasts an extended operational lifespan. In addition, it is worth
• Lithium nickel manganese cobalt oxide—Li(Ni,Mn,Co)O2 noting that LTO demonstrates safe operational performance even under
freezing temperature conditions [14]. Table 2 displays a comparison of
The commercialization of lithium nickel manganese cobalt oxide different types of batteries that could be suitable for electric vehicles
(LNMC) battery technology occurred in 2004. Additionally, LNMC ex­ [15].
hibits elevated power and energy density, along with enhanced The careful choice of materials is of utmost importance, particularly
longevity and performance. An increase in the proportion of manganese concerning the positive electrode. This component plays a critical role in
results in an augmentation of specific power, whereas an increase in the determining the battery's key properties, including power output, safety,
percentage of nickel leads to an augmentation of specific energy. cost, and longevity [16]. Energy storage systems play a crucial role in
the pursuit of a sustainable, dependable, and low-carbon energy future.
• Lithium nickel cobalt aluminium oxide—Li(Ni, Co, Al)O2 By improving the productivity and effectiveness of diverse energy-
generating and consumption processes, these systems are of utmost
Nickel-cobalt aluminium oxide (NCA) batteries were introduced in importance [17]. The proliferation of ESS is anticipated to experience
1999. The use of nickel as a cathode material reduces lithium cobalt significant growth in the foreseeable future due to technological ad­
oxide's cobalt dependence. Tesla, a prominent player in the automotive vancements and decreasing prices [18].
industry, is presently employing LNCA battery technology in the

Fig. 6. An overview of BMS.

5
S. Nyamathulla and C. Dhanamjayulu Journal of Energy Storage 86 (2024) 111179

Fig. 7. Cell monitoring block diagram of BMS.

Fig. 8. Voltage and current measurement block diagram of BMS.

3. Battery management systems (BMS) Managing, guarding, balancing, and regulating operations can all
benefit from the data gleaned by monitoring battery cells.
Battery management systems (BMSs) are systems that help regulate
battery function by electrical, mechanical, and cutting-edge technical 3.2. Voltage and current measurement
means [19]. By controlling and continuously monitoring the battery
storage systems, the BMS increases the reliability and lifespan of the Series and parallel battery cell connections to the battery bank pro­
EMS [20]. This is accomplished through a variety of control techniques, duce sufficient voltage and current. There are many voltage-measuring
including charge-discharge control, temperature control, cell potential, channels in EV battery packs due to the enormous number of cells in
current, and voltage monitoring [21]. The key components of the BMS series. It is impossible to estimate SoC or other battery states without a
for its efficient operation are represented in Fig. 6. precise measurement of a battery cell [23]. Using high-voltage current
sensors, the battery module's current is measured and then converted to
3.1. Battery cell monitoring a digital signal using an analog-to-digital converter (ADC), as repre­
sented in Fig. 8. The voltage and current measurements are then used to
In order to carry out the responsibilities of the BMS, it is necessary to calculate accurate estimates of SoC, SoH, and RUL [24].
have access to data on the charging and discharging, health, tempera­
ture, and problem diagnostics of the batteries, as represented in Fig. 7. 3.3. Data acquisition
While the battery is being charged or discharged, the cell might have a
variety of responses. So, it's important to constantly keep an eye on your The current, voltage, temperature, and state of charge (SoC) are only
batteries to learn more about their states and performance metrics [22]. a few of the characteristics of the battery pack that may be measured and

6
S. Nyamathulla and C. Dhanamjayulu Journal of Energy Storage 86 (2024) 111179

Fig. 9. Data acquisition block diagram of BMS.

Fig. 10. BMS-based cloud-integrated EV data acquisition framework.

estimated with the use of a data acquisition system (DAS). The DAS is an 3.5. State of charge (SoC)
integral part of the BMS, which has a microprocessor and programmed
instructions as shown in Fig. 9. The ADC component is responsible for The state of charge (SoC) can be defined as the ratio of the present
handling data conversion in the DAS. Exchange data and have conver­ accessible capacity to the maximum battery capacity. Where ‘Qo’, ‘Q’
sations with the BMS using a controller area network (CAN) bus and and ‘Qm’ represents the initial charge, quantity of electricity delivered
serial communication interface (SCI) modules. Fig. 10 shows a BMS that or supplied to the battery and maximum charge that can be stored in the
uses a cloud-based DAS platform to measure battery current, voltage, battery respectively. The state of charge may also be considered the
and temperature [24]. other way around and it is called the Depth of discharge (DoD). It can be
calculated as following equations.
3.4. Battery state estimation
(Qo + Q)
SoC (%) = × 100 (1)
Qm
Accurate battery status estimation is of utmost importance to effec­
tively estimate both battery charge and health. One way to figure out the DoD (%) = 100 − SoC (%) (2)
battery management system's monitoring parameters like state of charge
(SoC), state of health (SoH), remaining useful life (RUL), state of func­ The SoC value ranges from 0 to 100 %. If the SoC is 100 %, the
tion (SoF), state of performance (SoP), state of energy (SoE), state of battery is fully charged, whereas a SoC of 0 % indicates that the cell is
safety (SoS), and state of temperature (SoT) as shown in Fig. 11 [25]. totally discharged. Various techniques can be employed to estimate the
SoC, as seen in Fig. 12. The operational intricacies of these approaches
are elaborated upon in the subsequent discussion. In recent times, there

7
S. Nyamathulla and C. Dhanamjayulu Journal of Energy Storage 86 (2024) 111179

Fig. 11. BMS-supported battery state estimation framework.

Fig. 12. A classification of SoC estimation techniques.

has been significant interest in the application of machine learning and However, its primary drawback lies in the extended duration required to
deep learning techniques for SoC estimation. This is primarily due to its reach equilibrium. Hence, the use of an online estimate for SoC is
notable advantages, such as enhanced accuracy, greater learning ca­ deemed unsuitable. Hence, this approach is suitable only for applica­
pacity, superior generalization performance, and faster convergence tions with minimal power consumption. In addition, it is necessary to
speed [26]. conduct certain observations to quantify the discharge and charge
voltages. As seen in Fig. 13 from the article [18], batteries exhibit hys­
3.5.1. Conventional methods teresis characteristics, resulting in the charging process occurring at
high open circuit voltage (OCV) levels while the discharge process takes
• Open-circuit voltage method place at lower OCV levels. This phenomenon has been extensively
studied and documented [27].
The open-circuit voltage technique exhibits a notable degree of
precision, is readily implementable, and follows a direct approach. • Coulomb Counting (CC) method

8
S. Nyamathulla and C. Dhanamjayulu Journal of Energy Storage 86 (2024) 111179

Vt = Voc − Vdr − Vep (4)

where terminal voltage (Vt), open-circuit voltage (Voc), potential dif­

ference (Vdr), and electric potential (Vep). Methods for measuring the
SoC are analysed in Table 3.

3.5.2. Adaptive filter (AF) algorithm

• Kalman filter (KF) algorithm

The Kalman filter is a widely used and sophisticated technique that

finds frequent application in the fields of automotive engineering, nav­
Fig. 13. OCV vs SoC were tested at 25 C. ◦
igation tracking, and aerospace technology [34]. One notable charac­
teristic of the Kalman filter is its inherent ability to self-correct [35]. The
Table 3 Kalman filter linear model consists of a state equation that forecasts the
Comparison of common SoC estimate techniques. present state and a measurement equation that modifies the present
state [36]. These equations may be stated as follows:
SOC method Merits Demerits

OCV [27] • Simple to implement. • Offline State equation : xn+1 = An xn + Bn un + f n (5)

• Highly precise. • Low speed
• Cost-effective. Measurement equation : yn = Cn xn + Dn un + zn (6)
CC [28] • Low power required. • Low accuracy.
• Simple to implement. • Difficulty in defining the SOC The covariance matrices in consideration are denoted as A, B, C, and
initial values. D. The variable x stands for the system state, while the symbol for
EIS [30] • Low cost • Temperature effect process noise is f. The variable u stands for the control input, and the
• Highly accurate for stabilized • Aging effect.
impedance value.
variable y is a representation of the measurement input. Lastly, the
• Operates in online mode. measurement noise is denoted as z. The RC battery model proposed in
Model-Based • Operates in online mode. • It depends on the accuracy of the article [37] is known as the Kalman filter. In order to describe the
[33] • Highly accurate. the model. dynamic characteristics of batteries, the mathematical equations of the
RC model are transformed into a state-space model. The application of
The constant-current (CC) methodology is considered the most the Kalman filter results in a lower estimated root mean square (RMS)
straightforward method for estimating the SoC of a battery. This tech­ error for the SOC compared to the observed error [38].
nique offers the advantage of being easily implemented and requiring
minimal power calculations. The process of charging and draining the • Extended Kalman Filter (EKF)
battery is contingent upon the integration of the electric current with
respect to time [28]. The expression can be found in Eq. (3). The Extended Kalman Filter (EKF) has been utilized to operate in the
∫ context of nonlinear applications [39]. The use of partial derivatives and
i.ηdt first-order Taylor series expansion linearizes the battery model. The
SoC = 1 − (3)
Cm state-space model is linearized, and at each moment in time, it estab­
lishes an equation that correlates the projected battery value with the
where “ɳ” indicates the Coulombic efficiency, “i” indicates the current of observed voltage [40]. This process is conducted to accurately estimate
the battery, and “Cm ” indicates the total capacity. the restrictions for the state of charge (SoC). If the scheme exhibits a
high degree of nonlinearity, there is a possibility of encountering an
• Electrochemical Impedance Spectroscopy (EIS) error in the process of linearization [41]. Ultimately, the enhanced dual
AEKF technique was implemented, resulting in SoH and SoC estimate
To ensure the successful implementation of the Electrochemical errors that fell within the range of 1 % [42].
Impedance Spectroscopy (EIS) technique, it is crucial to utilize an
appropriate electrochemical model. Battery impedance is evaluated by • H ∞ filter
employing capacitances and inductances across a broad range of fre­
quencies [29]. Two capacitive arcs and one inductive arc operating at The proposed model utilizes a straightforward approach that does
both low and high frequencies are analogous to the described circuit not require knowledge of specific noise details or measurement features.
architecture. The model impedances are calculated using a method The system robustly operates under particular conditions by considering
called nonlinear least squares fitting (LSF). This method is used across just the time-varying properties of the battery. The accuracy of the
different states of charge (SoC). The aforementioned technique offers model is affected by hysteresis, aging, and temperature influences [43].
several advantages, including cost efficiency, online operability, and a The adaptive H ∞ filter was first described in [44] to estimate the SoC.
notable degree of precision [30]. This method evaluates system functions using a polynomial function and
compares them to the adaptive extended Kalman filter. The AHF
• Model-based SOC estimation demonstrated superior performance in terms of accuracy and computing
efficiency compared to alternative approaches [45].
Due to its inability to operate in real-time, the open-circuit voltage
method necessitates sufficient idle time for monitoring SoC [31]. • Sigma Point Kalman Filter
Consequently, its application is precluded during vehicle motion. Hence,
the creation of a battery model is crucial for the implementation of This technique is an alternative approach to state computation,
online SoC estimation in the context of online systems [32]. The battery characterized by its nonlinearity. In comparison to the extended Kalman
models that are often utilized consist of electrochemical models and filter, it is capable of producing more accurate results. Sigma-Point
equivalent circuit models. The concept can be articulated as follows: Kalman filter (SPKF) numerical approximation is sensitive [46]. The

9
S. Nyamathulla and C. Dhanamjayulu Journal of Energy Storage 86 (2024) 111179

Fig. 14. A neural network-based framework for SoC prediction

Table 4 Table 5
Comparison of adaptive filter SoC estimation techniques. Comparison of learning SoC estimation methods.
Method Merits Demerits Method Merits Demerits

KF [34] • Self correction nature • Can't directly estimate for SOC NN [50] • Accomplished work of batteries • Huge memory is required.
• Accurate • Complex calculations in nonlinear circumstances
• Intelligent control tool FL [51] • High accuracy • Complex
EKF [42] • Estimates the probability of • Limited durability • Effective in operation • Costly
nonlinear dynamic errors • Error in linearization • Storage is required
• Improves accuracy GA [54] • Accuracy is good • Computation is high
H ∞ filter • Cost of computation • Aging and temperature might • Robust • To obtain effective output
[44] • Precise performance cause accuracy values to deviate tuning of parameters is
SPKF [46] • Robustness • Heavy calculations required.
• Improvement in precision • Complicated SVM • Performs outstandingly for • Computation is high
[56] nonlinear models • Process error
• Performs well
programme selects sigma points with the same mean and covariance as
the model. The SPKF has the advantage of possessing a comparable
calculating method. In addition, the omission of Jacobian matrices in the sophisticated ANFIS [53].
the consideration of the Extended Kalman Filter (EKF) leads to a
reduction in complexity [47]. The SPKF exhibits enhanced precision, • Genetic algorithm (GA)
reduced memory consumption, and a decreased computing workload
[48]. The primary use of a genetic algorithm is the identification of the
optimal parameter to enhance the efficiency of the system. The use of
3.5.3. Learning algorithms this method has been seen in the fields of mathematics, physics, and
engineering, where it has been employed to ascertain nonlinear opti­
• Neural network (NN) algorithm mum parameters. The use of numerous driving cycles served to validate
the strategy. The results obtained from this method exhibited enhanced
An artificial neural network (ANN) is an algorithm that possesses the accuracy, with an error rate of <1 % [54].
ability to learn autonomously and exhibits intelligent behaviour. The
estimation of the state of charge (SoC) is achieved by utilizing trained • Support Vector Machine Algorithm (SVM)
data, even in the absence of prior knowledge of the original SoC data
[49]. The neural network topology, as shown in Fig. 14, comprises input, The Support Vector Machine (SVM) algorithm is a regression
hidden, and output layers. One advantage of a neural network (NN) is its approach that operates using a kernel function. The purpose of the
ability to effectively operate under non-linear battery conditions. One of kernel function is to transform non-linear data into a linear form,
the limitations associated with this approach is the substantial data allowing for more effective analysis in extreme measurements. In article
requirement for training as well as the need for significant memory [55], the support vector machine (SVM) approach was employed to
capacity to retain the acquired information [50]. In Table 4, an exami­ estimate the state of charge (SoC) of batteries. One of the advantages of
nation of several approaches for estimating the SoC using adaptive filters support vector machines (SVMs) is their ability to effectively handle
is presented. high-dimensional models and nonlinear data representations. The esti­
mation of the SoC is efficiently and precisely achieved through the use of
• Fuzzy logic algorithm training data. One limitation of this approach is the necessity for trial
and error, which may be time-consuming [56]. In Table 5, an exami­
The most prominent algorithm for expanding nonlinear, complicated nation of several methods used for estimating the state of charge algo­
prototypes using training data is FL. Defuzzification, rule-based inputs rithms is presented.
and outputs, and reasoning membership functions are used in fuzzy logic
[51]. The nonlinear model estimation function is strong. A complex 3.5.4. Nonlinear observer (NLO)
computation, dispensing unit, and memory storage are needed. FL was
used to estimate SoC using CC technique data [52]. SoC and impedances • Sliding Mode Observer (SMO)
are inputs at three frequencies in this approach. It forecasts SoC with a
maximum inaccuracy of ±5 %. Li-ion battery SoC is best estimated by The Sliding Mode Observer (SMO) algorithm is a training controller

10
S. Nyamathulla and C. Dhanamjayulu Journal of Energy Storage 86 (2024) 111179

Table 6 Table 8
Comparison of nonlinear observer SoC estimating techniques. SoC estimation hybrid methods analysis.
Method Merits Demerits Method Merits Demerits

SMO • Robustness • Difficult to alter switching gain CC and KF • Low power usage for • CC and KF errors need sophisticated
[57] • Stability improves [67] CC. computations.
NLO • Accuracy improves • Finding of gain matrix to minimize the • KF self-corrects and
[59] • Convergence speed error is challenging. is smart.
improves EKF [68] • Nonlinear dynamic • Weak resilience
error prediction. • Linearization mistake.
• Increases accuracy.
H ∞ filter • Accuracy high • Age and temperature affect accuracy.
Table 7 [68] • Strongness
Comparison of deep learning SoC estimate techniques. • Stability improves.
NAMHE • Working under • Running the programme increases
Algorithm Merits Demerits
[69] uncertain noise computational complexity and memory
LSTM [63] • Proven effectiveness in • A costly equipment is needed levels. utilization.
managing long-term to improve complex training • Improved stability
dependency. execution. and accuracy
• The calculation is less
demanding during the
online stage.
GRU [64] • Long-term sequential • Involves a significant quantity Table 9
dependencies are recorded. of training data and storage Overview of Li-ion battery SoC estimate techniques.
• LSTM gating mechanism space.
Method Merits Demerits
concerns resolved.
CNN–LSMT • Enhanced tracking • Complicated structure Conventional • Easy to set up • Not online-friendly
[66] accuracy method [28] • Low power use • Vulnerable to age and
• Strength in nonlinear • Simple to understand temperature changes
fitting. • Model accuracy is crucial
• Models are susceptible to
aging and temperature
that enhances the resilience and stability of a system in the presence of AF [35] • Effective filtering • High processing complexity.
model uncertainty and environmental disruptions. In article [57], a • High precision • Expensive
• insensitivity to initial • Big noise measurements are
sophisticated SMO was proposed as a means to address the nonlinearity SoC not possible
of battery dynamic characteristics. By incorporating an RC circuit into • High durability
the system, this was possible. This approach offers a means of regulating LA [50] • Decentralized models • Large training data
the discharge or charge rate during the conjunction time at a high level • High accuracy • Large memory units
• Inference based on rules Expensive processing
of sophistication. The constraints were extracted by employing a battery
•
• Capability for nonlinear • Time-consuming method
pulse and formulating the state equations based on the circuit model and mapping
terminal voltage [58]. NLO [61] • Stable • Low accuracy
• Strong tracking ability • Difficult to establish a
• Nonlinear observer (NLO) • Excellent nonlinear suitable gain matrix
processing potential • Not stable
Hybrid [67] • More efficient • Longer computing time
The SoC has been estimated using several observers, such as a linear • Durable • Complex calculations
observer and a nonlinear observer [59]. NLO-dependent SoC estimates • Accurate to a high
were initially introduced in article [60] using a similar first-order RC degree
circuit. Results from this model employing a driving cycle and a
discharge test were faster, more accurate, and less expensive than those LSTM networks can capture the spatial and temporal characteristics of
using extended KF and SMO [61]. The nonlinear observer SoC estima­
the battery data [66]. Several deep-learning SoC estimation techniques
tion technique comparison is shown in Table 6.
are compared and contrasted in Table 7.

3.5.5. Advanced SOC estimation techniques

• Hybrid methodologies

• Deep Learning Algorithm (DLA)

A hybrid algorithm combines features from many different algo­
rithms to achieve better performance and accuracy. Since it performs
The use of deep learning (DL) techniques has improved our capa­
extensive mathematical calculations, it needs a sizable memory unit.
bility to estimate SoC. Notable examples include recurrent neural net­ Contrarily, a hybrid approach achieves consistent and operational out­
works, convolutional neural networks, gated recurrent units, and long
comes while also reducing the BMS cost [67].
short-term memories. In terms of SoC estimates, the LSTM network ex­ The analysis of hybrid SoC estimation algorithms is mentioned in
cels because of its potent capacity for self-learning [62]. LSTM networks
Table 8, and the summary of the different SoC estimation methods for Li-
evaluate battery SoC using voltage, current, and temperature. In addi­ ion batteries is mentioned in Table 9.
tion, DNN encodes the battery's temperature-dependent behaviours into
DNN weights, enabling competitive estimation performance throughout • The nonlinear model of the Li-ion battery was applied to the H ∞
a wide temperature range [63]. Battery SoC at various temperatures is
filter and discrete-time KF. Compared to SMO-based estimating
estimated using GRU, and the efficiency of two commonly used lithium- models, this strategy increased accuracy by <1 %. EKF and EKF-
ion batteries is compared [64]. CNN is another promising deep-learning
assisted ANFIS RMSEs were compared. The hybrid technique
architecture. A convolutional neural network (CNN) and long short-term enhanced precision and accuracy while saving money [68].
memory network (LSTM) hybrid were presented in the article [65] to
• NAMHE was recommended for Li-ion batteries with unknown noise
mimic the intricate battery dynamics. The CNN was utilized to collect [69]. The suggested strategy had a lower SoC estimation error than
sophisticated spatial characteristics from the raw data. Both CNN and

11
S. Nyamathulla and C. Dhanamjayulu Journal of Energy Storage 86 (2024) 111179

Table 10 MHE, as shown by the simulation results [69]. Table 10 lists the
SoC monitoring issues, causes, and solutions. different SoC monitoring issues, causes, and solutions.
Ref. Issues Causes Solutions

[70] Temperature • Increased electrolyte • The best temperature 3.6. State of health (SoH)
development and range and battery
inconsistencies assist cycle charging rate The state of health (SoH) of a battery is the amount of usable
particle diffusion and are recognized. maximum capacity that is left over after cycling, which involves
mobility.
[71] Aging effect • Internal resistance and • Enhancing one
charging and discharging the battery many times [78]. Fig. 15 shows
capacitance restriction, battery different ways to figure out SoH. The estimation of SoH may be done
deterioration cause it. age, will provide an with the help of the following equations:
OCV curve model to
evaluate battery SoH. Qc
SoH(%) = × 100 (7)
[72] Cell unbalancing • Battery manufacturing • Passive and active cell Qn
and chemical properties balancing
may fluctuate when mechanisms were Rtermi − Rcu
discharging and proposed. SoH (%) = × 100 (8)
Rtermi − Rn
charging.
[73] Hysteresis effect • Impedance, • Li-ion cell hysteresis The symbol ‘Qc’ represents the current capacity of the battery,
electrochemical measurement whereas ‘Qn’ denotes the new battery capacity. After the battery life,
problems, concentration improves precision
‘Rtermi’ represents the ohmic internal resistance, ‘Rcu’ represents the
polarization, and energy despite its influence.
scattering in current state and ‘Rn’ represents the starting state. The SoH of a battery
development are the may be readily approximated by considering the battery's capacity
main causes. deterioration and internal resistance. There exist a range of techniques
[74] Battery • Modelling a battery is • ESC and higher-order
that have been devised to estimate the SoH of batteries. These meth­
modelling difficult due to its RC models were
complicated dynamics proposed
odologies are model-free, model-based, and data-driven. Electro­
and electrochemical chemical Impedance Spectroscopy (EIS) analysis is better than
environment. conventional approaches for model-free capacity and internal resistance
[75] Self-discharge • Lost lithium species and • An ECN model for SoC estimation. In contrast, model-based techniques use the equivalent cir­
SEI development cause estimation utilizing
cuit model and electrochemical model to estimate battery capacity and
self-discharge. prediction error
minimization was internal resistance [79].
developed.
[76] Charge and • The main indicator of • The current range of 3.6.1. SoH estimation experimental methods
discharge rate excessive discharge the Li-ion battery dis­
Experiments are usually done in labs since they require special
current in plastic Li-ion charged and charged.
batteries is phase
equipment and take time. They employ data and measures to assess
dispersion. battery aging. This section describes the primary experimental ap­
[77] Communication • An improved, even • Wireless technology proaches for battery SoH estimation.
charger is difficult to transfers data
create due to the non- between charger and
• Internal resistance measurement
uniform charging battery.
process.
The battery's internal resistance is an essential SoH indication that
determines its voltage drop when current is supplied. Many authors
studied internal resistance measurement methods. The current pulse is

Fig. 15. Classification of approaches for SoH estimation.

12
S. Nyamathulla and C. Dhanamjayulu Journal of Energy Storage 86 (2024) 111179

Table 11
SoH estimation experimental-based methods.
Methods Mode of Data calculation Merits Demerits Estimated error
operation

Internal resistance • Offline • Polynomial • Simple and direct method • It works in Offline mode only • <1 %
[80] exponential • Accurate • Slow
Internal • Offline • Exponential function • Improves the understanding of battery • Relies on battery chemistry data. • 2.1 %
impedance [81] deterioration causes.
• Accurate
Energy level [82] • Offline • Linear function and • Accurate • Not possible when the battery is • 9 % (21st
operations • Fast in response running (needs full charge). cycle)
• 1 % (after 8th
cycle)

Table 12
SoH estimation model-based methods.
Methods Mode of Data calculation Merits Demerits Estimated
operation errors

KF-base [78] • Online • Matrix operation • Accurate and Error bounds • Valid for nonlinear system • ±5 %
• Complex
• Computational effort is high.
• A high-performance controller is
required.
Least square-based [80] • Online • Linear equation • Precise • Relies in terms of accuracy on the • ±5 %
• Robust selected model
• Simple structure • A high-performance controller is
required.
Observer [81] • Offline • Polynomial • Accurate • A high-performance controller is • <1 %
exponential • Robust required.
• High cost
Electrochemical models • Offline • Exponential • Accurate • A high-performance controller is • 2.1 %
[82] function • Provides the degradation phenomena required.
knowledge of the battery. • Complex
• High computational effort

Fig. 16. Machine Learning algorithm functional block diagram.

13
S. Nyamathulla and C. Dhanamjayulu Journal of Energy Storage 86 (2024) 111179

Table 13
SoH estimation machine-learning methods.
Method Mode of operation Data calculation Merits Demerits Estimated
error

SVM • Offline (can be • Exponential function, • Accurate • Heavily dependent on the quality and • 2%
[78] adapted online) multiplication and accumulate. • Nonparametric quantity of the training data used.
• Robust • A high-performance controller is required.
FL [79] • Online • Exponential function, • Accuracy is high • Heavily dependent on the quality and • 1.4 %–9.2
multiplication and accumulate. • Robust quantity of the training data used. %
• Applicable for nonlinear • A high-performance controller is required.
systems
NN [80] • Online • Exponential function, vector and • Accurate • Heavily dependent on the quality and • <0.5 %
matrix operations • Requires less amount of data quantity of the training data used.
than Fuzzy Logic. • A high-performance controller is required.

the most typical approach based on Ohm's Law. After measuring the deterioration degrees [82]. The merits and demerits of the studied
battery's voltage drop for a particular current, it determines its internal experimental procedures are in Table 11.
resistance [80].
3.6.2. SoH estimation model-based methods
Vocv − Vbattery
Rbattery = (9) These models employ SoH indicators to characterize battery behav­
Ipulse
iour. Indicators are used to estimate battery statuses and performance.
where Rbattery is battery internal resistance, VOCV is open circuit voltage, The literature provides several strategies for identifying these signs. This
Vbattery is its voltage, and Ipulse is applied current. This approach is section outlines the primary model-based SoH estimation approaches.
extensively used in labs to accurately estimate battery internal resis­ The reviewed model-based methods are listed in Table 12 with their
tance in varied operating situations. Due to its time commitment, this merits and demerits.
procedure is better for stationery and laboratory work.
3.6.3. Battery machine learning SoH estimation methods
• Internal impedance measurement These approaches combine experimental and model-based method­
ologies. Training data, measurements, and models are used to estimate
Internal impedance is a battery's resistance and reactance. Age in­ battery SoH during learning. The literature has several machine-
creases a battery's intrinsic impedance, as proved. Hence, a battery SoH learning techniques for battery SoH prognostics. Fig. 16 shows a func­
indicator. EIS impedance measurement is the most commonly used tional block diagram of data driven techniques. Advancements in data
method to estimate the health condition of the battery [81]. Non- driven methods, such machine learning algorithms, have greatly
destructive approach evaluates electric system impedance by applying increased the accuracy and performance of SoC/SoH estimation. The
sinusoidal AC current and measuring response output voltage. Fre­ reviewed machine learning-based methods are listed in Table 13 with
quency determines impedance. Its key benefit is identifying battery their merits and demerits.
aging correctly.
3.6.4. Rapid EV battery SoH estimation methods
• Battery energy level The quick SoH estimate techniques in this study are electrical
parameters-based and material properties-based. There are two cate­
Energy storage capacity is a battery's capacity. As batteries age, this gories: pulse test technique (PTT) and electrochemical impedance
trait declines. The battery SoH can be best estimated by empirically spectrum (EIS) measurement, and ultrasonic inspection and a suggested
evaluating capacity declining over time. A lithium-ion battery was active acoustic emission (AE) detection technology [82]. Fig. 17 com­
charged and discharged till its end of life. The goal of this study is to pares four basic SOH estimate approaches from various angles for
determine battery charging capacity based on voltage for different practical application. These approaches can quickly estimate EV battery

Fig. 17. Comparison of SoH estimation methods in practical applications.

14
S. Nyamathulla and C. Dhanamjayulu Journal of Energy Storage 86 (2024) 111179

Fig. 18. Challenges and future scope of SoH estimation methods.

SOH online or offline. The problems and future work for improving SOH that the power demand is satisfied. Otherwise, it is “0”. Another over­
estimates for lithium-ion batteries in practical applications are pre­ head logical-variable equation characterized the SoF as follows:
sented in Fig. 18. Their uses and future scope are detailed below.
Pb − Pd
SoF = (10)
Pm − Pd
3.7. State of function (SoF) ‘Pb’ represents battery power, ‘Pd’ represents power demand, and
‘Pm’ represents maximum power (when SoC and SoH are “0” and the
The state of function (SoF), defined as the working state of a lithium- operating temperature is constant). State of charge SoC is always used to
ion battery pack under specific constraint conditions, is particularly represent the current status of a battery's charge, whereas SoH is used to
important. One of the most important responsibilities of the BMS is to show how the battery ages in comparison to a new one. Nonetheless,
evaluate the SoF. The SoF concept suited to a certain application's re­ when we need to characterize the battery pack function state under
quirements was presented. In some cases, none of the battery-pack status exact constraint circumstances, the state of function is the best option.
variables, such SoH, SoC, or voltage, can inform the system whether or The Fuzzy Logic Control Algorithm (FLCA) is the most recent
not the battery meets the requirements of the given application under approach for estimating SoF. The FLCA, an intellectual control method
real operating conditions [83]. used to estimate the SOF, has an essence. Some particular instances from
It represented “no” and “yes” with rational numbers “0” and “1”. The observed and simulated data are required to investigate the usefulness
SoF is “1” if the current-voltage exceeds the preset voltage, indicating

Fig. 19. RUL prediction methods.

15
S. Nyamathulla and C. Dhanamjayulu Journal of Energy Storage 86 (2024) 111179

Table 14
Comparison of RUL prediction methods.
Method Lithium-ion battery capacity Temperature Battery degradation details Accuracy
(Ah)

UKF [78] 0.9 25 ◦ C 100 cycles and 160 cycles MAPE 0.1611 and RMSE 0.01156 at 100 cycles.
MAPE 0.1857 and RMSE 0.02069 at 160 cycles.
GPM [79] 2 25 ◦ C 60 and 80 cycles RMSE 0.0158 at 60 cycles
RMSE 0.0130 at 80 cycles
UPF [80] 0.9 25 ◦ C 15 cycles and 32 cycles ER 4.08 %, RMSE 0.00423 at 15 cycles.
ER < 2.04 %, RMSE 0.00250 at 32 cycles
Bayesian [81] 2 4 ◦ C and 24 ◦ C 25 %, 50 %, 75 % and 90 % of actual RUL 74.7 %, 50 %, 25 % and 10.1 % at 25 %, 50 %, 75 % and 90 %.
life
RVM [82] 0.9 25 C
◦
104, 128, 146, 157, 170 and 183 Absolute prediction errors are 9, 7, 4, 0, 8, 1 with respect to cycles.
cycles.
WPME [83] 2 25 ◦ C 68 cycles Mean square error under 0.001
SVM [84] 2 4 ◦ C and 24 ◦ C 40, 60 and 80 cycles RMSE 0.1150, 0.0210 at 40th and 60th cycles respectively.
RMSE 0.0300 at 80th cycle.
Naïve Bayes 0.9 (type A) 25 ◦ C 15,32 cycles (Type A) SD 6,2 at 18,32 cycles for type A
[85] 1.1 (type B) 250 cycles (Type B) SD 35 at 250 cycles for type B
PSO [83] 0.9 25 C
◦
80, 95, 100, 126 cycles Mean prediction errors are 8, 9, 2, 3 cycles with respect to
inspection cycles.
LSTM [84] 2.7 25 ◦ C and 40 ◦ C 253 cycles and 354 cycles SD 28 and 11 at 253 and 354 cycles respectively.
AFS [85] 0.9 20 ◦ C–25 ◦ C 300 cycles Training RMSE 0.0836
Validation RMSE 0.1757

and significance. SoF is a battery's ability to complete a task. It describes results are more accurate and resilient. He et al. (2011) used Bayesian
how well the battery meets power demand. SoF is estimated using SoC, Monte Carlo (BMC) and Dempster-Shafer Theory (DST) to create a
SoH, and temperature [84]. The fuzzy logic control technique, adaptive double-index deterioration model that could forecast RUL. The battery's
characteristic maps, and similar circuit models may compute the SoF usable life determines optimal EV performance. Charged and discharged
[85]. KF and ANN algorithms are used in model-based SoF estimation for batteries degrade capacity, which can cause serious breakage, economic
accuracy. The fuzzy logic technique estimates the battery SoF using SoC, loss, and safety hazards. Therefore, EV technology must estimate battery
SoH, and C-rate parameters [86]. RUL to be safe, accurate, durable, and dependable. Continuous charging
and discharging leaves the battery at 70 % or 80 % of its initial capacity,
requiring replacement. Table 14 summarizes the comparison of various
3.8. Remaining useful life (RUL) RUL prediction methods.

RUL estimates the number of cycles until the battery's SoH hits 0 %. 3.9. Charge and discharge control
The optimal model for estimating RUL is not ubiquitous owing to data
unavailability, model complexity, and system limitations. This study The battery's measuring block digitizes analog measurements at each
divides RUL approaches into four categories: adaptive filter, intelligent, node for analysis of current, temperature, and voltage. To limit the
stochastic, and others as shown in Fig. 19 [84]. maximum charging and discharging currents, a capacity estimation
block is used. The cell balance block uses the results of the capacity
3.8.1. Adaptive filter technique estimation to regulate excessive discharging or charging [87]. Fig. 20
A non-linear time series prediction model is used with the UKF al­ demonstrates their method for controlling the charging and discharging
gorithm to provide a new technique for predicting RUL. Health prog­ of EVs using a systematic approach based on charging reliability in­
nosis can be improved by adaptive filtering. However, fluctuating dicators [88].
currents and temperatures might influence model accuracy. There are four types of charging strategies that may be characterized
as follows:
3.8.2. Intelligent techniques
Zhang et al. (2018) explored LSTM-based recurrent neural networks i. Constant Current (CC) Charging
for RUL prediction. Various lithium-ion cells were used to collect Constant-current charging entails sending a constant current to
experimental data at varying current rates and temperatures. Despite the battery during the charging process. The charging rate re­
offline training data, the model predicts RUL well. Wang et al. (2014) mains constant as the battery voltage increases. When the battery
suggested an SVM-based multi-step prediction model for reliable RUL voltage is low, this method is frequently utilized in the early
prediction. Characterizing the training dataset using working tempera­ stages of charging.
ture and energy efficiency. Results from trials indicate that the model ii. Constant Voltage (CV) Charging
accurately detects health features with few parameters. When charging at a constant voltage, the battery's voltage is
maintained as the charging current gradually decreases towards
3.8.3. Stochastic technique zero as the battery nears full charge. By controlling the voltage
To evaluate the deterioration of lithium-ion battery health, the sto­ between the battery terminals, this method protects the battery
chastic process is better characterized. The algorithm still has a problem from being overcharged.
in generating correct findings when taking into account the effect of iii. Constant Current/Constant Voltage (CC-CV) Charging
random current, time-varying temperatures, and self-discharge The CC-CV charging approach uses constant current and con­
characteristics. stant voltage. It quickly charges the battery with constant current
until a specified voltage is attained. After that, the charging
3.8.4. Others technique changes to a constant voltage to prevent overcharging and enable
According to Ng et al. (2014), a model for predicting RUL under the battery to fully charge.
varying current rates and ambient temperatures using Naive Bayes may
be used to lithium-ion batteries. When compared to SVM, the prediction

16
S. Nyamathulla and C. Dhanamjayulu Journal of Energy Storage 86 (2024) 111179

Fig. 20. Charge and discharge control block diagram of the BMS.

efficiency. MCC charging continuously injects multistage series

Table 15
current into the battery. MCC charges slower than CC-CV. MCC
Comparison of various charging strategies.
currently improves performance with fuzzy logic.
Criteria CC CV CC-CV MCC
charging charging charging charging
Charging strategies depend on battery type, chemistry, and perfor­
Efficiency [87] Moderate Low High Moderate mance goals. Charging efficiently, safely, and without overcharging
Lifespan [87] Low Medium High Medium
improves battery life. Table 15 lists the comparisons of the charging
Charging Speed High Medium Medium Low
[88] strategies.
Complexity [88] Small Small Medium More

3.10. Battery equalizer control

iv. Multistage Constant Current (MCC) Charging
Multistage constant-current charging charges the battery in The Battery Management System (BMS) is capable of safeguarding
stages with each current constant. Larger batteries, or battery the battery from irregularities resulting from both undercharging and
banks, employ this method. Managing heat can increase charging overcharging. This is achieved through the implementation of individ­
ual cell monitoring and charge equalization management. To optimize

Fig. 21. Charge equalization controllers for lithium-ion batteries.

17
S. Nyamathulla and C. Dhanamjayulu Journal of Energy Storage 86 (2024) 111179

Table 16
Comparison of operations various cell balancing methods.
Cell balancing Mechanism and operation Ref.
method

Passive balancing:
1. Resistor-Based • High-value resistors are connected in parallel with [90]
each cell.
• When a cell's voltage exceeds a certain threshold,
the resistor conducts and dissipates excess energy as
heat.
2. Zener Diode- • Zener diodes are connected in parallel with cells. [91]
Based • When a cell's voltage crosses a threshold, the Zener
diode conducts and allows current flow to balance
the voltage.
Active balancing:
1. Voltage-Based • Active electronics (DC-DC converters) monitor cell [92]
voltages.
• When an imbalance is detected, energy is
transferred from higher-voltage cells to lower-
voltage cells.
2. Charge • Active switches transfer charge directly between [90]
Transfer cells.
• Energy from higher-voltage cells is moved to lower- Fig. 22. Classification of temperature control methods.
voltage cells.
3. Capacitive • Capacitors are connected in parallel with cells. [93] performance, and when the temperature is too high, problems with
• When an imbalance occurs, energy is transferred
between cells through the capacitors.
thermal runaway and safety arise. Batteries lose capacity and function
4. Flyback • Specialized DC-DC converters store energy from [93] poorly when exposed to temperatures between 40 ◦ C and − 10 ◦ C.
Converter higher-voltage cells and release it to lower-voltage Therefore, the heat control of an EV's battery pack plays a vital role in
cells. real-time scenario [98].
5. Hybrid • Hybrid methods combine passive and active [94]
To maintain the battery at its ideal working temperature, a battery
Balancing balancing techniques to optimize efficiency and
cost-effectiveness. thermal management system (BTMS) must carry out essential functions
• For instance, a BMS might use passive balancing like heat dissipation through cooling, heat augmentation in the case of
most of the time and switch to active methods when low temperatures, and facilitating appropriate ventilation for exhaust
imbalances become significant. gases. The classification of BTMS may be based on the heat transfer
6. SOC Balancing • Similar to voltage-based or charge transfer [95]
balancing, but focused on equalizing the energy
medium, which includes air, liquid, and phase-change material (PCM)
content (SOC) of cells. [96]. An explosion ensues as a result of an imbalance in the electro­
chemical characteristics of a lithium-ion battery (LIB) caused by
elevated temperature. An explosion is triggered when the lithium-ion
battery (LIB) experiences a temperature rise, leading to the release of
Table 17
Performance comparison of various cell balancing methods. carbon monoxide (CO), acetylene (C2H2), and hydrogen sulfide (H2S)
from its internal chemical components [99]. Additionally, an internal
References Performance Passive balancing Active balancing
short circuit manifests inside the power circuit topology of the lithium-
indicator
ion battery (LIB). Fig. 22 illustrates several heat management tech­
[90] Efficiency Lower Higher
niques, and further elaboration may be found in reference [100].
[91] Speed of Balancing Slower Faster
[92] Complexity and Lower complexity Higher complexity Table 18 describes the temperature control techniques for BMS
Cost and cost and cost applications.
[93] Voltage Difference Limited to minor Suitable for large
Handling imbalances imbalances 3.12. Power/energy management control
[90] Applicability Smaller systems, Larger systems, high-
minor imbalances voltage cells
[94] Heat Generation Generates heat Minimizes heat Electric vehicle (EV) performance is dependent on several factors,
generation including energy storage, power management, and energy efficiency.
[95] Control Precision Limited Higher precision The energy storage control system of an electric vehicle has to be able to
handle high peak power during acceleration and deceleration if it is to
and sustain the consistent performance of the battery, it is imperative to effectively manage power and energy flow. There are typically two main
prioritise the equalization of voltage and charge across battery cells approaches used for regulating power and energy management (PEM)
[89]. The control of battery equalizer may be classified into two main [104]. By linking ESS, auxiliary ES, ICE, and generators together in a
categories: active charge equalization controllers and passive charge power transfer train, a low-level component control technique improves
equalization controllers, as seen in Fig. 21. Table 16 lists the comparison the performance and adaptability of the PEM [105].
of operations of various cell balancing methods, and Table 17 lists the
performance comparison of various cell balancing methods. 3.13. Fault diagnosis and assessment

Battery management systems for electric vehicles are required under

3.11. Operating temperature control a standard established by the International Electro-Technical Commis­
sion (IEC) in 1995 to include battery fault detection functionalities that
Longevity, energy conversion efficiency, and battery safety are just a can issue early alerts of battery aging and danger. It is common practice
few of the areas where temperature plays a major role [96]. Increasing to utilize analytical model-based, signal-processing, knowledge-based,
the battery's operating temperature, which degrades battery perfor­ and data-driven approaches in EV applications for problem diagnostics
mance, has been traced back to the quick charge-discharge cycle [97]. [106].
The operating temperature has an impact on the electrolyte's The model-based technique identifies the offending parameters by

18
S. Nyamathulla and C. Dhanamjayulu Journal of Energy Storage 86 (2024) 111179

Table 18
Temperature control techniques for BMS applications.
Temperature control method Description Applications References

Active Cooling:
Liquid Cooling Circulates cooling fluid through channels in a battery pack. EVs, PHEVs, grid storage [96]
Air Cooling Uses fans or blowers to direct airflow over the battery pack. EVs, consumer electronics, UPS [96]
Refrigeration Utilizes refrigeration systems to actively remove heat. High-performance EVs, data centres [97]

Passive cooling:
Heat Sinks Uses materials with high thermal conductivity to dissipate heat. EVs, laptops, electronic components [97]
Phase Change Materials (PCMs) Absorb/release heat by changing phase (solid to liquid). EVs, solar batteries, medical devices [98]
Thermal Insulation Provides thermal barriers to prevent external temperature changes. EVs, spacecraft, cold storage [98]

Heating:
Heating Elements Electric elements warm up the battery pack in cold conditions. EVs in cold climates, cold storage [98]
Exothermic Reactions Some chemistries release heat during operation for warming. EVs in cold climates, medical devices [98]

Cell Balancing for Temperature:

Passive Balancing Transfers energy between cells to equalize temperatures. EVs, consumer electronics [98]
Active Balancing Uses circuitry to redistribute energy for uniform temperatures. EVs, large-scale energy storage [98]

Temperature-Dependent Charging/Discharging:
Charging Rate Adjustment Adjusts charging rate based on battery temperature. EVs, grid storage, renewable energy [99]
Discharging Rate Adjustment Manages discharging rate based on temperature. EVs, grid stabilization, backup power [99]

Thermal Modelling and Prediction:

Thermal Models Predicts temperature changes under various conditions. EVs, energy management systems [99]
Predictive Algorithms Uses data to predict temperature changes and adapt control. EVs, smart grid, renewable integration [99]

Safety Measures:
Temperature Thresholds Sets limits; triggers actions like reducing power or cooling. EVs, stationary storage, aerospace [100]
Emergency Shutdown Shuts down battery if temperatures exceed critical levels. EVs, aerospace, critical systems [100]

User Interaction and Notifications:

Driver Alerts Notifies the driver if the battery temperature is unsafe. EVs, consumer electronics [101]

Adaptive Control:
Learning Algorithms Adapt strategies over time based on past scenarios. EVs, smart energy management [102]

Integrated Design:
System Integration Aligns thermal strategies with an overall vehicle and battery design. EVs, stationary storage, renewable energy [103]

comparing the residual signal to a predefined threshold. The problem is proposed approach for battery management is a data-driven and
that measurement and process noise might muddy the diagnostic results. customized strategy that leverages big data and cloud computing, as
The signal processing-based technique relies heavily on time-domain seen in Fig. 24.
analysis to acquire the test data necessary for fault analysis [107]. Unpredictably, the several currently promoted BMS each indepen­
Further, a knowledge-based approach to defect diagnostics employs dently perform the elemental abilities. Table 20 compares and contrasts
machine learning and expert systems, both of which may be used to various BMS products, and Table 21 compares the performance studies
estimate a battery's remaining useful life. In Fig. 23, a flowchart de­ among BMS components.
tailing their suggested method for problem identification in a lithium-
ion battery system [108]. The BMS runs a battery parameter estima­ 4. Applications of battery management system (BMS)
tion suite of tests in accordance with the recommendations made in
Table 19 [15]. Battery management systems (BMS) play a crucial role in the man­
agement of battery performance, safety, and longevity. Rechargeable
batteries find widespread use in several applications. Battery manage­
3.14. Communication and networking ment systems (BMS) have emerged as crucial components in several
domains due to their ability to efficiently monitor and control the per­
The communication system employs physical transmission mediums formance of batteries. The following are notable applications where
such as cables or data lines [109]. A fundamental BMS typically com­ BMS plays a critical role. Fig. 25 presents how BMS is grid-integrated
prises essential components such as a microcontroller, debugger, with different possible sources for power electronics converter appli­
Controller Area Network (CAN) bus, and host computer. The AS8505, cations and similarly, the PV-Battery integration block diagram for the
which is an integrated circuit designed for monitoring battery condition, grid is presented in Fig. 26.
establishes communication with the microcontroller by utilizing I/O The operational life of the battery in a photovoltaic (PV)-battery-
lines and a Controller Area Network (CAN) bus. This communication integrated system is significantly reduced, and its performance is
enables the regulation of cell data and facilitates the balancing process significantly affected due to repeated charging and discharging cycles.
[110]. ZigBee, Wi-Fi, GSM, Bluetooth, GPRS, and GPS have been iden­ This study presents a suggested intelligent power control technique for a
tified as potential technologies for battery monitoring [111]. The

19
S. Nyamathulla and C. Dhanamjayulu Journal of Energy Storage 86 (2024) 111179

Fig. 23. Fault diagnosis process of a battery system.

islanding techniques, to extend the battery's lifespan. An advanced BMS

Table 19
can handle multiple operations; hence, it was determined that the most
BMS parameter testing guidelines.
effective advancement of EV technology is shown in Fig. 27 for BMS-EV
S.No. Parameter of testing Standards and guidelines integration [113].
1. Cell balancing IEE 1679.1 In general, the applications of battery management systems span
2. Thermal management IEE 1679.1 across several industries and technologies, as shown in Fig. 28, with the
3. Over-discharge UL1973, NAVSEA S9310
primary objective of improving battery performance, ensuring safety,
4. Temperature range IEE 1679.1
5. Control of current IEC62619, NAVSEA S9310
and prolonging battery lifespan in different environments [114].
6. Cell operating range IEC62619, IEE 1679.1
7. Thermal runaway fault IEE 1679.1 5. BMS challenges and recommendations
8. Overheating control IEC62619
9. Heating and cooling control IEE 1679.1
10. Control of voltage IEC62619, UL1973, UL9540
Battery management systems (BMS) monitor and control battery
performance in electric vehicles, renewable energy systems, and
portable electronics. The recommendations for various open challenges
standalone PV battery system, aiming to enhance the battery's depend­ are mentioned in Fig. 29, and finally, a few add-on constraints are
ability throughout its operating lifespan. The control technique being mentioned in Fig. 30.
presented operates in two distinct regulatory modes, namely maximum
power point tracking (MPPT) mode and battery management system 5.1. BMS challenges
(BMS) mode. The unique controller employs an MPPT system to effec­
tively monitor and optimize the power output of the solar cells, maxi­ • Battery Storage Technology: Fast charging can lead to high current
mizing their energy harvesting potential across various air conditions. flow, which can cause health degradation and ultimately shorten
Instead, a backpropagation neural network (BPNN) algorithm has been battery life, impacting overall performance. Small batteries can be
used in the battery management system (BMS) mode to create a way to combined in series and parallel configurations to solve this issue.
estimate SoC [112]. This technique facilitates the effective management • Battery Balancing and Temperature Issues: Passive balancing and
of battery storage operations, including charging, discharging, and thermal management are important tasks in battery management

Fig. 24. Superior BMS design utilizing 5G for EVs.

20
S. Nyamathulla and C. Dhanamjayulu Journal of Energy Storage 86 (2024) 111179

Table 20 proper operational conditions, particularly for BMS peripheral con­

Comparative studies among BMS products. trol units.
Parameters Maxim DS2726 TI BQ78PL114 OZ890 [18,85] • Aging and Memory Effect: There are three main causes of battery
[18,85] [18,85] deterioration: internal resistance, capacitance loss, and overheating.
Cell constraints Voltage and Voltage, Current and In order to deal with memory's effects and possible imbalances, a
measured current temperature, voltage model that takes cell aging variables into account is required.
impedance, and • Battery Recycling and Reuse: Recycling batteries helps the environ­
current ment and expands recycling options. A BMS is essential for safe
Status estimation Not available SOC SOC
Protection Short circuit • Three power Short circuit
battery functioning throughout its lifespan.
current, over- field-effect current and over- • Battery Disposal Issues: Used batteries are considered hazardous
current, and transistors current trash and must be disposed of in a certain way to avoid potential
over-voltage • One secondary fires, chemical spills, and other catastrophes. Disposal is costly due to
safety output
transportation, treatment, and final disposal fees, as well as potential
fuse
Pack parameters Not available Not available Temperature regulatory issues.
measured
Communication Not available Power LAN and SM CAN 5.2. Recommendations
Bus
Non-dissipative Not available Inductive charge Not available
equalization Recommendations and highlights are provided for future research
Data collection Not available PC-based GUI only EEPROM and development scopes in the sustainable electric vehicle (EV) domain
Dissipative Charge shifting Not available External based on identified concerns and obstacles. Future advancements in
equalization resistance stable
lithium-ion batteries (LIBs) production and BMS technology have been
achieved in the following manner:
systems. They help to address weakened cells and ensure efficient
energy dissipation by using external resistors. • Enhancing Safety and Reliability: Use interlock circuits and insu­
• Battery Environmental Issues: EVs help reduce emissions, but the lation monitoring to improve battery safety and dependability,
negative impact of non-renewable resources arises when batteries following ISO 26262 PCB-to-connector lengths.
are not properly recycled or reach the end of their lifespan. • Algorithm Hybridization and Advanced Prognostic: Hybrid algo­
• Real-Time SOC and SOH Estimation: Present methods for estimating rithms need enhanced prognostics and health management to
battery SoC and SOH in practical situations are challenging due to monitor temperature, charge/discharge rate, depth of discharge
low-cost BMS limitations. Owners must choose between replacing (DOD), vibrations, system safety, dependability and longevity.
batteries, increasing their financial burden, or waiting for storms. • Advanced Thermal Management: Sensor-less sensing, electro­
• Battery Models: BMS batteries require precise testing in various en­ chemical impedance spectroscopy, and innovative internal and
vironments due to physical and data-driven techniques. external temperature management technologies can be implemented
• RUL Prediction Issues: Due to modelling constraints, system noise, to enhance battery temperature control.
and sensor quality, RUL cannot be accurately predicted using BMS. • Life Cycle Assessment and Aging Effect: LIBs should avoid rare,
Only suitable for specific battery types and unreliable health costly, poisonous, and hard-to-recycle materials because of their
indicators. complicated aging dynamics.
• Safe and Efficient Operation: Batteries face safety concerns due to • Enhancing LIBs Capacity and Fast Charging: Hidden variables affect
changing factors impacting reliability and stability and maintaining LIB capacity; anomaly detection and different driving methods in­
crease efficiency. Fast-charging and overcharging electric vehicles
require advanced battery management technologies.

Table 21
Performances studies among various BMS components.
BMS components Functions Algorithm/methods Target Outcomes

Monitoring [109] • Cell monitoring • Voltage and current measurement • Current, voltage, and • BMS performance is improved.
• Voltage and current • Voltage divider technique temperature monitoring
monitoring
Communication [109] • Monitor and protect • Microcontroller unit, Debugger, • To control the battery data and • Monitoring of the battery status using
• the battery and CAN bus. monitoring wired or wireless approach.
Fault diagnosis and • Battery protection • Knowledge-based methods, and • Protection of battery • Increased battery lifetime
protection [106–108] • Unbalance data-driven methods • Protection from temperature obtained
• Undercharge • Detection of faulty locations.
• Overcharge
• Overcurrent
State estimation [110] • SOC • Ampere-hour (Ah) • To minimize • Enhanced Vehicle
• SOH • OCV estimation error • Performance
• SOF • FLA • Accurate estimation
• To reduce the
computational cost
Control operation [101] • Charge and discharge • CC • To enhance • Controlled operation can be obtained.
control • CV performance
• Power/energy • CC-CV
management • MCC • To improve
• Temperature control • PEM durability

• To control energy flow

Data acquisition [111] • Data processing • CAN bus • Communication • BMS performance is improved.
• SCI

21
S. Nyamathulla and C. Dhanamjayulu Journal of Energy Storage 86 (2024) 111179

Fig. 25. BMS Grid-Integrated with different possible sources for Power Electronics converter applications.

Fig. 26. PV-Battery Integration to grid.

• Reuse and Recycling: Battery reuse research is essential to saving 6. Conclusions and future scope
energy, ensuring Earth's lithium-ion battery supply, and recovering
power. Non-profits and governments should work together to create This review presented a comprehensive analysis of several battery
cost-effective, eco-friendly approaches. storage technologies, materials, properties, and performances. This
• Wireless and Universal BMS: A universal open-source battery man­ article also provided a detailed explanation of the advanced techniques,
agement system requires adaptive methods, manufacturer collabo­ algorithms, controllers, and optimization methodologies utilized in
ration, and upgraded hardware and software to improve efficiency electric vehicles (EVs). This research work comprehensively investi­
and lower costs. gated the categorization of traditional and sophisticated SoC and SoH
• Battery state estimation: Advanced technologies and methodologies estimation methodologies as well as the associated advantages and
for EV state estimation necessitate the inclusion of three key pa­ drawbacks. This review article acknowledged that traditional methods
rameters: SoC, SoH, and RUL. are simple and easy to implement. The learning algorithm (LA) improves
the performance of a nonlinear dynamic modelling setup. The NLO

22
S. Nyamathulla and C. Dhanamjayulu Journal of Energy Storage 86 (2024) 111179

Fig. 27. BMS-EV Integration.

Fig. 28. Different applications of BMS.

technique is faster, cheaper to compute, more accurate, and more reli­ demonstrated the difficulty of EV adoption by addressing existing
able than its predecessors. This review analysis demonstrated that BMS problems and developing superior BMSs. This comprehensive resource
continues to confront several challenges despite the use of a wide range offers valuable insights for engineers, researchers, and EV manufac­
of appropriate algorithms and sophisticated approaches and models. turers and presents detailed analyses, applications, challenges, and
Battery management systems (BMSs) are discussed in depth, as are their recommendations relevant to the field.
applications in EVs and renewable energy storage systems. This review
covered topics ranging from voltage and current monitoring to the
6.1. Future scope
estimation of charge and discharge, protection, equalization of cells,
thermal management, and actuation of stored battery data. This review
After surveying the literature on BMS and SoC estimation methods in

23
S. Nyamathulla and C. Dhanamjayulu Journal of Energy Storage 86 (2024) 111179

Fig. 29. An overview of BMS Challenges

Fig. 30. Basic challenges and recommendations to BMS.

depth, this article proposes various directions for further study, as rep­ sustainable growth.
resented in Fig. 31.
There will be substantial growth in the battery and EV sectors due to CRediT authorship contribution statement
further research on BMSs employing cutting-edge intelligent algorithms
to enhance battery performance and longevity and guarantee EVs' safe Shaik Nyamathulla: Conceptualization, Data curation, Formal
and dependable operation. Additionally, the battery-related industry analysis, Methodology, Visualization, Writing – original draft, Writing –
may support long-term development objectives like pollution reduction, review & editing. C. Dhanamjayulu: Investigation, Project adminis­
clean energy, economic expansion, job creation, and the growth of the tration, Resources, Supervision, Validation, Writing – review & editing.
EV sector. Therefore, further innovations are needed to increase EV
performance in terms of developing precise battery monitoring and
control strategies, collaborating across borders, and fostering

24
S. Nyamathulla and C. Dhanamjayulu Journal of Energy Storage 86 (2024) 111179

Fig. 31. Future trends in advanced BMS for EV applications.

Declaration of competing interest [5] W. Barbosa, T. Prado, C. Batista, J.C. Câmara, R. Cerqueira, R. Coelho,
L. Guarieiro, Electric vehicles: bibliometric analysis of the current state of the art
and perspectives, Energies 15 (2022) 395.
The authors declare that they have no known competing financial [6] A.A. Habib, M.K. Hasan, M. Mahmud, S. Motakabber, M.I. Ibrahimya, S. Islam,
interests or personal relationships that could have appeared to influence A review: energy storage system and balancing circuits for electric vehicle
the work reported in this paper. application, IET Power Electron. 14 (2021) 1–13.
[7] H. Barkh, A. Yu, D. Friend, P. Shani, Q. Tu, O. Swei, Vehicle fleet electrification
and its effects on the global warming potential of highway pavements in the
Data availability United States, Resour. Conserv. Recycl. 185 (2022) 106440.
[8] S.S. Ravi, M. Aziz, Utilization of electric vehicles for vehicle-to-grid services:
progress and perspectives, Energies 15 (2022) 589.
No data was used for the research described in the article. [9] Nyamathulla Shaik, D. Chittathuru, A review of multilevel inverter topologies for
grid-connected sustainable solar photovoltaic systems, Sustainability 15 (2023)
13376.
Acknowledgments [10] R. Debnath, R. Bardhan, D.M. Reiner, J.R. Miller, Political, economic, social,
technological, legal and environmental dimensions of electric vehicle adoption in
The authors further thank the Vellore Institute of Technology in the United States: a social-media interaction analysis, Renew. Sustain. Energy
Rev. 152 (2021) 111707.
Vellore, India, for their assistance. [11] M.S.H. Lipu, et al., Battery management, key technologies, methods, issues, and
future trends of electric vehicles: a pathway toward achieving sustainable
development goals, Batteries 8 (9) (Sep. 2022) 119, https://doi.org/10.3390/
Funding statement
batteries8090119.
[12] M.S. Hossain Lipu, M.A. Hannan, T.F. Karim, A. Hussain, M.H.M. Saad, A. Ayob,
This research received no external funding. M.S. Miah, T.M. Indra Mahlia, Intelligent algorithms and control strategies for
battery management system in EVs: progress, challenges and future outlook,
J. Clean. Prod. 292 (2021) 126044.
References [13] M.S.H. Lipu, M. Faisal, S. Ansari, M.A. Hannan, T.F. Karim, A. Ayob, A. Hussain,
M.S. Miah, M.H.M. Saad, Review of electric vehicle converter configurations,
[1] S. Koohi-Fayegh, M.A. Rosen, A review of energy storage types, applications and control schemes and optimizations: challenges and suggestions, Electronics 10
recent developments, Journal of Energy Storage 27 (2020) 101047. (2021) 477.
[2] M.K. Hasan, M. Mahmud, A.A. Habib, S. Motakabber, S. Islam, Review of electric [14] A. Asekomeh, O. Gershon, S.I. Azubuike, Optimally clocking the low carbon
vehicle energy storage and management system: standards, issues, and energy mile to achieve the sustainable development goals: evidence from
challenges, J. Energy Storage 41 (2021) 102940. Dundee’s electric vehicle strategy, Energies 14 (2021) 842.
[3] K. Gholami, A. Azizivahed, A. Arefi, Risk-oriented energy management strategy [15] Vedhanayaki Selvaraj, Indragandhi Vairavasundaram, A comprehensive review of
for electric vehicle fleets in hybrid AC-DC microgrids, J. Energy Storage 50 state of charge estimation in lithium-ion batteries used in electric vehicles,
(2022) 104258. Journal of Energy Storage 72 (Part E) (2023) 108777.
[4] A.K.M.A. Habib, M.K. Hasan, G.F. Issa, D. Singh, S. Islam, T.M. Ghazal, Lithium- [16] M.A. Hannan, D.N.T. How, M.S.H. Lipu, M. Mansor, P.J. Ker, Z.Y. Dong, K.S.
ion battery management system for electric vehicles: constraints, challenges, and M. Sahari, S.K. Tiong, K.M. Muttaqi, T.M.I. Mahlia, et al., Deep learning approach
recommendations, Batteries 9 (3) (Feb. 2023) 152, https://doi.org/10.3390/ towards accurate state of charge estimation for lithium-ion batteries using self-
batteries9030152. supervised transformer model, Sci. Rep. 11 (2021) 19541.

25
S. Nyamathulla and C. Dhanamjayulu Journal of Energy Storage 86 (2024) 111179

[17] D.N.T. How, M.A. Hannan, M.S. Hossain Lipu, P.J. Ker, State of charge estimation [44] N. Wassiliadis, J. Adermann, A. Frericks, M. Pak, C. Reiter, B. Lohmann,
for lithium-ion batteries using model-based and data-driven methods: a review, M. Lienkamp, Revisiting the dual extended Kalman filter for battery state-of-
IEEE Access 7 (2019) 136116–136136. charge and state-of-health estimation: a use-case life cycle analysis, J. Energy
[18] T. Girijaprasanna, C. Dhanamjayulu, A review on different state of battery charge Storage 19 (2018) 73–87.
estimation techniques and management systems for EV applications, Electronics [45] V. Sangwan, R. Kumar, A.K. Rathore, State-of-charge estimation of Li-ion battery
11 (11) (Jun. 2022) 1795. at different temperatures using particle filter, J. Eng. 2019 (2019) 5320–5324.
[19] M.O. Qays, et al., An intelligent controlling method for battery lifetime increment [46] S. Schwunk, N. Armbruster, S. Straub, J. Kehl, M. Vetter, Particle filter for state of
using state of charge estimation in PV-battery hybrid system, Appl. Sci. 10 (24) charge and state of health estimation for lithium–iron phosphate batteries,
(Dec. 2020) 8799. J. Power Sources 239 (2013) 705–710.
[20] Yijie Zhang, Tao Ma, Hongxing Yang, Grid-connected photovoltaic battery [47] Y. He, X. Liu, C. Zhang, Z. Chen, A new model for State-of-Charge (SOC)
systems: a comprehensive review and perspectives, Appl. Energy 328 (2022) estimation for high-power Li-ion batteries, Appl. Energy 101 (2013) 808–814.
120182. [48] M. Charkhgard, M.H. Zarif, Design of adaptive H ¥ filter for implementing on
[21] M.A. Hannan, M.M. Hoque, A. Hussain, Y. Yusof, P.J. Ker, State-of-the-art and state-of-charge estimation based on battery state-of-charge-varying modelling,
energy management system of lithium-ion batteries in electric vehicle IET Power Electron. 8 (2015) 1825–1833.
applications: issues and recommendations, IEEE Access 6 (2018) 19362–19378. [49] M.-Y. Kim, C.-H. Kim, J.-H. Kim, G.-W. Moon, A chain structure of switched
[22] R.P. Narasipuram, S. Mopidevi, A technological overview & design considerations capacitor for improved cell balancing speed of lithium-ion batteries, IEEE Trans
for developing electric vehicle charging stations, J. Energy Storage 43 (2021) Ind Electron 61 (2014) 3989–3999.
103225. [50] M.H.A. Jamlouie, Accuracy improvement of SOC estimation in lithium-ion
[23] Z. Wang, G. Feng, D. Zhen, F. Gu, A. Ball, A review on online state of charge and batteries by ANFIS vs. ANN modeling of nonlinear cell characteristics, J. Energy
state of health estimation for lithium-ion batteries in electric vehicles, Energy Storage 6 (2021) 95–104.
Rep. 7 (2021) 5141–5161. [51] P. Singh, R. Vinjamuri, X. Wang, D. Reisner, Design and implementation of a
[24] M. Hannan, S. Wali, P. Ker, M.A. Rahman, M. Mansor, V. Ramachandaramurthy, fuzzy logic-based state-of-charge meter for Li-ion batteries used in portable
K. Muttaqi, T. Mahlia, Z. Dong, Battery energy-storage system: a review of defibrillators, J. Power Sources 162 (2006) 829–836.
technologies, optimization objectives, constraints, approaches, and outstanding [52] S. Malkhandi, Fuzzy logic-based learning system and estimation of state-of-charge
issues, J. Energy Storage 42 (2021) 103023. of lead-acid battery, Eng. Appl. Artif. Intel. 19 (2006) 479–485.
[25] I. Snihir, W. Rey, E. Verbitskiy, A. Belfadhel-Ayeb, P.H. Notten, Battery open- [53] I.-H. Li, W.-Y. Wang, S.-F. Su, Y.-S. Lee, A merged fuzzy neural network and its
circuit voltage estimation by a method of statistical analysis, J. Power Sources applications in battery state-of-charge estimation, IEEE Trans. Energy Convers. 22
159 (2006) 1484–1487. (2007) 697–708.
[26] Y. Zhang, W. Song, S. Lin, Z. Feng, A novel model of the initial state of charge [54] J. Xu, B. Cao, Z. Chen, Z. Zou, An online state of charge estimation method with
estimation for LiFePO4 batteries, J. Power Sources 248 (2014) 1028–1033. reduced prior battery testing information, Int. J. Electr. Power Energy Syst. 63
[27] F. Leng, C.M. Tan, R. Yazami, M.D. Le, A practical framework of electrical based (2014) 178–184.
online state-of-charge estimation of lithium ion batteries, J. Power Sources 255 [55] J.C.Á. Antón, P.J.G. Nieto, F.J. de Cos Juez, F.S. Lasheras, M.G. Vega, M.N.
(2014) 423–430. R. Gutiérrez, Battery state-of-charge estimator using the SVM technique, App.
[28] M. Li, Li-ion dynamics and state of charge estimation, Renew. Energy 100 (2017) Math. Model. 37 (2013) 6244–6253.
44–52. [56] X. Wu, L. Mi, W. Tan, J.L. Qin, M. Na Zhao, State of charge (SOC) estimation of
[29] J. Meng, D.-I. Stroe, M. Ricco, G. Luo, R. Teodorescu, A simplified model-based Ni-MH battery based on least square support vector machines, Adv. Mater. Res.
state-of-charge estimation approach for lithium-ion battery with dynamic linear 211–212 (2011) 1204–1209.
model, IEEE Trans Ind Electron 66 (2019) 7717–7727. [57] I.-S. Kim, Nonlinear state of charge estimator for hybrid electric vehicle battery,
[30] S. Cho, H. Jeong, C. Han, S. Jin, J.H. Lim, J. Oh, State-of-charge estimation for IEEE Trans. Power Electron. 23 (2008) 2027–2034.
lithium-ion batteries under various opeating conditions using an equivalent [58] N. Boizot, E. Busvelle, J.-P. Gauthier, An adaptive high-gain observer for
circuit model, Comput. Chem. Eng. 41 (2012) 1–9. nonlinear systems, Automatica 46 (2010) 1483–1488.
[31] A. Rahman, S. Anwar, A. Izadian, Electrochemical model parameter identification [59] J. Xu, C. Mi, B. Cao, J. Deng, Z. Chen, S. Li, The state of charge estimation of
of a lithium-ion battery using particle swarm optimization method, J. Power lithium-ion batteries based on a proportional integral observer, IEEE Trans Veh
Sources 307 (2016) 86–97. Technol 63 (2014) 1614–1621.
[32] K.D. Stetzel, L.L. Aldrich, M. Trimboli, G.L. Plett, Electrochemical state and [60] F. Xu, Y. Wang, X. Luo, Soft sensor for inputs and parameters using nonlinear
internal variables estimation using a reduced-order physics-based model of a singular state observer in chemical processes, Chin. J. Chem. Eng. 21 (2013)
lithium-ion cell and an extended Kalman filter, J. Power Sources 278 (2015) 1038–1047.
490–505. [61] I.-S. Kim, The novel state of charge estimation method for lithium battery using
[33] T. Bruen, J. Marco, Modelling and experimental evaluation of parallel connected sliding mode observer, J. Power Sources 163 (2006) 584–590.
lithium ion cells for an electric vehicle battery system, J. Power Sources 310 [62] B. Xia, C. Chen, Y. Tian, W. Sun, Z. Xu, W. Zheng, A novel method for state of
(2016) 91–101. charge estimation of lithium-ion batteries using a nonlinear observer, J. Power
[34] X. Lai, W. Gao, Y. Zheng, M. Ouyang, J. Li, X. Han, L. Zhou, A comparative study Sources 270 (2014) 359–366.
of global optimization methods for parameter identification of different [63] E. Chemali, P.J. Kollmeyer, M. Preindl, R. Ahmed, A. Emadi, P. Kollmeyer, Long
equivalent circuit models for Li-ion batteries, Electrochim. Acta 295 (2019) short-term memory networks for accurate state-of-charge estimation of Li-ion
1057–1066. batteries, IEEE Trans Ind Electron 65 (2018) 6730–6739.
[35] M.-K. Tran, A. Mevawala, S. Panchal, K. Raahemifar, M. Fowler, R. Fraser, Effect [64] D.N. How, M. Hannan, M.H. Lipu, K.S. Sahari, P.J. Ker, K.M. Muttaqi, State-of-
of integrating the hysteresis component to the equivalent circuit model of charge estimation of li-ion battery in electric vehicles: a deep neural network
Lithium-ion battery for dynamic and non-dynamic applications, J. Energy Storage approach, IEEE Trans. Ind. Appl. 56 (2020) 5565–5574.
32 (2020) 101785. [65] F. Yang, W. Li, C. Li, Q. Miao, State-of-charge estimation of lithium-ion batteries
[36] T.O. Ting, K.L. Man, E. Lim, M. Leach, Tuning of Kalman filter parameters via based on gated recurrent neural network, Energy 175 (2019) 66–75.
genetic algorithm for state-of-charge estimation in battery management system, [66] X. Song, F. Yang, D. Wang, K.-L. Tsui, Combined CNN-LSTM network for state-of-
Sci. World J. 2014 (2014) 176052. charge estimation of lithium-ion batteries, IEEE Access 7 (2019) 88894–88902.
[37] Z. Chen, Y. Fu, C. Mi, State of charge estimation of lithium-ion batteries in electric [67] A. Alfi, M. Charkhgard, M.H. Zarif, Hybrid state of charge estimation for lithium-
drive vehicles using extended Kalman filtering, IEEE Trans Veh Technol 62 ion batteries: design and implementation, IET Power Electron. 7 (2014)
(2013) 1020–1030. 2758–2764.
[38] Y. Fang, Q. Zhang, H. Zhang, W. Xu, L. Wang, X. Shen, F. Yun, Y. Cui, L. Wang, [68] S. Afshar, K. Morris, A. Khajepour, State-of-charge estimation using an EKF-based
X. Zhang, State-of-charge estimation technique for lithium-ion batteries by means adaptive observer, IEEE Trans Control Syst Technol 27 (2019) 1907–1923.
of second-order extended Kalman filter and equivalent circuit model: great [69] Z. Zhang, B. Xue, J. Fan, Noise adaptive moving horizon estimation for state-of-
temperature robustness state-of-charge estimation, IET Power Electron. 14 (2021) charge estimation of Li-ion battery, IEEE Access 9 (2021) 5250–5259.
1515–1528. [70] C. Zhao, B. Zhang, Y. Zheng, S. Huang, T. Yan, X. Liu, Hybrid battery thermal
[39] X. Liu, K. Li, J. Wu, Y. He, X. Liu, An extended Kalman filter based data-driven management system in electrical vehicles: a review, Energies 13 (2020) 6257.
method for state of charge estimation of Li-ion batteries, J. Energy Storage 40 [71] M.P. Bonkile, V. Ramadesigan, Physics-based models in PV-battery hybrid power
(2021) 102655. systems: thermal management and degradation analysis, J. Energy Storage 31
[40] P. Nian, Z. Shuzhi, Z. Xiongwen, Co-estimation for capacity and state of charge for (2020) 101458.
lithium-ion batteries using improved adaptive extended Kalman filter, J. Energy [72] C. Pastor-Fernández, T. Bruen, W.D. Widanage, M.A. Gama-Valdez, J. Marco,
Storage 40 (2021) 102559. A study of cell-to-cell interactions and degradation in parallel strings:
[41] R. Xiong, H. He, F. Sun, K. Zhao, Evaluation on state of charge estimation of implications for the battery management system, J. Power Sources 329 (2016)
batteries with adaptive extended Kalman filter by experiment approach, IEEE 574–585.
Trans Veh Technol 62 (2013) 108–117. [73] Y. Li, H. Guo, F. Qi, Z. Guo, M. Li, Comparative study of the influence of open
[42] S. Wang, C. Fernandez, L. Shang, Z. Li, J. Li, Online state of charge estimation for circuit voltage tests on state of charge online estimation for lithium-ion batteries,
the aerial lithium-ion battery packs based on the improved extended Kalman IEEE Access 8 (2020) 17535–17547.
filter method, J. Energy Storage 9 (2017) 69–83. [74] H. He, X. Zhang, R. Xiong, Y. Xu, H. Guo, Online model-based estimation of state-
[43] I.B. Espedal, A. Jinasena, O.S. Burheim, J.J. Lamb, Current trends for state-of- of-charge and open-circuit voltage of lithium-ion batteries in electric vehicles,
charge (SoC) estimation in lithium-ion battery electric vehicles, Energies 14 Energy 39 (2012) 310–318.
(2021) 3284.

26
S. Nyamathulla and C. Dhanamjayulu Journal of Energy Storage 86 (2024) 111179

[75] W. He, N. Williard, C. Chen, M. Pecht, State of charge estimation for electric [95] F. Ju, W. Deng, J. Li, Performance evaluation of modularized global equalization
vehicle batteries using unscented kalman filtering, Microelectron. Reliab. 53 system for lithium-ion battery packs, IEEE Trans Autom Sci Eng 13 (2016)
(2013) 840–847. 986–996.
[76] Y. Wang, B. Liu, P. Han, C. Hao, S. Li, Z. You, M. Wang, Optimization of an air- [96] A. Li, A.C.Y. Yuen, W. Wang, T.B.Y. Chen, C.S. Lai, W. Yang, W. Wu, Q.N. Chan,
based thermal management system for lithium-ion battery packs, J. Energy S. Kook, G.H. Yeoh, Integration of computational fluid dynamics and artificial
Storage 44 (2021) 103314. neural network for optimization design of battery thermal management system,
[77] L. Lavigne, J. Sabatier, J.M. Francisco, F. Guillemard, A. Noury, Lithium-ion Open Batteries 8 (2022) 69.
Circuit Voltage (OCV) curve modelling and its ageing adjustment, J. Power [97] A.A.H. Akinlabi, D. Solyali, Configuration, design, and optimization of air-cooled
Sources 324 (2016) 694–703. battery thermal management system for electric vehicles: a review, Renew.
[78] Gökhan Yüksek, Alkan Alkaya, A novel state of health estimation approach based Sustain. Energy Rev. 125 (2020) 109815.
on polynomial model for lithium-ion batteries, Int. J. Electrochem. Sci. 18 (5) [98] G. Xia, L. Cao, G. Bi, A review on battery thermal management in electric vehicle
(2023) 100111. application, J. Power Sources 367 (2017) 90–105.
[79] H. Lin, L. Kang, D. Xie, J. Linghu, J. Li, Online state-of-health estimation of [99] J. Kim, J. Oh, H. Lee, Review on battery thermal management system for electric
lithium-ion battery based on incremental capacity curve and BP neural network, vehicles, Appl. Therm. Eng. 149 (2019) 192–212.
Batteries 8 (2022) 29. [100] D. Hu, Y. Wang, J. Li, Q. Yang, J. Wang, Investigation of optimal operating
[80] M.S.H. Lipu, M.A. Hannan, A. Hussain, M.M. Hoque, P.J. Ker, M.H.M. Saad, temperature for the PEMFC and its tracking control for energy saving in vehicle
A. Ayob, A review of state of health and remaining useful life estimation methods applications, Energ. Conver. Manage. 249 (2021) 114842.
for lithium-ion battery in EVs: challenges and recommendations, J. Clean. Prod. [101] A. Marongiu, F.G.W. Nußbaum, W. Waag, M. Garmendia, D.U. Sauer,
205 (2018) 115–133. Comprehensive study of the influence of aging on the hysteresis behavior of a
[81] Nassim Noura, Loïc Boulon, Samir Jemeï, A review of battery state of health lithium iron phosphate cathode-based lithium ion battery—an experimental
estimation methods: hybrid electric vehicle challenges, World Electric Vehicle investigation of the hysteresis, Appl. Energy 171 (2016) 629–645.
Journal 11 (4) (2020) 66. [102] A. Barai, W.D. Widanage, J. Marco, A. McGordon, P. Jennings, A study of the
[82] Zuolu Wang, Xiaoyu Zhao, Fu Lei, Dong Zhen, Fengshou Gu, Andrew D. Ball, open circuit voltage characterization technique and hysteresis assessment of
A review on rapid state of health estimation of lithium-ion batteries in electric lithium-ion cells, J. Power Sources 295 (2015) 99–107.
vehicles, Sustainable Energy Technologies and Assessments 60 (2023) 103457 [103] G. Dong, J. Wei, C. Zhang, Z. Chen, Online state of charge estimation and open
(ISSN 2213-1388). circuit voltage hysteresis modeling of LiFePO4 battery using invariant imbedding
[83] P. Shen, M. Ouyang, L. Lu, J. Li, X. Feng, The co-estimation of state of charge, method, Appl. Energy 162 (2016) 163–171.
state of health, and state of function for lithium-ion batteries in electric vehicles, [104] Y. Cao, R.C. Kroeze, P.T. Krein, Multi-timescale parametric electrical battery
IEEE Trans Veh Technol 67 (2018) 92–103. model for use in dynamic electric vehicle simulations, IEEE Trans. Transp.
[84] D. Wang, F. Yang, L. Gan, Y. Li, Fuzzy prediction of power lithium ion battery Electrif. 2 (2016) 432–442.
state of function based on the fuzzy C-means clustering algorithm, World Electr. [105] Kamruzzaman, X. Zhang, M. Abdelmalak, D. Shi, M. Benidris, A data-driven
Veh. J. 10 (2019) 1. accurate battery model to use in probabilistic analyses of power systems,
[85] H.A. Gabbar, A.M. Othman, M.R. Abdussami, Review of battery management J. Energy Storage 44 (2021) 103292.
systems (BMS) development and industrial standards, Technologies 9 (2021) 28. [106] R. Xiong, W. Sun, Q. Yu, F. Sun, Research progress, challenges and prospects of
[86] P. Křivík, Methods of SoC determination of lead acid battery, J. Energy Storage 15 fault diagnosis on battery system of electric vehicles, Appl. Energy 279 (2020)
(2018) 191–195. 115855.
[87] D. Marcos, M. Garmendia, J. Crego, J.A. Cortajarena, Functional safety Bms [107] H. Kaplan, K. Tehrani, M. Jamshidi, A fault diagnosis design based on deep
design methodology for automotive lithium-based batteries, Energies 14 (2021) learning approach for electric vehicle applications, Energies 14 (2021) 6599.
6942. [108] L. Yao, Y. Xiao, X. Gong, J. Hou, X. Chen, A novel intelligent method for fault
[88] L. Helseth, The self-discharging of supercapacitors interpreted in terms of a diagnosis of electric vehicle battery system based on wavelet neural network,
distribution of rate constants, J. Energy Storage 34 (2021) 102199. J. Power Sources 453 (2020) 227870.
[89] Y. Ji, X. Chen, T. Wang, H. Ji, Y. Zhang, Q. Yuan, L. Li, Coupled effects of [109] N. Tudoroiu (Ed.), Battery Management Systems of Electric and Hybrid Electric
charge–discharge cycles and rates on the mechanical behavior of electrodes in Vehicles, MDPI, Basel, Switzerland, 2021 (ISBN 978-3-0365-1060-6).
lithium–ion batteries, J. Energy Storage 30 (2020) 101577. [110] Y. Wang, J. Tian, Z. Sun, L. Wang, R. Xu, M. Li, Z. Chen, A comprehensive review
[90] S. Singirikonda, Y. Obulesu, Active cell voltage balancing of electric vehicle of battery modeling and state estimation approaches for advanced battery
batteries by using an optimized switched capacitor strategy, J. Energy Storage 38 management systems, Renew. Sustain. Energy Rev. 131 (2020) 110015.
(2021) 102521. [111] P. Sivaraman, C. Sharmeela, IoT-based battery management system for hybrid
[91] M.M. Hoque, M.A. Hannan, A. Mohamed, Voltage equalization control algorithm electric vehicle, in: Artificial Intelligent Techniques for Electric and Hybrid
for monitoring and balancing of series connected lithium-ion battery, J. Renew. Electric Vehicles, Wiley, Hoboken, NJ, USA, 2020, pp. 1–16.
Sustain. Energy 8 (2016) 025703. [112] S. Nyamathulla, D. Chittathuru, S.M. Muyeen, An overview of multilevel inverters
[92] J. Gallardo-Lozano, E. Romero-Cadaval, M.I. Milanes-Montero, M.A. Guerrero- lifetime assessment for grid-connected solar photovoltaic applications, Electronics
Martinez, Battery equalization active methods, J. Power Sources 246 (2014) 12 (2023) 1944.
934–949. [113] Houria Assem, Toufik Azib, Farid Bouchafaa, Cherif Laarouci,
[93] Y.-X. Wang, H. Zhong, J. Li, W. Zhang, Adaptive estimation-based hierarchical Nasreddine Belhaouas, Amar Hadj Arab, Adaptive fuzzy logic-based control and
model predictive control methodology for battery active equalization topologies: management of photovoltaic systems with battery storage, International
part I–balancing strategy, J. Energy Storage 45 (2022) 103235. Transactions on Electrical Energy Systems 2023 (2023) 9065061 (18 pages).
[94] M. Hoque, M. Hannan, A. Mohamed, A. Ayob, Battery charge equalization [114] Xiaosong Hu, Fei Feng, Kailong Liu, Lei Zhang, Jiale Xie, Bo Liu, State estimation
controller in electric vehicle applications: a review, Renew. Sustain. Energy Rev. for advanced battery management: Key challenges and future trends, Renew.
75 (2017) 1363–1385. Sustain. Energy Rev. 114 (2019) 109334.

27