batteries

Article
A Data-Driven Online Prediction Model for Battery Charging
Efficiency Accounting for Entropic Heat
Xiaowei Ding 1 , Weige Zhang 2 , Chenyang Yuan 2 , Chang Ge 2 , Yan Bao 2 , Zhenjia An 3 , Qiang Liu 3 ,
Zhenpo Wang 1, *, Jinkai Shi 2 and Zhihao Wang 2

1 National Engineering Laboratory for Electric Vehicles, Beijing Institute of Technology, Beijing 100081, China;
[email protected]
2 School of Electrical Engineering, Beijing Jiaotong University, Beijing 100044, China;
[email protected] (W.Z.); [email protected] (C.Y.); [email protected] (C.G.);
[email protected] (Y.B.); [email protected] (J.S.); [email protected] (Z.W.)
3 Longruisanyou New Energy Automobile Technology Co., Ltd., Beijing 100075, China;
[email protected] (Z.A.); [email protected] (Q.L.)
* Correspondence: [email protected]

Abstract: This study proposes a charging efficiency calculation model based on an equivalent internal
resistance framework. A data-driven neural network model is developed to predict the charging
efficiency of lithium titanate (LTO) batteries for 5% state of charge (SOC) segments under various
charging conditions. By considering the impact of entropy change on the open-circuit voltage (OCV)
during the charging process, the accuracy of energy efficiency calculations is improved. Incorporating
battery data under various charging conditions, and comparing the predictive accuracy and com-
putational complexity of different hyperparameter configurations, we establish a backpropagation
neural network model designed for implementation in embedded systems. The model predicts the
energy efficiency of subsequent 5% SOC segments based on the current SOC and operating condi-
tions. The results indicate that the model achieves a prediction error of only 0.29% under unknown
charging conditions while also facilitating the deployment of the neural network model in embedded
systems. In future applications, the relevant predictive data can be transmitted in real time to the
cooling system for thermal generation forecasting and predictive control of battery systems, thereby
Citation: Ding, X.; Zhang, W.; Yuan,
enhancing temperature control precision and improving cooling system efficiency.
C.; Ge, C.; Bao, Y.; An, Z.; Liu, Q.;
Wang, Z.; Shi, J.; Wang, Z. A
Keywords: lithium titanate battery; neural network model; entropy heat; energy efficiency
Data-Driven Online Prediction Model
for Battery Charging Efficiency
Accounting for Entropic Heat.
Batteries 2024, 10, 350. https://
doi.org/10.3390/batteries10100350
1. Introduction
1.1. Background and Motivation
Academic Editor: Ottorino Veneri
The development of the electric vehicle and energy storage system industry has led
Received: 19 July 2024 to the widespread adoption of lithium-ion batteries (LIBs) with higher energy density
Revised: 19 September 2024 and greater storage. Lithium titanate batteries (LTO) are a new type of LIB that utilizes
Accepted: 27 September 2024 Li4Ti5O12 instead of graphite for the negative electrode material, enhancing safety be-
Published: 2 October 2024 cause of a significant increase in internal resistance during thermal runaway, preventing
explosions [1–5]. LTO batteries offer higher theoretical capacity and longer life cycles
than conventional graphite-anode lithium batteries. Consequently, LTO batteries have the
Copyright: © 2024 by the authors.
potential to become the negative electrode material of choice for the next generation of
Licensee MDPI, Basel, Switzerland.
LIBs, finding applications in new energy vehicles, electric motorcycles, and other areas that
This article is an open access article demand high safety, stability, and durability. The energy efficiency of LIBs has been a crucial
distributed under the terms and focus of recent research [6–8]. The accurate prediction of battery heat generation relies on
conditions of the Creative Commons precise computation of the battery’s energy efficiency, enabling the transmission of relevant
Attribution (CC BY) license (https:// data to the cooling system (coolant flow rate) for effective heat dissipation control. This
creativecommons.org/licenses/by/ approach aims to narrow the temperature control interval of the battery system, enhancing
4.0/). temperature control accuracy and the heat dissipation system’s efficiency. However, during

Batteries 2024, 10, 350. https://doi.org/10.3390/batteries10100350 https://www.mdpi.com/journal/batteries

Batteries 2024, 10, 350 2 of 17

battery charging, factors such as temperature, charging rate (C-rate), state of charge (SOC),
and capacity decay affect efficiency. Estimating real-time efficiency becomes pivotal for
evaluating and enhancing battery performance under various working conditions [9,10].
In addition, LTO batteries show a temperature drop during charging at room temperature,
caused by a negative entropy heat coefficient resulting in heat absorption. In calculating
energy loss during this process, it is imperative to consider irreversible heat from polar-
ization and ohmic effects, alongside reversible heat from internal chemical (entropy) and
electrochemical reactions. Therefore, a comprehensive analysis of entropy heat variation
on battery internal energy is necessary for accurately assessing the energy efficiency of LTO
battery charging.
In the current research on battery energy efficiency, the ratio of discharge energy to
charge energy is defined, which is used to evaluate energy loss. Coulomb efficiency is
the discharge-to-charge capacity ratio and it is crucial for evaluating battery performance.
Liu et al. [11] employed a theoretical computational approach to determine the energy
conversion efficiency of LTO batteries at various discharge current rates. In practical
applications, the battery’s energy efficiency may deviate from theoretical assumptions
due to operational conditions. Researchers have investigated the relationship between
battery efficiency and factors such as SOC and current. Li et al. [12] studied the energy
efficiency of LIBs utilized as microgrid energy storage devices, discovering that the overall
energy efficiency depends on the energy efficiency under charging, discharging, and
combined charge–discharge conditions. The relationship between energy efficiency and
current rate was analyzed through nonlinear curve fitting and polynomial modeling.
Meanwhile, Lai et al. [13] investigated the correlation between SOC and Coulomb efficiency,
conducting long-term constant voltage charging simulations to simulate the battery’s
open-circuit condition. The study accurately measured and analyzed the relationships
between battery-side reaction current, SOC, and temperature. Subsequently, the Coulomb
efficiency was evaluated under different aging levels, temperature, and SOC. Kang et al. [14]
developed a quantitative relationship between the battery’s OCV and SOC to determine the
energy efficiency of nickel–metal hydride and LIBs. The study contributed to establishing
fundamental test procedures for studying the energy efficiency of LIBs. While these
studies define energy and Coulomb efficiencies and their relationship with SOC, the
continuous real-time estimation of one-way energy efficiency during battery operation
remains unexplored, but it is essential for real-time performance monitoring of the battery.
Wang et al. [15] presented common approaches for studying entropy and heat changes
in lithium batteries during operation. As entropy and heat generation fluctuate with battery
operating conditions, the authors conducted a quantitative investigation to enhance the
management and application of LIBs, particularly high-power batteries. Alkanent et al. [16]
demonstrated that at low discharge currents, entropy effects dominate heat loss, resulting
in battery cooling. During charging, entropy change occurs with a negative absolute value,
leading to additional heat generation in addition to irreversible heat, confirmed through
calorimetric experiments. Doh et al. [17] proposed that battery heating primarily depends
on the impressed current, internal resistance, temperature, and entropy of the battery. Due
to the simultaneous dependence of entropic heat on SOC and temperature, the authors
propose a novel approach for entropy measurement and heat generation calculation under
adiabatic conditions. However, the existing literature lacks an analysis of the effect of
entropy changes on the energy efficiency of battery charging.
Accurate efficiency prediction enhances battery charging performance, provides data
support for the charging optimization strategies for high-power lithium-ion batteries, and
facilitates the online prediction of charging efficiency under various working conditions, all
of which are crucial for enhancing charging efficiency within designated charging durations.
However, the nonlinear heat generation that is sensitive to environmental conditions poses
challenges for accurate prediction of battery charging efficiency, prompting the use of
artificial neural networks (ANNs) [18]. ANN prediction has been widely adopted in LIBs
due to its efficacy in addressing highly nonlinear issues such as battery modeling, SOC
Batteries 2024, 10, 350 3 of 17

estimation, load prediction, and battery temperature forecasting [19–23]. From the existing
literature, the backpropagation (BP) neural network stands out for its superior nonlinear
mapping capabilities, particularly in solving nonlinear prediction and multi-dimensional
problems [24]. Neural networks are integrated into embedded systems to enable real-time
prediction of charging efficiency in a Battery Management System (BMS). However, these
embedded systems have limited computing resources, necessitating neural networks to
achieve sufficient accuracy and computational speed while employing fewer computational
resources. The BP neural network offers lower computational complexity and memory
requirements than other architectures, making it more suitable for implementation in
embedded systems. Therefore, our study chooses BP to model the nonlinear relationship
between input and output parameters for charging efficiency prediction.

1.2. Contributions and Innovation Points

This study derives a formula for one-way charging efficiency based on the equivalent
model of battery internal resistance, incorporating the effects of entropy heat. The formula
of entropy heat coefficient is utilized to modify the OCV, enabling a more accurate charging
efficiency computation. Given the excellent rate performance and temperature charac-
teristics of lithium titanate batteries, their charging efficiency varies significantly across
different conditions. To investigate this phenomenon, we conducted charging experiments
and collected data on the batteries, which allowed us to derive charging efficiency curves
considering the SOC, current, temperature, and entropic heat. A BP neural network predic-
tion model is then established, with the starting SOC, ending SOC, temperature, voltage,
and current as inputs, and the charging efficiency during the SOC charging segment under
the given working conditions as the output. The model parameters are optimized by
comparing prediction accuracy and computational complexity across various hyperparam-
eter configurations. This real-time prediction of changes in battery heat generation across
different time scales could be integrated into future heat dissipation systems, allowing
for the transmission of relevant real-time data to the heat dissipation system, predicting
and controlling the heat dissipation (coolant flow rate) of the battery system, refining
temperature control intervals, and improving both the precision and efficiency of heat
dissipation systems.

1.3. Section Arrangement

The remainder of this paper is organized as follows. Section 2 introduces the model
construction of a lithium battery considering entropy heat and the derivation computation
method of charging efficiency. Section 3 establishes the BP neural network prediction
model, compares various model parameters, and proposes the optimal prediction approach.
Section 4 summarizes the conclusions and outlines future work. Figure 1 illustrates the
overall structural framework of this study.
Batteries 2024,
Batteries 2024, 10,
10, 350
x FOR PEER REVIEW 44 of 18
of 17

Figure 1. Structural framework of

Figure 1. of this
this study.
study.

2. A Battery
2. A Battery Internal
Internal Resistance
Resistance Model
Model Incorporating
Incorporating EntropicEntropic HeatHeat Effects
Effects
2.1. OCV Correction Based on Entropic Heat
2.1. OCV Correction Based on Entropic Heat
Battery charging efficiency is determined by OCV characteristics, with the efficiency
Battery charging efficiency is determined by OCV characteristics, with the efficiency
computation formula derived from the equivalent circuit model. Thus, establishing the
computation formulaisderived
OCV characteristics essentialfromwhen theanalyzing
equivalent thecircuit model.circuit
equivalent Thus,model.
establishing
OCV thede-
OCV characteristics is essential when analyzing the equivalent
pends on the SOC or the state of energy (SOE), represented by OCV-SOC or OCV-SOE circuit model. OCV de-
pends on the SOC
characteristics, or the state
respectively, of energy
in this study. The(SOE), represented
conventional by OCV-SOC
approach involves or averaging
OCV-SOE
measured charge and discharge voltages with low-current charge and discharge. averaging
characteristics, respectively, in this study. The conventional approach involves However,
measured charge and discharge voltages with low-current charge
the significant temperature variation during LTO battery charging affects its internal energy, and discharge. How-
ever, the significant temperature variation during LTO battery charging
reflected in OCV changes with temperature via the entropy heat coefficient. As a result, the affects its internal
energy, reflected
OCV-SOC in OCV changes
curve changes throughout withthetemperature via the entropy
charging process. heat coefficient.
This requires modifyingAs thea
result, the OCV-SOC curve changes throughout the charging
conventional OCV-SOC curve to account for temperature changes affecting the battery’s process. This requires mod-
ifying the
internal conventional
energy, ensuringOCV-SOC curve to account
accurate efficiency for temperature
computation. Therefore,changes
the OCV-SOC affecting the
curve
battery’s
is modified internal
based energy, ensuringheat
on the entropic accurate efficiency
derivation formula.computation. Therefore, the OCV-
SOCConsidering
curve is modified based on
the charging the entropic
process of a LTOheat derivation
battery, formula.
the electric charge migration occurs
Considering the charging process of a LTO battery,
under the influence of the potential difference applied by the external the electric charge migration
charging oc-
device.
This signifies that external work ∆W is performed on the system. At the same time, the
curs under the influence of the potential difference applied by the external charging de-
vice.battery
LTO This signifies thatby
is affected external
entropy heat,Δleading
work W is performed on the system.
to a temperature decreaseAt the same
during time,
charging.
the LTO battery
Assuming is affected
the ambient by entropy
temperature heat, leading
remains constant, to aa temperature
temperature difference
decrease during
arises
charging.the
between Assuming the ambient
battery system and itstemperature
surroundings, remains constant,
causing a temperature
the system to absorb heat ∆q.
difference
arises betweenthe
Consequently, theinternal
battery energy
system of andtheitsbattery
surroundings,
increases.causing
This can thebesystem
quantifiedto absorb
usingheat
the
Δq . Consequently,
mathematical the
expression internal
of the energy
first law ofofthe battery
thermodynamics.increases. This can be quantified us-
ing the mathematical expression of the first law of thermodynamics.
∆U = ∆q + ∆W (1)
Δ U = Δq + ΔW (1)
The change in internal energy of the system is represented by ∆U; the heat absorbed
The change in internal energy of the system is represented by Δ U ; the heat absorbed
by the system is denoted as ∆q; and the work carried out on the system by the charging
by the system is denoted as Δq ; and the work carried out on the system by the charging
device is denoted as ∆W.
device is denoted
During as ΔW process,
the charging . the work carried out on the system primarily consists of
During the
the electrical charging
energy inputprocess,
from thethe work carried
charging devicesout andonthe thework
system primarily
related to the consists
change
of the electrical energy input from the charging devices and the work
in the battery’s volume. In consideration of the size of LTO batteries and their operating related to the change
in the battery’s volume. In consideration of the size of LTO batteries
environment, the work carried out by the system due to volume expansion can be neglected. and their operating
environment, the work
For a reversible carried
battery out when
system, by theone system
mole due to volume
of reaction expansion
occurs, can be
the decrease ne-
in the
glected.
molar Gibbs free energy of the system is equal to the maximum work carried out by the
Batteries 2024, 10, 350 5 of 17

system in terms of volume. According to Faraday’s law of electrolysis, the relationship

between the standard molar Gibbs free energy of formation and the reversible electromotive
force (EMF) of the galvanic cell under conditions of constant temperature and pressure can
be established as follows:
∂G
∆r Gm = ( ) = −zFE (2)
∂ξ T,p
where z represents the number of charges transferred in the electrode reaction, F denotes
the Faraday constant, ξ represents the extent of the electrode reaction, and E is the EMF of
the battery.
In a closed system with constant pressure and temperature, where only volumetric
work is present, the system is homogeneous, and its composition remains unchanged.
According to the fundamental thermodynamic equation S = −(∂G/∂T ) p , in a reversible
battery reaction, there exists non-volumetric electrical work. However, due to the process-
independence of state functions, if the initial and final states of two processes are identical,
the change in thermodynamic state functions will be the same, regardless of the presence of
non-volumetric work. Therefore, the standard molar entropy of formation for a reversible
battery can be calculated analogously to the fundamental thermodynamic equation.

∂∆r Gm ∂E
∆r Sm = ( ) = zF ( ) (3)
∂T p ∂T p

Under constant temperature and pressure conditions, a relationship can be estab-

lished between the standard molar enthalpy of formation of a reversible battery and its
electromotive force.
∂E
∆r Hm = ∆r Gm + T∆r Sm = −zFE + zTF ( ) (4)
∂T p

where ∆r Hm represents the standard molar enthalpy of formation, ∆r Sm represents the

standard molar entropy of formation, and T denotes the temperature of the battery. The
change in the internal energy of the battery during charging can be expressed as follows:

dT ∂E dT
∆r Um = ∆r Hm + mC p = −zFE + zTF ( ) + mC p (5)
dt ∂T p dt

where C p represents the isobaric heat capacity of the battery. By combining this with
Equation (1), the heat transfer from the battery system to its surroundings can be determined
as follows:
∂E dT
∆q = ∆U + ∆W2 = −zFE + zTF ( ) + mC p + IU (6)
∂T p dt
where U represents the battery charging voltage, I represents the battery charging current,
T is the actual temperature of the LTO battery, and E is the open-circuit voltage (Uocv ) of
the LTO battery. The first term on the right-hand side of equation represents the electrical
energy stored in the battery, the second term corresponds to the heat change due to entropy
variation during the chemical reaction, and the third term accounts for the heat stored in
the battery. In this section, we primarily focus on the charging efficiency of the LTO battery.
Therefore, the heat generation rate of the reversible battery system can be further derived
as follows:
∂Uocv ∂Uocv
q′ = − IUocv + IT ( ) + IU = I (U − Uocv ) + IT ( ) (7)
∂T p ∂T p

The battery’s entropy heat coefficient, denoted as (∂Uocv /∂T ) p , is determined by the
battery’s SOC. Therefore, within the same SOC range, OCV changes with temperature.
During constant current charging, entropy heat varies with temperature, which causes
the heat generation rate of the battery to fluctuate accordingly. This is reflected in the
 The battery’s entropy heat coefficient, denoted as (∂Uocv / ∂T ) p , is determined by the
battery’s SOC. Therefore, within the same SOC range, OCV changes with temperature.
Batteries 2024, 10, 350 During constant current charging, entropy heat varies with temperature, which causes the
6 of 17
heat generation rate of the battery to fluctuate accordingly. This is reflected in the temper-
ature dependence of the OCV as determined by the entropy heat coefficient. The entropy
temperature
heat representsdependence
the thermal ofenergy
the OCV as determined
generated by changesby the in entropy heat internal
the battery’s coefficient. The
energy,
entropy
and this heat represents
factor the thermalwhen
must be considered energycalculating
generatedenergyby changes in the To
efficiency. battery’s
obtaininternal
a more
energy, and
accurate this factor
estimation of must
battery be efficiency
consideredaccounting
when calculating energy
for entropy efficiency.
heat variation, To the
obtain
rela-a
more accurate estimation of battery efficiency accounting for entropy
tionship between the OCV and temperature for LTO batteries can be corrected using the heat variation, the
relationship
following between the OCV and temperature for LTO batteries can be corrected using
equation:
the following equation:
∂U (T , SOC )
U ocv (T , SOC ) = U ocv (Tref , SOC ) + (T − Tref ) × ( ocv ref )p (8)
∂T( Tre f , SOC )
∂Uocv
Uocv ( T, SOC ) = Uocv ( Tre f , SOC ) + ( T − Tre f ) × ( ) (8)
∂T p
where U ocv (T , SOC ) represents the battery’s open-circuit voltage at the corresponding
temperature,
where Uocv ( T,and
SOCT) denotes the the
represents actual temperature
battery’s of the voltage
open-circuit LTO battery.
at the Ucorresponding
ocv (Tref , SOC )
is
temperature, and T denotes the actual temperature of the LTO battery.
the reference open-circuit voltage measured at a constant temperature ofocv U ( T
25 °C, SOC )
re f and T is
,
the reference open-circuit voltage measured at a constant temperature of 25 ◦ C, and Trereff
is the reference
is the reference temperature
temperature of of 25
25 ◦°C. The corrected
C. The corrected open-circuit
open-circuit voltage
voltage accounts
accounts for the
for the
entropy changes in
entropy changes in the
the battery’s
battery’s open-circuit
open-circuit voltage
voltage due
due toto temperature
temperature variations.
variations.

2.2. Improved Internal Resistance Model for Lithium-Ion Batteries

Selecting an
an appropriate
appropriateequivalent
equivalentcircuit
circuitmodel
modelisiscrucial
crucialinin deriving
deriving thethe charging
charging ef-
efficiency equation.
ficiency equation. Common
Common equivalent
equivalent circuit
circuit models
models include
include Rint,
Rint, Thevenin,
Thevenin, PNGV,
PNGV, n-RC,n-
RC,
and and fractional-order
fractional-order models models[25]. [25].
SinceSince our study
our study considers
considers the battery
the battery chargingcharging effi-
efficiency
ciency within
within a constant
a constant currentcurrent
charging charging
system andsystemthe and
shortthe short polarization
polarization duration,duration, the
the internal
internal
resistance resistance
model ismodel
chosen.is This chosen.
modelThisis model
widelyisutilized
widelyinutilized
studying in studying battery
battery charging
charging processes
processes employingemploying
the constant the constant
current current
charging modecharging
and the mode andresistance
internal the internal modelre-
sistance
is chosenmodel is chosen
to examine to examine
the effect the resistance
of internal effect of internal
on batteryresistance on battery perfor-
performance.
mance. As an illustration, we demonstrate the modeling of internal resistance. The internal
resistance
As anequivalent
illustration,circuit model (Figure
we demonstrate the 2) utilizes Ohm’s
modeling law resistance.
of internal to establishThe the internal
voltage
and current relationship (Equation (9)) in a battery. In constant current
resistance equivalent circuit model (Figure 2) utilizes Ohm’s law to establish the voltage charging mode, the
current
and remains
current constant, (Equation
relationship and OCV varies (9)) in with SOC (Section
a battery. In constant2.1).current
The model resistance
charging mode, R
reflects
the the variation
current in the internal
remains constant, and OCV resistance
varies of with
the battery. The internal
SOC (Section 2.1). Theresistance
modelofre-a
lithium-ion
sistance batterythe
R reflects is influenced
variation in bythe
a combination of factors,
internal resistance of including
the battery. temperature,
The internal SOC,
re-
aging, current
sistance intensity, and
of a lithium-ion batterythe design and materials
is influenced of the battery
by a combination itself [26–28].
of factors, Based
including on
tem-
the experimental conditions in this study, it can be concluded that
perature, SOC, aging, current intensity, and the design and materials of the battery itselfthe internal resistance
of the battery
[26–28]. Basedchanges nonlinearly with
on the experimental respectintothis
conditions battery temperature,
study, charging that
it can be concluded current,
the
and SOC.
internal resistance of the battery changes nonlinearly with respect to battery temperature,
charging current, and U ch ( t ) = Uocv ( T, SOC ) + I ( t ) × R ( T, SOC, I )
SOC. (9)

Figure 2.
Figure Internal resistance
2. Internal resistance of
of equivalent
equivalent circuit
circuit model.
model.

Under constant current conditions, the SOC varies linearly with the runtime. Further-
more, the applied current during constant current charging and discharging is directly
proportional to the charging rate, as described in Equation (10). For a 25 Ah LTO battery, a
charging current of 25 A corresponds to 1 C-rate.
 Uch ( t ) =Uocv (T , SOC ) +I ( t ) × R (T , SOC, I ) (9)

Under constant current conditions, the SOC varies linearly with the runtime. Fur-
Batteries 2024, 10, 350
thermore, the applied current during constant current charging and discharging is di-
7 of 17
rectly proportional to the charging rate, as described in Equation (10). For a 25 Ah LTO
battery, a charging current of 25 A corresponds to 1 C-rate.
SOC computation formula at rated capacity C0 is as follows:
SOC computation formula at rated capacity C0 is as follows:
100Z tt
SOC ( t ) = SOC ( t − 1)100 I ( t ) dt
SOC (t) = SOC (t − 1) + C0· t −1 I (t)dt
+ ⋅ (10)
(10)
C0 t −1
By conducting 0.05 C constant current charge and discharge tests on a LTO battery,
By conducting 0.05 C constant current charge and discharge tests on a LTO battery,
the battery’s voltage variation curves were obtained during both the charging and dis-
the battery’s voltage variation curves were obtained during both the charging and dis-
charging processes.After
charging processes. Afteraveraging
averagingand andinverting
invertingthe thecurves,
curves, the
the original
original OCVOCV curve
curve was was
derived [29]. The OCV acquired by this approach does not take into
derived [29]. The OCV acquired by this approach does not take into account the influ- account the influence
of
encetemperature
of temperaturechanges during
changes duringthetheactual
actualcharging
charging process
process at at different
differenttemperatures.
temperatures.
Temperature variationsduring
Temperature variations duringthethecharge
chargeand anddischarge
dischargeprocess
processaffect
affect the
the reversible
reversible elec-
elec-
trochemical reactions within the battery, leading to changes in its internal
trochemical reactions within the battery, leading to changes in its internal energy. Therefore, energy. There-
fore,
whenwhen calculating
calculating one-way
one-way charging
charging efficiency,
efficiency, it is essential
it is essential to accountto account
for theseforinternal
these in-
ternal
energyenergy
changes.changes.
From From the perspective
the perspective of electrical
of electrical work,work, the effect
the effect of entropy
of entropy heat onheat
on efficiency
efficiency is considered
is considered through
through adjustments
adjustments to thetoOCV.
the OCV.
Figures 3–5
3–5 compare
compare thethe OCV-SOC
OCV-SOCcurves curvesofofan anLTO
LTObattery
batteryobtained
obtainedfrom
from a low-
a low-
current experiment at room temperature (25 ◦ C) and the OCV-SOC curves corrected using
experiment at room temperature (25 °C) and the OCV-SOC curves corrected using
the entropy
the entropy heat
heat coefficient
coefficient from
from Equation
Equation (8) (8) under
under 12 12different
differentconstant
constantcurrent
currentcharging
charg-
ing conditions.
conditions.

OCV 0°C_0.5C 0°C_1C 0℃_2C 0℃_4C

Original OCV Original OCV
2.55 Revised OCV 2.55 Revised OCV
2.50 2.50
2.45
OCV(V)

OCV(V)
2.45
2.40 2.40
2.35
2.35
2.30
2.30
2.25 (a) (b)
2.25
0 20 40 60 80 100 0 20 40 60 80 100
SOC(%) SOC(%)

Original OCV Original OCV

2.55 Revised OCV 2.55 Revised OCV
2.50 2.50
2.45
OCV(V)

OCV(V)

2.45
2.40 2.40
2.35
2.35
2.30
2.30
2.25 (c) (d)
2.25
0 20 40 60 80 100 0 20 40 60 80 100
SOC(%) SOC(%)
Figure 3.3. Comparison
ComparisonofofOCV-SOC
OCV-SOCcurves
curvesbefore
beforeand
andafter
afterOCV
OCVcorrection at at
correction (a)(a) 0 ◦ C_0.5
0 °C_0.5 C; (b)
C; 0
°C_1 ◦C; (c) 0 °C_2 ◦C; (d) 0 °C_4◦C.
(b) 0 C_1 C; (c) 0 C_2 C; (d) 0 C_4 C.
Batteries 2024, 10, 350
x FOR PEER REVIEW 8 8of
of 18
17
Batteries 2024, 10, x FOR PEER REVIEW 8 of 18

OCV 25℃_0.5C 25℃_1C 25℃_2C 25℃_4C

OCV 25℃_0.5C 25℃_1C 25℃_2C 25℃_4C
Original OCV Original OCV
2.55 Original OCV
Revised OCV 2.55 Original OCV
Revised OCV
2.55 Revised OCV 2.55 Revised OCV
2.50 2.50
2.50 2.50
2.45

OCV(V)

OCV(V)
2.45
2.45

OCV(V)

OCV(V)
2.45 2.40
2.40 2.40
2.40 2.35
2.35 2.35
2.35 2.30
2.30 2.30
2.30 (a) 2.25 (b)
2.25 (a) 2.25 (b)
2.25 0 20 40 60 80 100 0 20 40 60 80 100
0 20 40 60 80 100 0 20 40 60 80 100
SOC(%) SOC(%)
SOC(%) SOC(%)
Original OCV Original OCV
2.55 Original OCV
Revised OCV 2.55 Original OCV
2.55 2.55 Revised OCV
2.50 Revised OCV Revised OCV
2.50
2.50 2.50
2.45
OCV(V)

OCV(V)
2.45
2.45
OCV(V)

OCV(V)
2.40 2.45
2.40 2.40
2.35 2.40
2.35 2.35
2.30 2.35
2.30 2.30
2.25 (c) 2.30 (d)
2.25 (c) 2.25 (d)
0 20 40 60 80 100 2.25 0 20 40 60 80 100
0 20 40 60 80 100 0 20 40 60 80 100
SOC(%) SOC(%)
SOC(%) SOC(%)
Figure
Figure 4.4.Comparison
ComparisonofofOCV-SOC
OCV-SOC curves before
curves andand
before afterafter
OCVOCV
correction at (a)at
correction 25(a) 25 ◦ C_0.5
°C_0.5 C; (b) 25
C;
Figure
°C_1 C;
◦
4.(c)
Comparison
25 °C_2 C;
◦
of OCV-SOC
(d) 25 °C_4 C.
◦
curves before and after OCV correction at (a) 25 °C_0.5 C; (b) 25
(b) 25 C_1 C; (c) 25 C_2 C; (d) 25 C_4 C.
°C_1 C; (c) 25 °C_2 C; (d) 25 °C_4 C.
OCV －20℃_0.5C －20℃_1C －20℃_2C －20℃_4C
OCV －20℃_0.5C －20℃_1C －20℃_2C －20℃_4C
Original OCV Original OCV
2.55 Original OCV
Revised OCV
2.55 Original OCV
Revised OCV
2.55 Revised OCV
2.55 Revised OCV
2.50 2.50
2.50 2.50
2.45
OCV(V)

OCV(V)

2.45
2.45
OCV(V)

OCV(V)

2.45 2.40
2.40 2.40
2.40 2.35
2.35 2.35
2.35 2.30
2.30 2.30
2.30 (a) 2.25 (b)
2.25 (a) 2.25 (b)
2.25 0 20 40 60 80 100 0 20 40 60 80 100
0 20 40 60 80 100 0 20 40 60 80 100
SOC(%) SOC(%)
SOC(%) SOC(%)
Original OCV Original OCV
2.55 Original OCV
Revised OCV 2.55 Original OCV
Revised OCV
2.55 Revised OCV 2.55 Revised OCV
2.50 2.50
2.50 2.50
2.45
OCV(V)

OCV(V)

2.45
2.45
OCV(V)

OCV(V)

2.40 2.45
2.40 2.40
2.35 2.40
2.35 2.35
2.30 2.35
2.30 2.30
2.25 (c) 2.30 (d)
2.25 (c) 2.25 (d)
0 20 40 60 80 100 2.25 0 20 40 60 80 100
0 20 40 60 80 100 0 20 40 60 80 100
SOC(%) SOC(%)
SOC(%) SOC(%)
Figure 5. Comparison of OCV-SOC curves before and after OCV correction at (a) −20 °C_0.5 C; (b)
Figure
Figure
−20 °C_15. Comparison
Comparison
5. C; (c) −20 °C_2 ofofOCV-SOC
C;OCV-SOC curves
C. before
curves
(d) −20 °C_4 beforeand
andafter OCV
after correction
OCV at at
correction (a)(a) 20 ◦ C_0.5
−20−°C_0.5 C; (b)
C;
−20 °C_1
(b) − ◦ C; (c)C;−20
20 C_1 (c) °C_2
−20 C;
◦ (d)C;
C_2 −20
(d)°C_4
−20C.C_4 C.
◦

At ambient temperatures of 0 ◦°C and −20 °C, the corrected OCV curve of the LTO
◦ C, the
At ambient
At ambient temperatures
temperatures of C and −
of 0 °C −2020 °C, the corrected
corrected OCV
OCV curve
curve of
of the
the LTO
battery across the full SOC range was found to be higher than the original OCV curve, and
battery across
battery acrossthe
thefull
fullSOC
SOCrange
range was
was found
found to higher
to be be higher thanthan the original
the original OCVOCV curve,
curve, and
the difference between them tended to decrease in the high SOC range. This phenomenon
the difference between them tended to decrease in the high SOC range. This phenomenon
Batteries 2024, 10, 350 9 of 17

and the difference between them tended to decrease in the high SOC range. This phe-
nomenon occurs because the entropy heat coefficient of the LTO battery is negative in the
0–80% SOC range, and the ambient temperature is lower than the reference temperature,
resulting in the corrected OCV being higher than the original OCV. In the high SOC range,
the entropy heat coefficient becomes positive, causing the gap between the corrected and
original OCV curves to narrow, although the corrected curve remains above the original.
It can be inferred that when the battery temperature exceeds the reference temperature,
the actual OCV curve shifts downward. Additionally, as the battery temperature deviates
further from the reference temperature, the difference between the actual OCV and the
experimentally measured original OCV increases. From Figure 4, it is evident that during
charging at 25 ◦ C, the difference between the corrected and original OCV is minimal in the
low SOC range. However, as charging progresses and the battery continues to accumulate
heat, the corrected OCV becomes higher than the original OCV. It is noteworthy that the
correction of OCV does not vary significantly with the charging rate. This is because the
temperature rise of the battery at different charging rates under the same ambient tempera-
ture is much smaller than the impact of ambient temperature changes on the OCV. Under
actual charging conditions, the battery temperature differs from that during low-current
charge and discharge tests, necessitating corrections to the battery OCV.

2.3. Derivation of Charging Efficiency Considering Entropic Heat

Battery efficiency mainly includes capacity efficiency, energy efficiency, and voltage
efficiency, among others. The battery’s charging/discharging efficiency (energy efficiency)
reflects the extent of energy loss during the charging and discharging process. Efficiency
can be categorized by the direction of energy flow, including charging, discharging, and
round-trip efficiency. One-way efficiency includes both charging and discharging efficiency,
while round-trip efficiency refers to charging and discharging within the same SOC range.
If the SOC ranges during the charge and discharge half cycles differ, the resulting round-
trip efficiency becomes invalid. Round-trip efficiency can be easily determined from
the measured current and voltage. In contrast, determining one-way efficiency requires
additional data, as it pertains to quantities that cannot be measured directly. Round-trip
energy efficiency is expressed as follows:

Edis
ηcycle,E = (11)
Ech

where Edis and Ech are the discharging and charging energy, respectively, both of which
can be calculated from the following general expression:
Z t
E= U (t) I (t)dt (12)
0

where U and I represent the discharge/charge voltage and current, respectively; t represents
the charge/discharge time. The round-trip Coulomb efficiency is expressed as follows:

C dis
ηcycle,I = (13)
C ch

where C dis and C ch represent the amount of discharging and charging charge, respectively,
determined as follows: Z t
C= I (t)dt (14)
0
where I represents the measured discharge/charge current; t represents the duration of
discharge/charge. The round-trip voltage efficiency is defined as follows:

U dis
ηcycle,U = (15)
U ch
Batteries 2024, 10, 350 10 of 17

where U dis and U ch represent the average discharge and charge voltage, respectively.
Round-trip efficiency is calculated by multiplying the charging and discharging efficiencies.

ηcycle = ηch · ηdis (16)

One-way efficiency during charging and discharging is heavily influenced by condi-

tions such as charging/discharging current and the SOC range. These conditions often
vary significantly between the charging and discharging phases, making it difficult for
round-trip efficiency to distinguish between them. Therefore, understanding the loss dur-
ing each individual cycle (charging or discharging) necessitates knowledge of the one-way
efficiency value.
Following the model design in Section 2.2, the efficiency can be determined by calcu-
lating the one-way charging voltage relative to the OCV and integrating it over time to
obtain the efficiency.
One-way charging voltage efficiency is defined as follows:
Rt
UOCV dt
ηch = R0 t (17)
0 Uch dt

One-way discharging voltage efficiency is expressed as follows:

Rt
Udis dt
ηdis = R t0 (18)
0 UOCV dt

When calculating the one-way charging/discharging efficiency, it is important to

ensure that the time range for voltage integration is consistent. The calculated efficiency
should correspond to the one-way charging/discharging efficiency within this time range.
Based on the internal resistance equivalent model Equation (9) in Section 2.2, the
calculation formula for the one-way charging efficiency of the battery in continuous time
is derived. R t2
t1 UOCV ( T, SOC ) × Idt
ηch = R t (19)
2
( U ( T, SOC ) × I + I 2 R ) dt
t 1
OCV i

The one-way charging efficiency of a lithium-ion battery is closely related to the

OCV and internal resistance at the current charging moment. Using laboratory battery
testing equipment, a 25 Ah LTO battery was charged at ambient temperatures of −20 ◦ C,
0 ◦ C, and 25 ◦ C, with charging rates of 0.5 C, 1 C, 2 C, and 4 C, respectively. Data such
as charging voltage, current, charging time, and temperature were collected under each
charging condition. The battery SOC was calculated using Equation (10), and the corrected
OCV data were obtained in conjunction with Section 2.1. The one-way charging efficiency
was then computed under 12 orthogonal charging conditions. It is worth noting that,
theoretically, battery efficiency changes continuously over the course of charging. However,
considering the stability of charging efficiency over short time intervals and the practical
significance of efficiency prediction, we discretized the one-way charging efficiency at
every 5% SOC increment. The average charging efficiency within each 5% SOC interval
was calculated using Equation (19). This efficiency value represents the charging efficiency
within the corresponding SOC range.
Figures 6–8 compare the charging efficiency of the battery under 12 different charging
conditions, before and after OCV correction. As the charging rate increases and the ambient
temperature decreases, the charging efficiency across different SOC segments declines
significantly. Before OCV correction, the charging efficiency varies noticeably across the
different conditions, particularly under the −20 ◦ C_4 C condition, where the average
charging efficiency is only 82%. After comparing the battery charging efficiency before and
after OCV correction, a marked improvement is observed. This change is attributed to the
variation in electrochemical potential (the internal energy of the battery system) caused by
Batteries 2024, 10, x FOR PEER REVIEW 11 of 18

Batteries 2024, 10, 350 charging efficiency is only 82%. After comparing the battery charging efficiency before 11 of 17
and after OCV correction, a marked improvement is observed. This change is attributed
to the variation in electrochemical potential (the internal energy of the battery system)
caused by temperature
temperature during the
during the charging charging
process process
of LTO of LTO
batteries. Thebatteries.
change in The change
energy in en-
efficiency
explains
ergy the variation
efficiency explainsin the
the battery’s
variationinternal energy, asinternal
in the battery’s the unique thermodynamic
energy, as the unique entropy
ther-
heat coefficient
modynamic of LTO
entropy leads
heat to an endothermic
coefficient of LTO leads reaction
to anduring charging.
endothermic A portion
reaction duringof
the reversible
charging. heat stored
A portion of the within the heat
reversible battery contributes
stored within theto battery
improving chargingtoefficiency.
contributes improv-
Under
ing the 0 ◦ C_0.5
charging C condition,
efficiency. Under thethe0charging
°C_0.5 Cefficiency
condition, increases by nearly
the charging 1%. It is
efficiency evident
increases
that OCV correction has a significant impact on the calculation of unidirectional
by nearly 1%. It is evident that OCV correction has a significant impact on the calculation charging
efficiency
of under different
unidirectional charging operating
efficiency conditions. However,
under different given that
operating the LTO battery
conditions. However, can
only charge
given upLTO
that the to 25% SOCcan
battery under
onlythe −20 ◦up
charge C_4toC25%
condition,
SOC underthe efficiency dataCare
the −20 °C_4 limited,
condition,
andefficiency
the this condition is excluded
data are limited, from further
and this analysis.
condition is excluded from further analysis.

Figure
Figure 6.
Batteries 2024, 10, x FOR PEER REVIEW 6. Comparison
Comparison of
of efficiency
efficiency before
before and
and after
after OCV
OCV correction
correction at
at (a)
(a) 00 °C_0.5
◦ C_0.5 C;
C; (b)
(b) 00 12
°C_1
of C;
◦ C_1 C;
18
(c)
(c) 00 °C_2
◦ C_2 C;
C; (d)
(d) 00 °C_4
◦ C_4 C.
C.

Figure 7. Cont.
Batteries 2024, 10, 350 12 of 17

Figure
Figure 7.7. Comparison
Comparison of of efficiency
efficiencybefore
beforeand
andafter
afterOCV
OCVcorrection
correctionatat
(a)(a)
2525 °C_0.5
◦ C_0.5 C;C;
(b)(b)
25 25 °C_1
◦ C_1 C;
Figure
C; (c) 25
◦
7. °C_2
Comparison
C; (d) 25
◦
of efficiency
°C_4 C. before and after OCV correction at (a) 25 °C_0.5 C; (b) 25 °C_1
(c) 25 C_2 C; (d) 25 C_4 C.
C; (c) 25 °C_2 C; (d) 25 °C_4 C.

Figure 8. Comparison of efficiency before and after OCV correction at (a) −20 °C_0.5 C; (b) −20 °C_1
Figure
C; 8.
8. Comparison
(c) −20
Figure of
of efficiency
°C_2 C; (d) −20
Comparison before and
°C_4 C. before
efficiency and after
after OCV
OCVcorrection
correctionatat(a)
(a)−−20
20 ◦°C_0.5 C; (b)
C_0.5 C; (b) −
−20
20°C_1
◦ C_1
C; (c) −20 °C_2 C; (d) −20 °C_4 C.
C; (c) −20 ◦ C_2 C; (d) −20 ◦ C_4 C.

3. Data-Driven Charging Efficiency Prediction Method

The data analysis in Section 2 determined the charging efficiency at each 5% SOC
segment under 11 fixed working conditions. However, real-world charging processes
involve varying operational conditions and result in dynamic variations in battery tempera-
ture. Consequently, a predictive method is necessary to estimate charging efficiency across
diverse working conditions. Using the aforementioned data, Section 3 proposes employing
a neural network to predict charging efficiency, with consideration of entropy heat changes
across various working conditions.
Feedforward Neural Networks (FNNs), Recurrent Neural Networks (RNNs), and
Long Short-Term Memory Networks (LSTMs) have been widely applied in battery state
prediction. FNNs are primarily used for processing static data, making them suitable
for tasks such as pattern recognition and regression. In contrast, LSTMs and RNNs are
better suited for time series data, as they can retain historical information [30,31]. Since
 employing a neural network to predict charging efficiency, with consideration of entropy
heat changes across various working conditions.
Feedforward Neural Networks (FNNs), Recurrent Neural Networks (RNNs), and
Long Short-Term Memory Networks (LSTMs) have been widely applied in battery state
Batteries 2024, 10, 350 prediction. FNNs are primarily used for processing static data, making them suitable for
13 of 17
tasks such as pattern recognition and regression. In contrast, LSTMs and RNNs are better
suited for time series data, as they can retain historical information [30,31]. Since the effi-
ciency
the prediction
efficiency in this in
prediction study
this is a one-step
study forecastforecast
is a one-step that doesthatnot relynot
does onrely
previous battery
on previous
states, and considering the nonlinear relationship between battery efficiency
battery states, and considering the nonlinear relationship between battery efficiency and and key pa-
rameters such as SOC, current, and temperature, this study adopts a feedforward
key parameters such as SOC, current, and temperature, this study adopts a feedforward neural
network
neural with the
network backpropagation
with (BP) algorithm,
the backpropagation i.e., a BP
(BP) algorithm, i.e.,neural network.
a BP neural network.

3.1. Model
3.1. Model Development
Development and
and Training
TrainingSetting
Setting
Based on
Based on the
the analysis
analysis inin Section
Section 2.3,
2.3, itit isisevident
evidentthat
thatbattery
battery charging
charging efficiency
efficiency isis
closely related
closely related to
tothe
theOCV
OCVand andinternal
internalresistance
resistanceduring duringthe thecharging
chargingprocess.
process. Internal
Internal
resistance
resistanceexhibits
exhibitsaanonlinear
nonlinearrelationship
relationshipwith withkey keyparameters
parameterssuch suchasasSOC,
SOC,current,
current,and
and
temperature.
temperature. Battery
Battery terminal
terminal voltage,
voltage, inin particular,
particular,isis aacritical
critical parameter
parameterfor foridentifying
identifying
the
the charging
chargingprocess.
process. The
The dataset
dataset used
used inin this
thisstudystudywas
wascollected
collected from
frombattery
batterycharging
charging
experiments
experiments under various conditions, with charging rates of 0.5 C, 1 C, 2 C, and44C,
under various conditions, with charging rates of 0.5 C, 1 C, 2 C, and C,and
and
ambient temperatures of − 20 ◦ C, 0 ◦ C, and 25 ◦ C. The input parameters for the BP neural
ambient temperatures of −20 °C, °C, °C. parameters for the BP neural
network
networkmodel
modelinclude
includestarting
startingtemperature,
temperature,charging chargingcurrent,
current,voltage,
voltage,starting
startingSOC,
SOC,and
and
ending
endingSOC,
SOC,while
whilethetheoutput
outputparameter
parameteris is the the battery charging
battery charging efficiency
efficiencycorresponding
corresponding to
5% SOC
to 5% SOCsegments.
segments.Each training
Each dataset
training spans
dataset spansa 5%aSOC interval
5% SOC between
interval the starting
between and
the starting
ending SOC,SOC,
and ending with with
adjacent training
adjacent sets sliding
training sets slidingby 1%by SOC. BasedBased
1% SOC. on thisonconfiguration,
this configura-a
neural network architecture was developed, consisting of a five-input,
tion, a neural network architecture was developed, consisting of a five-input, single-out- single-output BP
neural
put BPnetwork prediction
neural network model, as
prediction illustrated
model, in Figurein9.Figure 9.
as illustrated

Figure9.9. Structure
Figure Structureof
ofBP
BPnetwork
networkSOC
SOCsegment
segmentcharging
chargingefficiency
efficiencyprediction
predictionmodel.
model.

Consideringthe
Considering thediversity
diversityofof ambient
ambient temperature
temperature andand charging
charging raterate in real-world
in real-world effi-
efficiency
ciency prediction,
prediction, studies
studies training
training neural
neural networks
networks withwith charging
charging datadata across
across all work-
all working
ing conditions
conditions are limited.
are limited. ThisThis study
study divides
divides allall data
data (11(11groups)
groups)under
underdifferent
differentworking
working
conditions
conditionsinto
intotraining
trainingsets
sets(8(8groups)
groups)and
andtest
testsets
sets(3(3groups)
groups) based
based onon
the working
the working condi-
con-
tions to validate
ditions thethe
to validate method’s
method’s generalization capacity.
generalization capacity.TheThepractical prediction
practical scenarios
prediction of
scenarios
the user
of the usually
user fallfall
usually within thethe
within working conditions
working conditionsof the battery’s
of the limits
battery’s (Table
limits 1). 1).
(Table

Table1.1.Division
Table Divisionof
ofneural
neuralnetwork
networkdatasets.
datasets.

TrainingSet
Training Set TestSet
Test Set
−20
−20 C_0.5 C, −20 C_1 C, −20 ◦°C_2
°C_0.5
◦ C, −20 °C_1
◦ C, −20 C_2 C,C,0 0◦ C_4
°C_4C,C,
◦ ◦ ◦ ◦
00 ◦°C_0.5 C,00◦ C_1
C_0.5 C, °C_1C,C,
0 ◦0C_2
°C_2
C C
25
25°C_0.5 C, 25
C_0.5 C, 25 °C_1
C_1 C,C,2525C_2
°C_2C, C,
25 25C_4°C_4
C C

3.2. Comparison of Efficiency Prediction Results

The hyperparameters of a neural network act as control knobs that regulate its struc-
ture, functionality, and efficiency. Common hyperparameters include the learning rate,
number of iterations, number of hidden layers, number of neurons in hidden layers, acti-
vation functions, optimizers, and network initialization methods. Selecting appropriate
hyperparameters is crucial for the training and performance of a neural network. For
example, the optimization algorithm directly affects the model’s convergence speed and
performance, while the number of hidden layers and neurons determines the structural
complexity of the model. Given that battery efficiency prediction involves nonlinear func-
tion approximation problems [32,33], our study adopts a feedforward neural network with
a single hidden layer as the basic framework. We combine four optimization algorithms
suited for function approximation problems, three commonly used activation functions, and
two initialization methods for neural networks while varying the number of neurons in the
 ample, the optimization algorithm directly affects the model’s convergence speed and per-
formance, while the number of hidden layers and neurons determines the structural com-
plexity of the model. Given that battery efficiency prediction involves nonlinear function
approximation problems [32,33], our study adopts a feedforward neural network with a
Batteries 2024, 10, 350 single hidden layer as the basic framework. We combine four optimization algorithms 14 of 17
suited for function approximation problems, three commonly used activation functions,
and two initialization methods for neural networks while varying the number of neurons
hidden layer. This
in the hidden study
layer. investigates
This the impact
study investigates theofimpact
different
of hyperparameters—activation
different hyperparameters—
functions,
activation optimization algorithms,algorithms,
functions, optimization number ofnumber
hidden oflayer neurons,
hidden layerand initialization
neurons, and ini-
methods—on prediction outcomes. Figure 10 illustrates the configuration of various
tialization methods—on prediction outcomes. Figure 10 illustrates the configuration neuralof
network hyperparameters.
various neural network hyperparameters.

Figure10.
Figure 10. Hyperparameter
Hyperparameteroptimization
optimizationspace.
space.

In this study,
In study, gridgridsearch
searchwas wasemployed
employed to to
optimize
optimize 72 different hyperparameter
72 different hyperparameter con-
configurations.
figurations. TheThe neural
neural network
network was
was trainedusing
trained usingaabatch
batch learning
learning mode, where the the
weights
weights were updated
updated after afterall
allinputs
inputsinin thethe training
training setset
werewere applied.
applied. Mean Mean squared
squared error
error
(MSE)(MSE) was as
was used used
the as the performance
performance function, function,
and an and earlyanstopping
early stopping
mechanismmechanism
was im-
was implemented to prevent overfitting. Under two different data
plemented to prevent overfitting. Under two different data partitioning strategies, the av-partitioning strategies,
the average
erage prediction
prediction accuracyaccuracy
for eachforhyperparameter
each hyperparameter configuration
configuration was scored.
was scored. The
The predic-
− 3 − 5
prediction
tion error range (5 × 10(5 ,×1 10
error range −3 −5 × 10 ) corresponds
× 10 , )1corresponds to scores to scores 0between
between and 100.0 and 100.
Figure
Figure11 11illustrates
illustratesthe thevariation
variation in inprediction
prediction accuracy
accuracyacross different
across hyperparam-
different hyperpa-
eter configurations.
rameter configurations. TableTable
2 presents the optimal
2 presents hyperparameter
the optimal hyperparametersettings for efficiency
settings for effi-
prediction. The sigmoid activation function offers the advantage of
ciency prediction. The sigmoid activation function offers the advantage of a limited outputa limited output range,
ensuring data stability during transmission. The scaled conjugate
range, ensuring data stability during transmission. The scaled conjugate gradient (SCG) gradient (SCG) algorithm
does not require
algorithm does not linerequire
searchline
in each
searchiteration,
in each which
iteration, enhances numericalnumerical
which enhances stability. stabil-
Com-
pared to activation functions, optimization algorithms, and the
ity. Compared to activation functions, optimization algorithms, and the number of neu- number of neurons in the
hidden
rons in layers, the initialization
the hidden method is less
layers, the initialization sensitive
method is lesstosensitive
prediction error for efficiency.
to prediction error for
The Nguyen–Widrow initialization method constrains each
efficiency. The Nguyen–Widrow initialization method constrains each neuron within neuron within a specific rangea
through weight initialization, aiming to accelerate model convergence.
specific range through weight initialization, aiming to accelerate model convergence. However, under
unknown
However, training conditions
under unknown in the test
training set, the random
conditions in the test initialization method
set, the random improves
initialization
the
method improves the generalization and prediction accuracy of the neural network. in
generalization and prediction accuracy of the neural network. Using 8 neurons the
Using
hidden layer has proven to be the optimal choice for efficiency prediction,
8 neurons in the hidden layer has proven to be the optimal choice for efficiency prediction, as too many
neurons may lead to overfitting, while too few may fail to capture the non-linear rela-
as too many neurons may lead to overfitting, while too few may fail to capture the non-
tionships of efficiency accurately. For the neural network used in this study, the input
linear relationships of efficiency accurately. For the neural network used in this study, the
layer consists of 5 neurons, the hidden layer has 8 neurons, and the output layer contains
input layer consists of 5 neurons, the hidden layer has 8 neurons, and the output layer
1 neuron. The time complexity can be expressed as O(5 × 8 + 8 × 1) = O(48), indicating
contains 1 neuron. The time complexity can be expressed as O(5 × 8 + 8 × 1) = O(48),
that the network performs approximately 48 multiplication and addition operations during
each forward pass. The space complexity is O(5 × 8 + 8 × 1 + 8 + 1) = O(57), meaning
that the network requires approximately 57 storage units for weights and biases. Given
the network’s small size, the computational and storage demands per input remain fixed.
Figure 12a,b depicts the predicted efficiency curves, actual curves, and error curves for
SOC segment charging efficiency under 0 ◦ C_0.5 C, 0 ◦ C_1 C, and 0 ◦ C_2 C conditions,
respectively. The average relative error for efficiency prediction using this approach is
0.29%. Under this optimal hyperparameter configuration, the neural network accurately
predicted the battery charging efficiency for the 5% SOC charging segment, confirming the
superiority of the feedforward neural network.
 fixed. Figure 12a,b depicts the predicted efficiency curves, actual curves, and error curves
for SOC segment charging efficiency under 0 °C_0.5 C, 0 °C_1 C, and 0 °C_2 C conditions,
Batteries 2024, 10, x FOR PEER REVIEW 15 of 18
respectively. The average relative error for efficiency prediction using this approach is
0.29%. Under this optimal hyperparameter configuration, the neural network accurately
Batteries 2024, 10, 350 predicted the battery charging efficiency for the 5% SOC charging segment, confirming 15 of 17
the superiority
indicating of the
that the feedforward
network neural
performs network. 48 multiplication and addition oper-
approximately
ations during each forward pass. The space complexity is O(5 × 8 + 8 × 1 + 8 + 1) = O(57),
meaning that the network requires approximately 57 storage units for weights and biases.
Given the network’s small size, the computational and storage demands per input remain
fixed. Figure 12a,b depicts the predicted efficiency curves, actual curves, and error curves
for SOC segment charging efficiency under 0 °C_0.5 C, 0 °C_1 C, and 0 °C_2 C conditions,
respectively. The average relative error for efficiency prediction using this approach is
0.29%. Under this optimal hyperparameter configuration, the neural network accurately
predicted the battery charging efficiency for the 5% SOC charging segment, confirming
the superiority of the feedforward neural network.

Figure 11.
Figure 11. Prediction
Prediction accuracy
accuracy score
score of
of hyperparameter
hyperparameter configuration:
configuration: (a)
(a) random
random and
and (b)
(b) Nguyen–
Nguyen–
Widrow.
Widrow.

Table 2. Optimal efficiency prediction hyperparameter configuration.

Optimal Structure of Neural Network

Optimization algorithm Scaled conjugate gradient
Activation function Sigmoid
Initialization method Random
Number of neurons 8
Figure 11. Prediction accuracy
Neural network score of hyperparameter configuration:
structure Feedforward(a) random and (b) Nguyen–
neural network
Widrow.

Figure 12. Under charging conditions of 0 °C _0.5 C, 0 °C _1 C, and 0 °C _2 C (a) predicted charging
segment efficiency of the battery’s SOC based on the optimal neural network hyperparameter con-
figuration; (b) relative error bar chart.

Figure 12. Under charging conditions of 0 °C _0.5 C, 0 °C _1 C, and 0 °C _2 C (a) predicted charging
Figure 12. Under charging conditions of 0 ◦ C _0.5 C, 0 ◦ C _1 C, and 0 ◦ C _2 C (a) predicted charging
segment efficiency of the battery’s SOC based on the optimal neural network hyperparameter con-
segment efficiency of the battery’s SOC based on the optimal neural network hyperparameter
figuration; (b) relative error bar chart.
configuration; (b) relative error bar chart.

Finally, the neural network was integrated into an STM32F407 embedded system,
which operates at a clock frequency of 168 MHz and has 192 KB of memory. The neural
Batteries 2024, 10, 350 16 of 17

network model was converted into the TensorFlow Lite format and deployed onto the de-
velopment board. Upon completing the training of the network, the established prediction
model was able to perform real-time predictions. The experimental results demonstrate
that the parameter generation and calculation process of the efficiency prediction model, de-
veloped in this study, takes approximately 0.034 ms on the embedded system. This process
can run within the BMS program, facilitating online predictions in the BMS environment.

4. Conclusions and Final Discussion

The internal energy of the battery is affected by changes in temperature, which di-
rectly impact its OCV. This energy component cannot be ignored when calculating battery
efficiency. In this study, we derived the battery charging efficiency using relevant formu-
las from electrochemical and internal resistance models. The results indicate that as the
deviation between the battery temperature and the reference temperature increases, the
discrepancy between the actual OCV and the experimentally measured OCV becomes more
pronounced. After correcting for OCV, the calculated charging efficiency improved by
approximately 1% compared to the uncorrected values. The five-input, single-output neural
network model developed in this study accurately predicted charging efficiency in 5% SOC
segments, with an average relative error of 0.29%. This demonstrates the effectiveness of
the feedforward neural network in predicting battery efficiency when the hyperparameters
are appropriately configured. Finally, we integrated the prediction model into an embed-
ded system. Our battery efficiency prediction method provides a foundation for further
research into heat generation predictions. Future work will focus on real-time prediction of
heat generation in battery modules/systems across various time scales, transmitting the
predicted data to the cooling system. This enhancement will improve thermal control in
battery modules/systems, reduce temperature variation, and increase both temperature
control accuracy and cooling system efficiency.

Author Contributions: Conceptualization, W.Z. and Z.W. (Zhenpo Wang); Methodology, X.D.;
Software, X.D. and C.Y.; Validation, C.Y., C.G. and X.D.; Data Curation, J.S. and Z.W. (Zhihao Wang);
Writing—Original Draft Preparation, C.Y. and C.G.; Writing—Review and Editing, Z.A. and Q.L.;
Supervision, Y.B. All authors have read and agreed to the published version of the manuscript.
Funding: This research received no external funding.
Data Availability Statement: The original contributions presented in this study are included in this
article; further inquiries can be directed to the corresponding author.
Conflicts of Interest: The authors Zhenjia An and Qiang Liu were employed by the company
Longruisanyou New Energy Automobile Technology Co., Ltd. The remaining authors declare that
this research was conducted in the absence of any commercial or financial relationships that could be
construed as potential conflicts of interest.

References
1. Yao, L.; Aziz, J.; Kong, P.; Idris, N.; Alsofyani, I. Modeling of lithium titanate battery for charger design. In Proceedings of the 2014
Australasian Universities Power Engineering Conference (AUPEC), Perth, WA, Australia, 28 September–1 October 2014; pp. 1–5.
[CrossRef]
2. Yue, Y. Study on Fast Charging Method of Lithium Titanate Battery. In Proceedings of the 2023 8th Asia Conference on Power
and Electrical Engineering (ACPEE), Tianjin, China, 14–16 April 2023; pp. 549–553. [CrossRef]
3. Wang, Z.; Zhou, X.; Zhang, W.; Sun, B.; Shi, J.; Huang, Q. Parameter sensitivity analysis and parameter identifiability analysis of
electrochemical model under wide discharge rate. J. Energy Storage 2023, 68, 107788. [CrossRef]
4. Zhou, X.; Wang, Z.; Zhang, W.; Sun, B.; Su, X.; Shi, J.; Huang, Q. Construction of simplified impedance model based on
electrochemical mechanism and identification of mechanism parameters. J. Energy Storage 2024, 76, 109673. [CrossRef]
5. Liu, S.; Jiang, J.; Shi, W.; Ma, Z.; Wang, L.; Guo, H. Butler–volmer-equation-based electrical model for high-power lithium titanate
batteries used in electric vehicles. IEEE T. Ind. Electron. 2015, 62, 7557–7568. [CrossRef]
6. Ahmadi, L.; Fowler, M.; Young, S.; Fraser, R.; Gaffney, B.; Walker, S. Energy efficiency of Li-ion battery packs re-used in stationary
power applications. Sustain. Energy Technol. 2014, 8, 9–17. [CrossRef]
7. Xiao, J.; Li, Q.; Bi, Y.; Cai, M.; Dunn, B.; Glossmann, T.; Liu, J.; Osaka, T.; Sugiura, R.; Wu, B.; et al. Understanding and applying
coulombic efficiency in lithium metal batteries. Nat. Energy 2020, 5, 561–568. [CrossRef]
Batteries 2024, 10, 350 17 of 17

8. Lin, Z.; Li, D.; Zou, Y. Energy efficiency of lithium-ion batteries: Influential factors and long-term degradation. J. Energy Storage
2023, 74, 109386. [CrossRef]
9. Yang, F.; Wang, D.; Zhao, Y.; Tsui, K.; Bae, S. A study of the relationship between coulombic efficiency and capacity degradation
of commercial lithium-ion batteries. Energy 2018, 145, 486–495. [CrossRef]
10. Zhu, W.; Zhu, Y.; Davis, Z.; Tatarchuk, B. Energy efficiency and capacity retention of Ni–MH batteries for storage applications.
Appl. Energy 2013, 106, 307–313. [CrossRef]
11. Liu, C.; Wang, S.; Zhang, C.; Fu, H.; Nan, X.; Yang, Y.; Cao, G. High power high safety battery with electrospun Li3V2(PO4)3
cathode and Li4Ti5O12 anode with 95% energy efficiency. Energy Stor. 2016, 5, 93–102. [CrossRef]
12. Li, K.; Tseng, K. Energy efficiency of lithium-ion battery used as energy storage devices in micro-grid. In Proceedings of
the IECON 2015-41st Annual Conference of the IEEE Industrial Electronics Society, Yokohama, Japan, 9–12 November 2015;
pp. 005235–005240.
13. Lai, X.; Zhou, L.; Zhu, Z.; Zheng, Y.; Sun, T.; Shen, K. Experimental investigation on the characteristics of coulombic efficiency of
lithium-ion batteries considering different influencing factors. Energy 2023, 274, 127408. [CrossRef]
14. Kang, J.; Yan, F.; Zhang, P.; Du, C. Comparison of comprehensive properties of Ni-MH (nickel-Metal Hydride) and Li-ion
(lithium-ion) batteries in terms of energy efficiency. Energy 2014, 70, 618–625. [CrossRef]
15. Wang, S. Entropy and heat generation of lithium cells/batteries. Chin. Phys. B 2016, 25, 010509. [CrossRef]
16. Jalkanen, K.; Aho, T.; Vuorilehto, K. Entropy change effects on the thermal behavior of a LiFePO4 /graphite lithium-ion cell at
different states of charge. J. Power Sources 2013, 243, 354–360. [CrossRef]
17. Doh, H.; Ha, Y.; Eom, S. Entropy measurement of a large format lithium ion battery and its application to calculate heat generation.
Electrochim. Acta 2019, 309, 382–391. [CrossRef]
18. Lai, Y.; Du, S.; Ai, L.; Ai, L.; Cheng, Y.; Tang, Y.; Jia, M. Insight into heat generation of lithium ion batteries based on the
electrochemical-thermal model at high discharge rates. Int. J. Hydrogen Energy 2015, 40, 13039–13049. [CrossRef]
19. Shi, J.; Zhang, W.; Bao, Y.; Gao, D.; Wang, Z. Load Forecasting of Electric Vehicle Charging Stations: Attention Based Spatiotempo-
ral Multi-Graph Convolutional Networks. IEEE Trans. Smart Grid 2023, 15, 3016–3027. [CrossRef]
20. Nascimento, R.; Corbetta, M.; Kulkarni, C.; Viana, F. Hybrid physics-informed neural networks for lithium-ion battery modeling
and prognosis. J. Power Sources 2021, 513, 230526. [CrossRef]
21. Wang, F.; Zhi, Q.; Zhao, Z.; Zhai, Z.; Liu, Y.; Xi, H.; Wang, S.; Chen, X. Inherently Interpretable Physics-Informed Neural Network
for Battery Modeling and Prognosis. IEEE Trans. Neural Netw. Learn. Syst. 2023, 1–15. [CrossRef]
22. Dong, C.; Wang, G. Estimation of power battery SOC based on improved BP neural network. In Proceedings of the 2014 IEEE
International Conference on Mechatronics and Automation, Tianjin, China, 3–6 August 2014; pp. 2022–2027. [CrossRef]
23. Wang, Y.; Chen, X.; Li, C.; Yu, Y.; Zhou, G.; Wang, C.; Zhao, W. Temperature prediction of lithium-ion battery based on artificial
neural network model. Appl. Therm. Eng. 2023, 228, 120482. [CrossRef]
24. Liu, Y.; Zhang, L.; Jiang, J.; Wei, S.; Liu, S.; Zhang, W. A Data-Driven Learning-Based Continuous-Time Estimation and Simulation
Method for Energy Efficiency and Coulombic Efficiency of Lithium Ion Batteries. Energies 2017, 10, 597. [CrossRef]
25. Xiong, R.; Cao, J.; Yu, Q.; He, H.; Sun, F. Critical review on the battery state of charge estimation methods for electric vehicles.
IEEE Access 2017, 6, 1832–1843. [CrossRef]
26. Chen, Y.; Yang, G.; Zhu, H.; Sun, X. An open circuit voltage and internal resistance estimation method of lithium-ion batteries
with constant current tests. Trans. China Electrotech. Soc. 2018, 33, 3976–3988.
27. Liu, X.; Chen, Z.; Zhang, C.; Wu, J. A novel temperature-compensated model for power Li-ion batteries with dual-particle-filter
state of charge estimation. Appl. Energy 2014, 123, 263–272. [CrossRef]
28. Chen, L.; Zhang, M.; Ding, T.; Wu, S.; Li, Y.; Liang, G.; Li, H.; Pan, H. Estimation the internal resistance of lithium-ion-battery
using a multi-factor dynamic internal resistance model with an error compensation strategy. Energy Rep. 2021, 7, 3050–3059.
[CrossRef]
29. Plett, G. Extended Kalman filtering for battery management systems of Li PB-based HEV battery packs, Part 2. Modeling and
identification. J. Power Sources 2004, 134, 262–276. [CrossRef]
30. Hu, Y.; Wang, Z. Study on SOC Estimation of Lithium Battery Based on Improved BP Neural Network. In Proceedings of the 2019
8th International Symposium on Next Generation Electronics (ISNE), Zhengzhou, China, 9–10 October 2019; pp. 1–3. [CrossRef]
31. Shrestha, A.; Mahmood, A. Review of Deep Learning Algorithms and Architectures. IEEE Access 2019, 7, 53040–53065. [CrossRef]
32. Sheng, L.; Zhang, Z.; Su, L.; Zhang, H.; Zhang, H.; Li, K.; Fang, Y.; Ye, W. Experimental and numerical approach for analyzing
thermal behaviors of a prismatic hard-cased lithium-ion battery. J. Energy Storage 2021, 35, 102313. [CrossRef]
33. Foresee, F.; Hagan, M. Gauss-Newton approximation to Bayesian learning. In Proceedings of the International Conference on
Neural Networks (ICNN’97), Houston, TX, USA, 9–12 June 1997; Volume 3, pp. 1930–1935.

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