Charging and discharging model of lithium-ion battery for charge equalization control
using particle swarm optimization algorithm
M. M. Hoque, M. A. Hannan, and A. Mohamed
Citation: J. Renewable Sustainable Energy 8, 065701 (2016); doi: 10.1063/1.4967972
View online: http://dx.doi.org/10.1063/1.4967972
View Table of Contents: http://aip.scitation.org/toc/rse/8/6
Published by the American Institute of Physics
JOURNAL OF RENEWABLE AND SUSTAINABLE ENERGY 8, 065701 (2016)
Charging and discharging model of lithium-ion battery
for charge equalization control using particle swarm
optimization algorithm
M. M. Hoque,1,2,a) M. A. Hannan,3,a) and A. Mohamed1
1
Department of Electrical, Electronics and Systems Engineering, Universiti Kebangsaan
Malaysia (UKM), Bangi, 43600 Selangor, Malaysia
2
Department of Applied Physics, Electronics and Communication Engineering,
University of Chittagong, Chittagong 4331, Bangladesh
3
Department of Electrical Power Engineering, College of Engineering,
Universiti Tenaga Nasional, Kajang, 43000 Selangor, Malaysia
(Received 8 June 2016; accepted 4 November 2016; published online 15 November 2016)
This study presents a charging and discharging controller of a lithium-ion battery
for charge equalization control of a battery storage system using the particle
swarm optimization (PSO) algorithm. The charge equalization controller is
designed using a bidirectional flyback DC–DC converter for exchanging the
amount of energy from a battery series stack to an overdischarged cell to be
charged and vice versa. The constant current–constant voltage charge proportio-
nal–integral (PI) control and discontinuous current mode control are applied to
charge and discharge the lithium-ion battery on a flyback converter operation.
This proposed system utilizes the PSO algorithm to optimize the values of the PI
controller parameters. Optimization results produce the ideal values of the PI con-
troller parameters with minimum error indices, thereby regulating the pulse-width
modulation to the MOSFET switching drive of the flyback converter and upgrad-
ing the battery charge performance for charge equalization. The PSO algorithmic
approach-based developed system is proven to be robust and competent for high-
tech storage systems toward the advancement of sustainable electric vehicle tech-
nologies and renewable source of applications. Published by AIP Publishing.
[http://dx.doi.org/10.1063/1.4967972]
NOMENCLATURE
C charge/discharge rate
CC constant current
CV constant voltage
DC duty control
DCM discontinuous current mode
EV electric vehicle
Ki integral constant
Kp proportional constant
HEV hybrid electric vehicle
MAE mean absolute error
MSE mean square error
OCV open circuit voltage
PI proportional integral
PSO particle swarm optimization
a)
Authors to whom correspondence should be addressed. Electronic addresses: [email protected] and [email protected]
1941-7012/2016/8(6)/065701/16/$30.00 8, 065701-1 Published by AIP Publishing.
065701-2 Hoque, Hannan, and Mohamed J. Renewable Sustainable Energy 8, 065701 (2016)
PWM pulse width modulation
RMSE root mean square error
SD standard deviation
SOC state of charge
TE trial & error
I. INTRODUCTION
Electric vehicles (EVs) are currently regarded as green and energy efficient modes of trans-
portation. Accordingly, energy storage and management challenges are needed to be developed
toward the sustainable development of EVs.1–3 Most EVs, such as hybrid EVs (HEVs), plug-in
HEVs, and battery EVs, are designed using a lithium-ion battery powered storage system that
offers high densities of capacity and power, low memory effect, long life span, high terminal
voltage, and low discharge rate compared with other batteries.4–6 However, lithium-ion batteries
are costly and need an optimal charging system and charge equalization control among a series
connected battery cells. Moreover, lithium-ion batteries require protection from overcharging
and overdischarging.4,7 Consequently, the price of EVs is compromised because of the enor-
mous production and the use of lithium-ion batteries for EV storage application.8
A series stack of lithium-ion battery cells with high power and voltage ratings are needed
to power an EV’s drive system.7 This battery storage system loses its charge balancing proper-
ties when charging and discharging after several charge–discharge cycles. This imbalance con-
dition results in the degradation of the storage system’s quality, as well as the reduction of
battery life and overall performance.4,9–11 The charge equalization control may equalize the
charge levels among the battery cells and protect the battery from overcharging and overdi-
scharging, thereby enhancing its life and storage efficiency.12,13
The efficient charging and discharging control for lithium-ion batteries is required during
charge equalization for long durability and good performance.14,15 Thus, the constant current–
constant voltage (CC–CV) control,16–18 reflex control,19 and pulse control20 are applied in
charging lithium-ion batteries.21 The CC–CV control with the flyback DC–DC converter-based
battery charging is promising because of the galvanic isolation between the output and input
circuits, extensive power range, low cost, and simple design.17,22 The CC–CV control in the fly-
back converter is implemented to charge the battery at constant current using closed loop duty
control (DC) in charging lithium-ion batteries. The charging remains at constant voltage until
the current reaches a defined value using closed loop proportional-integral (PI) control.7,17,18
The discharge control of lithium-ion batteries for charge equalization is commonly executed in
the discontinuous current mode (DCM) with constant duty control.7,18
This study proposes a single stage flyback DC–DC converter for the charge equalization of
the lithium-ion battery by charging the overdischarged cell with an optimal CC–CV charge PI
controller based on the particle swarm optimization (PSO) algorithm, as well as by discharging
the overcharged cell with the DCM control. The PSO algorithm based optimal CC–CV charge
PI controller for charging the lithium-ion battery contributes to produce better and more effi-
cient charging results than that of DC and trial-and-error (TE) PI controls.17 The target level of
charging and discharging of the lithium-ion battery is assigned to the average voltage/charge
level of the battery stack because it is applied for charge equalization.
The remainder of this paper is organized as follows. Section II presents the battery charge
equalization control model, which includes the charging and discharging model of the lithium-
ion battery, flyback DC–DC converter, and charge–discharge process models. Section III nar-
rates the charge and discharge control of charge equalization using the PSO algorithm with
system descriptions, PI controller design, PSO algorithm, PSO algorithm approach optimal
CC–CV charge PI control process, and optimization limitation. Section IV analyzes the results
to provide the improved outcome. Finally, Section V contains the concluding remarks on the
PSO algorithm approach lithium-ion battery charging for charge equalization as well as presents
the direction of future research.
065701-3 Hoque, Hannan, and Mohamed J. Renewable Sustainable Energy 8, 065701 (2016)
II. BATTERY CHARGE EQUALIZATION CONTROL
The block diagram of the battery charge equalization control is shown in Figure 1. The
proposed system mainly focuses on the charge and discharge control of the lithium-ion battery
for charge equalization. The lithium-ion battery charging and discharging circuit for the charge
equalization control entails designing an optimal lithium-ion battery, flyback DC–DC converter,
and charge and discharging control models. This section presents the design model of the pro-
posed system to achieve efficient charge and discharge of the lithium-ion battery for the charge
equalization of the series-connected battery cells in the EV applications. Section III presents the
model description and process algorithm.
A. Charging and discharging model of lithium-ion battery
A lithium-ion battery of 15.5 Ah and 3.7 V nominal is selected for the EV storage applica-
tion. The modified Shepherd model equation is used to develop the charge and discharge model
of the lithium-ion battery. This model has no algebraic loop problem and may be modified with
battery specifications for another battery model development. However, this battery model is
considered to be operated at normal temperature. In fact, the proposed control model is imple-
mented to charge and discharge the lithium-ion battery for charge equalization; the heat man-
agement feature for temperature control might be added in the battery management system.
Equations (1) and (2) present the charge and discharge models, respectively, of the lithium-ion
battery7,23
Q Q
VB ch ¼ E0 K i K it þ AeBit i R; (1)
it þ 0:1Q Q it
Q Q
VB disch ¼ E0 K i K it þ AeBit i R; (2)
Q it Q it
where VB disch is the battery discharge voltage, VB ch is the battery charge voltage, E0 is the con-
stant voltage (V), R is the internal resistance (X), K is the polarization constant (Ah1) Ðor resis-
t
tance (X), i is the battery current (A), it is the extracted battery capacity (Ah) (i.e., 0 iðtÞdt),
Q is the maximum battery capacity (Ah), A is the exponential voltage (V), B is the exponential
capacity (Ah1), and i* is the low frequency current (A).
The open circuit voltage (OCV) characteristic of the lithium-ion battery model is shown in
Figure 2.23 The relationship between the OCV and the state of charge (SOC) (see Equation (3))
is formed by the curve fitting approximation of the lithium-ion battery characteristics. This rela-
tionship is utilized in estimating the charge level of the lithium-ion battery during the charge
equalization operation24–27
VOC ¼ 21:049 SOC5 57:837 SOC4 þ 62:228 SOC3
32:997 SOC2 þ 8:9149 SOC þ 2:824: (3)
FIG. 1. Battery charge equalization control block.
065701-4 Hoque, Hannan, and Mohamed J. Renewable Sustainable Energy 8, 065701 (2016)
FIG. 2. OCV versus SOC characteristics of the lithium-ion battery model.
B. Flyback DC-DC converter
The optimal design of the flyback DC–DC converter is necessary for the proposed charge
equalization system in charging and discharging the lithium-ion battery. The converter carries
the equalization current to exchange the desired amount of energy from the overdischarged cell
to the battery stack and vice versa. Consistent with the flyback converter, energy is stored in
the primary coil during the ON stage of the converter MOSFET switch (0 < t DT).
Thereafter, the stored energy is transferred to the secondary coil during the OFF stage of the
MOSFET switch and the ON condition of the rectifier power diode (DT < t T). This system
operates the flyback converter in the DCM mode for forward access and discharge control, as
well as in the CC–CV charge PI control for feedback operation. The maximum value of the
switching duty, Dm, is calculated using Equation (4). The electric components of the flyback
converter, such as the mutual inductance, Lm, transformation ratio, n, and filter capacitor, Cf, of
the converter, are computed using Equations (5)–(7), respectively,7,17,18,22
Vds nom Vin min
Dmax ¼ ; (4)
Vds min þ Vds nom Vin max
ðg Vin min Dmax Þ2
Lm ¼ ; (5)
2 Pin f
N1 g Vin min Dmax
n¼ ¼ ; (6)
N2 ðV0 Vf Þð1 Dmax Þ
g Pin Dmax
Cf ¼ ; (7)
ðV0 Vf Þ2 f
where Vin_min is the minimum voltage for the battery cell pack; Vin_max is the maximum voltage
for the battery pack; Vds_min is the minimum voltage stress of the switch; Vds_nom is the nominal
voltage stress of the switch (30% margin of the maximum drain–source voltage); Pin is the
input power; f is the switching frequency; N1 and N2 are the number of primary and secondary
turns, respectively; V0 is the output voltage; and Vf is the forward voltage of the rectifier power
diode. g is the converter efficiency, which regulates the required converter output power by set-
ting the values of mutual inductance, transformation ratio, and filter capacitor of the power
converter.
065701-5 Hoque, Hannan, and Mohamed J. Renewable Sustainable Energy 8, 065701 (2016)
The total power loss and efficiency of the converter are obtained using Equations (8)
and (9), respectively. Equations (10) and (11) define the maximum voltage stresses in the pri-
mary and secondary sides of the converter, respectively,18
PCL Total ¼PL DS þ PL SW þ PL Pri þ PL Sec þ PL D þ PL Lm ; (8)
Pout Pin PCL Total
g¼ ¼ ; (9)
Pin Pin
VS DS max ¼Vin max þ n Vout ; (10)
Vin max
VS D max ¼ þ Vout ; (11)
n
where PCL_Total is the total converter power loss, PL_DS is the MOSFET conduction loss, PL_SW
is the switching loss, PL_Pri is the primary winding loss, PL_Sec is the secondary winding loss,
PL_D is the diode conduction loss, PL_Lm is the mutual inductor resistance loss, VS_DS_max is the
maximum voltage stress on the MOSFET switch, Vin_max is the maximum input in primary
side, Vout is the output voltage, and VS_D_max is the maximum voltage stress on the diode.
C. DCM control
The flyback DC–DC converter primarily processes the power transmission as a forward
process during charging, as well as a discharging process of the lithium-ion battery in the DCM
control mode. The flyback converter in the DCM control operates in the three states of the
MOSFET switch, namely, the ON, OFF, and RESET states. The key waveforms of the DCM
control are shown in Figure 3. During the ON state, the currents in the primary side and the
magnetizing coil of the transformer increase with a slope as defined by Equation (12). The volt-
age across the magnetizing inductor is equal to the input. Energy is stored in the primary coil
without conduction to the secondary side of the transformer with a voltage stress to the diode,
as defined by Equation (11)17,18
di Vin
¼ : (12)
dt Lm
During the OFF state, the diode conducts and the stored energy is transferred to the sec-
ondary circuit to charge the battery. Thereafter, the primary current drops to zero. The current
through the magnetizing coil decreases with a negative slope (see Equation (13)). The voltage
stress of the MOSFET switch is defined by Equation (10). The secondary current peaks at the
primary peak current level multiplied by the transformation ratio and then falls with a negative
slope (see Equation (14))17,18
di n Vout
¼ ; (13)
dt Lm
di n2 Vout
¼ : (14)
dt Lm
During the RESET state of the MOSFET switch, the primary and secondary sides of the
transformer are not connected; thus, the magnetizing coil neither stores nor delivers energy.
The voltage stresses to the switch and diode are the same as their side values.17,18
D. CC-CV control
The CC–CV control is highly effective for charging the lithium-ion battery because of the
depth of charging for a long discharge cycle without the memory effect.7,17 This control
065701-6 Hoque, Hannan, and Mohamed J. Renewable Sustainable Energy 8, 065701 (2016)
FIG. 3. Key waveforms of the DCM control of the flyback DC–DC converter.
regulates the lithium-ion battery charging process in such a way that the battery is charged to a
desired level for charge equalization with constant current (see Figure 4). Thereafter, the charg-
ing process remains with constant voltage until the current decreases to approximately 50% of
its normal rate or to any specified level.7,17
The empty battery should be charged using the trickle current charge control method until
the battery voltage reaches the threshold level. Otherwise, the internal impedance of the battery
may be affected by SOC and temperature.17 This study proposes a PSO algorithm-based PI con-
troller for the CC–CV control to charge the lithium-ion battery using a charge equalization
application. This study may minimize the error of feedback control and produce accurate pulse-
width modulation (PWM) for the converter MOSFET switching.
III. CHARGE AND DISCHARGE OF THE CHARGE EQUALIZATION USING THE PSO
ALGORITHM
The proposed system comprises a series-connected lithium-ion battery stack, flyback
DC–DC power converter, DCM controller, PSO algorithm-based CC–CV charge PI controllers,
and a select switch for the application of the battery charge equalization (see Figure 5). A bat-
tery cell from the stack that is detected as overdischarged during charge equalization should be
charged to reach the average charge level of the entire battery stack.7,13,15 The power converter
enables the equalization/charging current to transfer the required energy from the battery stack
065701-7 Hoque, Hannan, and Mohamed J. Renewable Sustainable Energy 8, 065701 (2016)
FIG. 4. CC–CV control mechanism for the lithium-ion battery charging.
to the detected overdischarged cell by establishing an electrical connection between the con-
verter and the detected cell. Thereafter, the charge equalization starts to charge the cell through
the PSO algorithm-based PI controller for the CC charge. First, the charging continues until the
cell’s charge reaches the average charge level of the battery stack. Thereafter, the CV charge
FIG. 5. PSO algorithm-based proposed lithium-ion battery charging and discharging for charge equalization.
065701-8 Hoque, Hannan, and Mohamed J. Renewable Sustainable Energy 8, 065701 (2016)
control is activated. A select switch controls the flow of the PWM signal for the CC and CV
charges. The PSO algorithm optimizes the parameters of the PI controller by minimizing the
error in the operations of the CC and CV charges. By contrast, the discharging of the lithium-
ion battery for charge equalization is done by the DCM control of the flyback power converter
when a detected overcharged cell from the stack needs to transfer its excess energy to the bat-
tery stack.7,17 The proposed charge control with charge equalization efficiently implements the
charging, discharging, and charge equalization processes. This process has no energy dissipation
and heat issues. The control algorithm and process may be made on demand. The controller
may be utilized as charge equalization control for any range of power and voltage constraints
in various fields.
A. PI controller model
The PI controller is universally well-known for closed loop control, which is flexible in
terms of operation and ease of tuning. This controller is applied to stabilize the output of the
process system with minimal error for any change of process system behavior. Equation (15) in
the discrete form and Equation (16) in the z-transform define the model characteristics of the PI
controller for digital implementation28
X
k
ek þ ek1
uk ¼ Kp ek þ Ki ei ¼ Kp ek þ Ki uk1 þ Ts ; (15)
i¼1
2
0 1
Ts Ts
ðz þ 1ÞTs @ Kp þ Ki 2 z Kp þ Ki 2 A
Uz ¼ Kp Ez þ Ki Ez ¼ Ez ; (16)
2ðz 1Þ z1
where u is the control input to process plant; Kp and Ki are the proportional and integral gains,
respectively; e is the tracking error, which is the difference between the desired and actual val-
ues; Ts is the sample rate; and U and E are the parameters in the z-transform corresponding to
u and e, respectively.
B. PSO algorithm
The PSO algorithm is a stochastic optimization algorithm that is contingent on population as
a swarm29 and is applied to iteratively optimize a candidate solution as a particle of a given prob-
lem.30–32 This algorithm searches the optimal solution of swarm particles through the particle’s
movement in the searchspace with its optimized position and velocity. The PSO algorithm opti-
mizes the solution by obtaining the minimum value of the objective function through an iteration
process. Moreover, the PSO algorithm uses three quantities, namely, swarm size (N), iteration
number (T), and dimension (D). Its process may be easy and short by making the searchspace
finite and fixing the D intervals and a few learning and weight factors. This algorithm has a faster
convergence property than other global optimization algorithms.33–36 The proposed system
charges the lithium-ion battery using the PSO algorithm approach optimal CC–CV charge PI con-
trol for charge equalization in a series battery stack. The values of the PI controller parameters
(i.e., proportional constant (Kp) and integral constant (Ki)) are optimized in this process.
1. PSO for the optimal CC-CV charge PI control process
The charge equalization controller identifies a battery cell as overdischarged during the
charge equalization of the series lithium-ion battery stack. Thereafter, the system enables the
detected cell to be charged with the amount of charge needed to reach the average charge of the
battery stack through the flyback DC–DC converter based on the proposed charge control algo-
rithm. The PSO algorithm approach CC–CV charge optimal PI control flow process is depicted
in Figure 6. The entire process is executed to set the N, T, D, and space range qualities. The
swarm positions are randomly generated using the parameters’ boundary range, and the initial
065701-9 Hoque, Hannan, and Mohamed J. Renewable Sustainable Energy 8, 065701 (2016)
FIG. 6. Proposed PSO-based process algorithm for lithium-ion battery charging.
pbest is set. Thereafter, the fitness values are calculated for each particle using the objective func-
tion (i.e., mean absolute error (MAE)) (see Equation (17)). The error value is the difference
between the feedback voltage/current value (Yi) and the respective reference value (Yr). The PSO
algorithm uses the error value in the form of MAE and produces optimal values of PI controller
parameters by obtaining a minimum MAE value through a number of iterations. The iteration
065701-10 Hoque, Hannan, and Mohamed J. Renewable Sustainable Energy 8, 065701 (2016)
then begins to compute the best fitness value, to determine the best pbest among all particles’
pbest, as well as to assign it as gbest, where “pbest” is the local best value in each iteration and
“gbest” is the global best value after maximum iteration. Subsequently, the particles’ velocities
are calculated and the swarm positions are updated based on Equations (18) and (19). The swarm
positions are verified whether they are within the search space. Thereafter, the fitness values are
recalculated for each particle using the objective function. The improved fitness value is com-
pared with previous values, and pbest is updated constantly until the maximum iteration number
is reached. Finally, the minimum fitness value of the objective function is selected with the best
swarm position when the fitness value reaches at its minimal value. Otherwise, the optimization
requires more iteration to obtain the minimal fitness value and the best particles. The optimal
parameter values of Kp and Ki are outputted to the PI controller.28,30–35 Thus, the CC–CV charge
control is executed for charging the lithium-ion battery and performing charge equalization with
the best and optimized PI controller parameter values
1X M
MAE ¼ jYr Yi j; (17)
M i¼1
vkþ1 ¼ w vk þ c1 r1k ðpbestk xk Þ þ c2 r2k ðgbestk xk Þ; (18)
xkþ1 ¼ xk þ vkþ1 ; (19)
where Yr is the reference value; Yi is the true value; M is the total number of samples; vk þ 1 is
the new velocity of the particle; vk is the present velocity of particle; xk is the present swarm
position; xk þ 1 is the updated swarm position; c1 and c2 are the learning factors, respectively;
w is the weight factor; and r1 and r2 are the random numbers between (0, 1).
2. Optimization limitation
For the PI controller parameter optimization, the minimum and maximum ranges of Kp and
Ki as the D intervals are fixed to define the minimum search-space. The ranges reduce the num-
ber of iterations through the determination of the possible ranges of Kp and Ki using the trial-
and-error method. The swarm particle values are verified frequently whether they cross the
search-space limit. Otherwise, the optimization may deviate from the divergence. The simula-
tion model is simplified when the PSO algorithm is running to reduce the optimization time.
Although the optimized results present a good outcome, the ideal result is not attained because
of the random generation of particle values with a few mathematical formulas.
IV. RESULTS AND DISCUSSIONS
The PSO algorithm is implemented to charge the lithium-ion battery in charge equalization
application of the series-connected 90 lithium-ion battery cells of 15.5 Ah and 3.7 V nominal. The
DCM control for discharging the lithium-ion battery is also implemented. The CC–CV charge con-
trol is employed when charging the lithium-ion battery using the PI controller with optimized Kp
and Ki parameters by means of the PSO algorithm. Table I shows that the values of the flyback
converter parameters for charging and discharging are determined by considering the input power
of 10 W and switching the frequency of 1000 Hz. All parts are connected together with the optimal
parameter values (see Figure 5 in Section III) to charge and discharge the lithium-ion battery. The
results of the proposed system’s performance are discussed in Subsections IV A–IV D.
A. Lithium-ion battery model
The voltage profiles of the lithium-ion battery model at different pulse current charge rates
are shown in Figure 7. These profiles are measured with a pulse current of 1C, C/3, and C/5
charge and discharge rates for 600 s at 600 s intervals. The waveforms in Figure 7 show the uni-
form characteristics in all charge and discharge profiles. In the pulse charge/discharge rate, it is
065701-11 Hoque, Hannan, and Mohamed J. Renewable Sustainable Energy 8, 065701 (2016)
TABLE I. Design parameters of the flyback DC–DC converter.
Type Parameter Value
Input power, Pin 10 W
Switching frequency, f 1000 Hz
Sampling time, Ts 1 104 s
Charging converter Transformation ratio, N1:N2 110:5
Mutual inductance, Lm 1.68 103H
Filter capacitance, C1 5.0 103F
Maximum MOSFET switch stress, VS_DS_max 833 V
Maximum rectifier diode stress, VS_D_max 40 V
Discharging converter Transformation ratio, N1:N2 5:110
Mutual inductance, Lm 2.35 106H
Filter capacitance, C2 2.0 103F
Maximum MOSFET switch stress, VS_DS_max 47 V
Maximum rectifier diode stress, VS_D_max 690 V
proved that the lithium-ion battery model performs without any major change of chemicals and
memory effect.
B. PI controller parameters optimization results
The CC–CV charge PI controller parameters are optimized by means of the PSO algorithm
(see Figure 5 in Section III). The result obtains the minimum MAE value to improve the lithium-
ion battery charging. Thus, the equalization controller receives the regulatory PWM signal for the
CC–CV charge controls to drive the MOSFET switching drive in the flyback DC–DC converter.
The proposed PSO algorithm approach CC–CV charge PI control uses 50 iterations to attain the
best results and the minimum values of the objective function (see Figure 8). Table II presents
the performance results of the proposed algorithm. The optimization results show that the PSO
algorithm approach CC–CV charge control has the best results in achieving the minimum error
indices’ values of MAE, mean square error (MSE), root mean square error (RMSE), and standard
FIG. 7. Voltage profile of the lithium-ion battery model at different charge rates (a) and different charge rates (b).
065701-12 Hoque, Hannan, and Mohamed J. Renewable Sustainable Energy 8, 065701 (2016)
FIG. 8. Optimization response curves: CC objective function (a) and CV objective function (b).
deviation (SD). The results using the PSO algorithm are superior to that of the trial and error
(TE) values because the search space of the PSO algorithm is fixed with the boundary values of
PI controller parameters based on the trial and error values to obtain the best or optimal values.
C. Flyback DC-DC converter outputs
The flyback DC–DC converter carries the equalization current to transfer the required
energy from the battery stack to the overdischarged battery cell for charge equalization and
vice versa. The PSO algorithm approach CC–CV charge PI controller assists the optimally reg-
ulated PWM switching to the MOSFET switch drive, thereby obtaining the robustness of the
charge controller using the flyback DC–DC converter in the charge equalization application.
Figure 9 shows the outputs of the flyback DC–DC converter when charging the overdischarged
lithium-ion battery in the battery stack. The input voltage for the 90 series battery stack to the
converter is 345 V. The output current for charging is controlled at a constant 0.5 C rate of the
lithium-ion battery (i.e., 7.75 A). The voltage stress on the MOSFET switch and rectifier diode
is approximately 650 V and 32 V, respectively. The results of the converter show that charging
the lithium-ion battery for charge equalization is safe and robust.
D. Outputs of the optimized lithium-ion battery charging and discharging
Figure 10 presents the voltage and current responses during the charging of the lithium-ion
battery on charge equalization. The output characteristics of the battery charge are demonstrated
for a consecutive control of the CC and CV charges. The overdischarged lithium-ion battery of
3.71 V charges to the average voltage level of 3.85 V among the 90 series battery cells at a CC
charge rate of 0.5 C. Thereafter, the battery charges at a CV charge of 3.85 V until the current
reaches a specified level. Figures 10(a) and 10(b) show the voltage and current responses of the
lithium-ion battery charging at CC charge with the duty control and at CV charge with the TE-
based PI control, respectively. The voltage and current characteristics of the lithium-ion battery
TABLE II. Performance results of the PSO–CC and CV.
CC CV
Charging !/Error # DC PI (TE) PI (PSO) PI (TE) PI (PSO)
5
MAE 0.1350 0.0081 0.0075 6.6376 10 4.9481 105
MSE 0.0205 0.0026 0.0013 2.2341 108 1.6654 108
RMSE 0.1430 0.0508 0.0354 1.4947 104 1.2905 104
4
SD 0.1430 0.0508 0.0354 1.4947 10 1.2905 104
065701-13 Hoque, Hannan, and Mohamed J. Renewable Sustainable Energy 8, 065701 (2016)
FIG. 9. Flyback DC–DC converter outputs during the lithium-ion battery charging.
charging at CC charge with TE-based PI control and CV charge with TE-based PI control are
shown in Figures 10(c) and 10(d), respectively. The outputs of battery charging at CC and CV
charges are presented in Figures 10(e) and 10(f), respectively, by employing the PSO
algorithm-based optimized PI controller. The results prove that the PI controller that applied the
lithium-ion battery charging system is robust and efficient in the charge equalization applica-
tion. Table II shows that the PSO algorithm is employed with minimal error values of different
indices.
Figure 11 shows the outputs of the voltage [Figure 11(a)] and current [Figure 11(b)]
characteristics during the lithium-ion battery discharge in the DCM mode. The overcharged
FIG. 10. Outputs of the lithium-ion battery charging voltage response (a) and current response (b) at CC with duty con-
trol and CV with TE-based PI control; voltage response (c) and current response (d) at CC with TE-based PI control and
CV with TE-based PI control; voltage response (e) and current response (f) at CC and CV both with PSO-based PI
control.
065701-14 Hoque, Hannan, and Mohamed J. Renewable Sustainable Energy 8, 065701 (2016)
FIG. 11. Output of the lithium-ion battery discharging voltage response (a) and current response (b) in the DCM mode.
lithium-ion battery of 3.89 V discharges to the average voltage level of 3.85 V among the
90 series battery cells at the DCM control mode with a current value of approximately 0.2 C
(i.e., 3.1 A). The results ascertain the smooth discharging of the lithium-ion battery that can be
pragmatic in the application of charge equalization using the flyback DC–DC converter.
V. CONCLUSION
This study proposes a PSO algorithmic approach-based CC–CV charge PI controller for
charging and discharging the lithium-ion battery in charge equalization. The model of the
charge and discharge control system, which includes the lithium-ion battery model, flyback
DC–DC converter model, and DCM control and CC–CV control models, is utilized in this
study. Moreover, the proposed charge and discharge control of the charge equalization control-
ler using the PSO algorithm is explained. The developed PSO-based CC–CV charge PI control-
ler is utilized for the equalization control of the lithium-ion battery in power and voltage rating
applications. The results are explained on the basis of the performances of the developed sys-
tem. The optimization of the PI controller parameters for the CC and CV charge controls is
shown to obtain the best objective function with minimal MAE in 50 iterations. Moreover, the
characteristics of the flyback converter are presented in a good stress value within the tolerance
level. The voltage and current responses of charging the overdischarged and discharging the
overcharged lithium-ion battery cells show performance consistency, thereby proving the quality
of the developed charge equalization controller using the flyback DC–DC converter. The pro-
posed charge and discharge control for lithium-ion battery equalization application performs
with minimum error indices’ values and obtains better results than other controllers, which
leads to extended battery life. In this study, an optimal charging and discharging model of the
lithium-ion battery for charge equalization is developed. This control model will experience an
experimental test to evaluate the performance of the battery in terms of life cycle and effi-
ciency. The developed model along with the equalization controller also will be implemented
afterward in prototype as lab scale for testing and validation to apply in real EV and other
energy storage systems applications.
065701-15 Hoque, Hannan, and Mohamed J. Renewable Sustainable Energy 8, 065701 (2016)
ACKNOWLEDGMENTS
This work was supported by the Ministry of Science, Technology and Innovation, Malaysia
under Grant No. 06-01-02-SF1060, UKM Grant No. DIP-2015-012, and the Ministry of Higher
Education, Malaysia under the Commonwealth Scholarship and Fellowship Plan in part.
1
X. Hu, J. Jiang, B. Egardt, and D. Cao, “Advanced power-source integration in hybrid electric vehicles: Multicriteria opti-
mization approach,” IEEE Trans. Ind. Electron. 62, 7847–7858 (2015).
2
M. A. Hannan, F. A. Azidin, and A. Mohamed, “Hybrid electric vehicles and their challenges: A review,” Renewable
Sustainable Energy Rev. 29, 135–150 (2014).
3
X. Hu, N. Murgovski, L. M. Johannesson, and B. Egardt, “Optimal dimensioning and power management of a fuel cell/
battery hybrid bus via convex programming,” IEEE/ASME Trans. Mechatron. 20, 457–468 (2015).
4
E. Samadani, M. Mastali, S. Farhad, R. A. Fraser, and M. Fowler, “Li-ion battery performance and degradation in electric
vehicles under different usage scenarios,” Int. J. Energy Res. 40, 379–392 (2016).
5
A. Panday and H. O. Bansal, “Energy management strategy for hybrid electric vehicles using genetic algorithm,”
J. Renewable Sustainable Energy 8, 015701 (2016).
6
S. F. Schuster, M. J. Brand, P. Berg, M. Gleissenberger, and A. Jossen, “Lithium-ion cell-to-cell variation during battery
electric vehicle operation,” J. Power Sources 297, 242–251 (2015).
7
M. M. Hoque, M. A. Hannan, and A. Mohamed, “Voltage equalization control algorithm for monitoring and balancing of
series connected lithium-ion battery,” J. Renewable Sustainable Energy 8, 025703 (2016).
8
A. Cheung, Battery Innovation: Incremental or Disruptive? (Bloomberg New Energy Finance, Summit 2012, New York,
2012). See http://about.bnef.com/summit/content/uploads/sites/3/2013/11/BNEF_2012_03_19__University_Battery_
Innovation.pdf for lithium-ion battery pack cost and production.
9
Y. Wang, C. Zhang, and Z. Chen, “An adaptive remaining energy prediction approach for lithium-ion batteries in electric
vehicles,” J. Power Sources 305, 80–88 (2016).
10
M. Sedghi, A. Ahmadian, E. Pashajavid, and M. Aliakbar-Golkar, “Storage scheduling for optimal energy management
in active distribution network considering load, wind, and plug-in electric vehicles uncertainties,” J. Renewable
Sustainable Energy 7, 033120 (2015).
11
S. Chen, H. Hu, C. Wang, G. Wang, J. Yin, and D. Cao, “(LiFePO4-AC)/Li4Ti5O12 hybrid supercapacitor: The effect of
LiFePO4 content on its performance,” J. Renewable Sustainable Energy 4, 033114 (2012).
12
S. Hameer and J. L. Niekerk, “A review of large-scale electrical energy storage,” Int. J. Energy Res. 39, 1179–1195
(2015).
13
C. C. Hua, Y. H. Fang, and P. H. Li, “Charge equalization for series-connected LiFePO4 battery strings,” IET Power
Electron. 8, 1017–1025 (2015).
14
T. T. N. Nguyen, H. G. Yoo, S. K. Oruganti et al., “Neuro-fuzzy controller for battery equalisation in serially connected
lithium battery pack,” IET Power Electron. 8, 458–466 (2015).
15
M. Uno and K. Tanaka, “A single-switch equalization charger using multiple stacked buck-boost converters for series-
connected energy-storage modules,” Electr. Eng. Jpn. 182, 57–67 (2013).
16
H. Fakham, D. Lu, and B. Francois, “Power control design of a battery charger in a hybrid active PV generator for load-
following applications,” IEEE Trans. Ind. Electron. 58, 85–94 (2011).
17
P. C. P. Chao, W. D. Chen, and R. H. Wu, “A battery charge controller realized by a flyback converter with digital pri-
mary side regulation for mobile phones,” Microsyst. Technol. 20, 1689–1703 (2014).
18
M. K. Kazimierczuk, Pulse-width Modulated DC–DC Power Converters, 1st ed. (Wiley, UK, 2008).
19
L. R. Chen, N. Y. Chu, C. S. Wang, and R. H. Liang, “Design of a reflex based bidirectional converter with the energy
recovery function,” IEEE Trans. Ind. Electron. 55, 3022–3029 (2008).
20
C. P. C. Paul, W. D. Chen, and C. W. Cheng, “A fast charging algorithm for an intelligent PV system with capability of
on-line temperature compensation,” Microsyst. Technol. 19, 1289–1306 (2013).
21
S. Yu, M. Q. Nguyen, and W. Choi, “A novel soft-switching battery charge/discharge converter with the zero voltage dis-
charge function,” IEEE Trans. Power Electron. 31, 5067–5078 (2016).
22
M. Uno and A. Kukita, “Single-switch single-transformer cell voltage equalizer based on forward–flyback resonant
inverter and voltage multiplier for series-connected energy storage cells,” IEEE Trans. Veh. Technol. 63, 4232–4247
(2014).
23
O. Tremblay, L. A. Dessaint, and A. I. Dekkiche, “A generic battery model for the dynamic simulation of hybrid electric
vehicles,” in Proceedings of the IEEE Vehicle Power and Propulsion, Arlington, TX (2007), pp. 284–289.
24
Z. Xu, S. Gao, and S. Yang, “LiFePO4 battery state of charge estimation based on the improved Thevenin equivalent cir-
cuit model and Kalman filtering,” J. Renewable Sustainable Energy 8, 024103 (2016).
25
H. Aung and K. S. Low, “Temperature dependent state-of-charge estimation of lithium ion battery using dual spherical
unscented Kalman filter,” IET Power Electron. 8, 2026–2033 (2015).
26
Y. Zou, X. Hu, H. Ma, and S. E. Li, “Combined State of Charge and State of Health estimation over lithium-ion battery
cell cycle lifespan for electric vehicles,” J. Power Sources 273, 793–803 (2015).
27
R. Xiao, J. Shen, X. Li, W. Yan, E. Pan, and Z. Chen, “Comparisons of modeling and state of charge estimation for
lithium-ion battery based on fractional order and integral order methods,” Energies 9, 184 (2016).
28
P. D. Ngo and Y. C. Shin, “Gain estimation of nonlinear dynamic systems modeled by an FBFN and the maximum output
scaling factor of a self-tuning PI fuzzy controller,” Eng. Appl. Artif. Intell. 42, 1–15 (2015).
29
J. Kennedy and R. Eberhart, “Particle swarm optimization,” in Proceedings of IEEE International Conference on Neural
Networks, UWA, Perth, Western Australia (IEEE, 1995), pp. 1942–1948.
30
M. A. Rahman, S. Anwar, and A. Izadian, “Electrochemical model parameter identification of a lithium-ion battery using
particle swarm optimization method,” J. Power Sources 307, 86–97 (2016).
31
N. F. A. Hamid, N. A. Rahim, and J. Selvaraj, “Solar cell parameters identification using hybrid Nelder-Mead and modi-
fied particle swarm optimization,” J. Renewable Sustainable Energy 8, 015502 (2016).
065701-16 Hoque, Hannan, and Mohamed J. Renewable Sustainable Energy 8, 065701 (2016)
32
T. Sudhakar Babu, K. Sangeetha, and N. Rajasekar, “Voltage band based improved particle swarm optimization tech-
nique for maximum power point tracking in solar photovoltaic system,” J. Renewable Sustainable Energy 8, 013106
(2016).
33
X. Hu, S. Li, and H. Peng, “A comparative study of equivalent circuit models for Li-ion batteries,” J. Power Sources 198,
359–367 (2012).
34
I. Tabet, K. Touafek, N. Bellel, N. Bouarroudj, A. Khelifa, and M. Adouane, “Optimization of angle of inclination of the
hybrid photovoltaic-thermal solar collector using particle swarm optimization algorithm,” J. Renewable Sustainable
Energy 6, 053116 (2014).
35
M. Seyedmahmoudian, S. Mekhilef, R. Rahmani, R. Yusof, and A. Asghar Shojaei, “Maximum power point tracking of
partial shaded photovoltaic array using an evolutionary algorithm: A particle swarm optimization technique,”
J. Renewable Sustainable Energy 6, 023102 (2014).
36
X. Hu, S. Li, H. Peng, and F. Sun, “Robustness analysis of State-of-Charge estimation methods for two types of Li-ion
batteries,” J. Power Sources 217, 209–219 (2012).
