Journal of Energy Storage 52 (2022) 104813

Contents lists available at ScienceDirect

Journal of Energy Storage

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

Research Papers

Robust nonlinear control of battery electric vehicle charger in grid to vehicle

and vehicle to grid applications
Ijaz Ahmed b , Hafiz Mian Muhammad Adil a , Shahzad Ahmed b , Iftikhar Ahmad b ,∗,
Zubair Rehman b
a
Electrical Engineering Department, Lahore University of Management Sciences (LUMS), Lahore, Pakistan
b School of Electrical Engineering and Computer Science, National University of Sciences and Technology (NUST), Islamabad, Pakistan

ARTICLE INFO ABSTRACT

Keywords: The battery electric vehicle (BEV) charger provides an effective two-way interconnection between the grid and
DC–DC power converter the vehicle in both grid to vehicle (G2V) and vehicle to grid (V2G) configurations. However, the control of
Grid to vehicle power flow in either of the configurations is a challenging task. In this study, a supertwisting sliding mode-
Vehicle to grid
based controller (ST-SMC) is designed for the control of the bi-directional power converter in a BEV charger
Supertwisting sliding mode controller
for tracking the desired current and output voltage of the charger in both G2V and V2G operations. The
Battery electric vehicles (BEVs)
Integral backstepping sliding mode controller
proposed controller reduces the chattering effect, which is a major drawback in power converters. The robust
integral backstepping controller (IBS-SMC) is also designed in this paper for comparison purposes. The stability
of the system is analyzed using the Lyapunov stability criterion. The proposed controllers are simulated in
Matlab/Simulink environment. The proposed controller shows better dynamic performance as compared to
the IBS-SMC controller. Hardware in loop (HIL)-based experimental verification has been done by using the
Delfino F28369D dual core microcontroller in order to prove the effectiveness of the proposed controller.

1. Introduction into unidirectional and bidirectional chargers. An EV charger usually

consists of two converters: a power factor adjusted AC–DC converter
Recent technological developments have extraordinarily contributed and a DC–DC converter for battery connection. These converters are
to the widespread use of battery electric vehicles around the globe. often unidirectional, they can only charge in the G2V mode. The
However, as the number of BEVs on the road increases, the load on converter must be bidirectional and have a suitable power rating to
the power system grows as well. On the other hand, electric vehicles allow V2X technologies [4]. The stored energy in the vehicle battery
(EVs) may be thought as ‘‘mobile power banks’’ since they reserve a
is sent to the grid in V2G technology, while in V2H technology, the
significant quantity of energy that can be utilized to help the grid [1].
energy is provided to the house loads. In G2V mode, the bidirectional
The idea of vehicle-to-X (V2X), which transfers power from onboard
AC–DC converter acts as a rectifier with sinusoidal current absorption
battery to infrastructure, is likely to gain popularity in the near future
[2]. V2X describes the power flow from vehicles to other devices, whereas the reversible DC–DC converter operates in the buck mode.
and thus it can be a vehicle to load (V2L), vehicle to home (V2H) On the other hand, the AC–DC converter acts as an inverter in the
or vehicle to grid (V2G) as shown in Fig. 1. As an emergency power V2G operating mode while the reversible DC–DC converter works in
source, V2L is intended to immediately transfer electricity from a boost mode [5]. The general block diagram for the bidirectional DC–
vehicle to an appliances. V2H and V2G technologies are being applied, DC converter is shown in Fig. 2. The converter states 𝑦1 , 𝑦2 , and 𝑦3
with the goals of management, control of power distribution and grid are inputs to the controller block, which generates the control signal 𝑢.
power stability in the event of an outage. Several demonstration tests The pulse width modulation (PWM) block receives it as an input, which
have been conducted, and some of these technologies are currently transforms the signal 𝑢 into the switching signals 𝑢1 and 𝑢2 .
in use [3]. A reliable charging infrastructure is essential for electric For the control of charging units for BEVs, many linear controllers,
vehicles. Based on the power transmission mechanism, EV chargers such as proportional integral (PI) controllers and linear quadratic based
may be divided into two categories: conductive and inductive. In
optimum controllers [5–8], have been proposed in the literature. The PI
addition, based on the direction of power transfer, they are divided

∗ Corresponding author.
E-mail addresses: [email protected] (I. Ahmed), [email protected] (H.M.M. Adil), [email protected] (S. Ahmed),
[email protected] (I. Ahmad), [email protected] (Z. Rehman).

https://doi.org/10.1016/j.est.2022.104813
Received 11 November 2021; Received in revised form 14 April 2022; Accepted 2 May 2022
Available online 16 May 2022
2352-152X/© 2022 Elsevier Ltd. All rights reserved.
I. Ahmed et al. Journal of Energy Storage 52 (2022) 104813

For the regulation of the output voltage of the BEV charger, a

robust nonlinear controller is required to be designed to regulate the
output voltage and track the reference currents in G2V and V2G in the
presence of external disturbances with reduced chattering. ST-SMC is
used for the control of various applications in various works [21,22],
and it involves the continuous approximation of switching control.
Therefore, ST-SMC has been proposed in this study to fulfill the afore-
mentioned control objectives, with the control signal being translated
into duty cycles by a pulse width modulator. The following are the key
contributions of this paper:

• In contrast to most of the linear controllers available in the

literature, a robust nonlinear higher order sliding mode controller
(ST-SMC) has been designed.
• The effect of disturbances has been damped using the designed
nonlinear controller which undoubtedly improves the perfor-
mance and increase the life cycle of the battery.
• The effect of chattering has been minimized, which again reduces
the heat and power losses in the system.
• The dynamic response of the system has been improved which is
crucial for improving the operating requirements of G2V an V2G
modes.
Fig. 1. Power flow in V2X technologies.
• HIL experiments have been carried out for the validation of the
proposed controller in real time scenarios.

controller [9] has proposed designing a charger for an EV using a SEPIC The rest of the paper is organized in the following fashion: In Sec-
converter. These controllers are designed to manage the dynamics of tion 2, averaged mathematical modeling of a dual stage bi-directional
the converter in a BEV charger. The dynamic performance of these buck-boost converter is discussed. The design of controllers, as well as
controllers is higher, but their operation is limited to a single operating their stability analysis, is covered in Section 3. Section 4 includes the
point where the system is linearized. Moreover, they are not robust simulation results: Section 4.1 and Section 4.2 present the simulation
against external disturbances. results for the G2V and V2G modes, respectively. Section 4.3 compares
In the literature, soft computing-based controllers have also been the proposed nonlinear controllers, and Section 4.4 shows the HIL sim-
developed. In [10], a fuzzy PI controller has been designed for a multi- ulations of the proposed ST-SMC for BEV chargers. Section 5 includes
input DC–DC converter to control the speed profile of an EV. A fuzzy the conclusion of the paper.
logic-based controller has also been proposed for the BEV charger [11],
which does not require the use of a mathematical model and is based 2. Averaged mathematical modeling of DC-DC converter
on human reasoning. Each control variable is assigned a membership
The power conversion unit in the system is assumed to be in con-
function with a value between 0 and 1. Their effectiveness is dependent
tinuous conduction mode. However, the detailed derivation illustrating
on the availability of correct system information, which is not always
the various modes of power conditioning circuitry, has been derived as
the case. Similarly, in recent literature, many database techniques [12–
follows:
14] for estimation of the state of health (SoH) of batteries are presented
for various applications, including electric vehicles however, these are
2.1. Description of the system
data dependent approaches and do not include the nonlinear control of
the G2V and V2G operations.
The electrical circuit of the bidirectional DC–DC converter which is
Many nonlinear controllers have also been developed to handle
connected to the battery is illustrated in Fig. 3 [23–26]. This converter
the nonlinear dynamics of the BEV charger. An Output feedback con-
behaves like a half bridge, consisting of two switches 𝑆1 , 𝑆2 , the filter-
troller [15] has been proposed for the DC–DC converter, It ensures
ing inductance 𝐿 and capacitance 𝐶. Inductance 𝐿, along with internal
the system’s global asymptotic stability, but does not show dynamic resistance 𝑟, is shown in Fig. 4. The output voltage of the converter
performance effectively. Furthermore, it is not robust against external is used to charge the battery. This converter acts as a buck converter
disturbances. Backstepping (BS) has been proposed [16] for the control in the G2V operation mode. It is used to control the battery current
of power converters in a BEV charging unit that has sufficient dynamic and voltage during the charging stage. During V2X operating mode,
performance. However, there is some steady state error and it is not the converter acts as a boost converter to scale up the battery voltage
robust against external disturbances. The addition of a switching con- to a suitable DC-bus voltage in order to ensure smooth operation.
trol law in the control strategy of BS makes it robust against external
disturbances, but at the cost of chattering, which introduces heat and 2.2. Modeling of the system
power losses in the system [17].
The main disadvantage of all of these controllers is that they are 2.2.1. Battery
not robust against external disturbance and they do not ensure finite Several models for batteries exist in the literature [27–29]. How-
time convergence to their desired trajectories. An IBS-SMC [18] has ever, for simplicity, the basic RC model for the battery has been taken.
been designed for controlling the dynamics of the AC-DC converter in a The electrical equivalent circuit of the battery is shown in Fig. 5.
BEV charger. These controllers have finite time convergence and better It consists of series and parallel equivalent resistances 𝑅𝑆 and 𝑅𝑃
dynamic performance compared to other controllers. In [19], a sliding respectively. 𝐶𝐵𝑎𝑡 is the capacitance of the battery.
mode-based controller for a charger is proposed that is robust against
uncertainties and disturbances of only some particular types. On the 2.2.2. Bidirectional DC-DC converter
other hand, the sliding trajectory shows the chattering effect, which is This converter operates in the following modes during V2G and
the main disadvantage of the sliding mode controller [20]. G2V.

2
I. Ahmed et al. Journal of Energy Storage 52 (2022) 104813

Fig. 2. Control function block diagram of DC–DC converter.

Fig. 6. Buck operating mode.

2.2.2.1. Buck mode. In this mode, the switch 𝑆1 is controlled by a

PWM signal 𝑢1 , whereas the switch 𝑆2 remains open for duty cycle 𝑢2 ,
which is equal to zero. The DC output link of the charger subsequently
Fig. 3. Bidirectional DC–DC power converter. transmits the electrical energy to the battery. In this operating mode,
the converter scales down the output voltage of the charger to a level
that is required for charging of the battery. The overall system will
operate in G2V mode. The corresponding electrical circuit of the con-
verter in this case is illustrated in Fig. 6. The following switching model
can be derived from this figure after applying the volt-second balance
theorem to the inductor and the charge-second balance theorem to the
capacitors:
𝑑𝑖𝐿
𝐿 = −𝑟𝑖𝐿 − 𝑉𝐵𝑎𝑡 + 𝑢1 𝑉𝐷𝐶 (1)
𝑑𝑡
𝑑𝑉𝐵𝑎𝑡 1 1
𝐶 = 𝑖𝐿 𝑉 + 𝑉 (2)
Fig. 4. Inductor with internal resistance. 𝑑𝑡 𝑅𝑆 𝐵𝑎𝑡 𝑅𝑆 𝐶
𝑑𝑉 1 1 1
𝐶𝐵𝑎𝑡 𝑐 = 𝑉 − 𝑉 − 𝑉 (3)
𝑑𝑡 𝑅𝑆 𝐵𝑎𝑡 𝑅𝑆 𝐶 𝑅𝑃 𝐶

2.2.2.2. Boost mode. In this mode, the switch 𝑆2 remains off for the
duty cycle of 𝑢2 and the switch 𝑆1 remains open for 𝑢1 equal to zero. The
energy is transferred from the battery to the DC bus. In this operating
mode, the DC–DC converter will be used to boost the battery voltage to
the desired voltage value, while the whole system will operate in V2G
mode. The corresponding electrical circuit of the converter in this case
is illustrated in Fig. 7. The following switching model can be obtained
after applying the volt-second balance theorem to the inductor and the
charge-second balance theorem to the capacitors:
𝑑𝑖𝐿
𝐿 = −𝑟𝑖𝐿 − 𝑉𝐵𝑎𝑡 + (1 − 𝑢2 )𝑉𝐷𝐶 (4)
𝑑𝑡
𝑑𝑉𝐵𝑎𝑡 1 1
𝐶 = 𝑖𝐿 𝑉 + 𝑉 (5)
𝑑𝑡 𝑅𝑆 𝐵𝑎𝑡 𝑅𝑆 𝐶
𝑑𝑉 1 1 1
𝐶𝐵𝑎𝑡 𝑐 = 𝑉 − 𝑉 − 𝑉 (6)
Fig. 5. Battery electric model.
𝑑𝑡 𝑅𝑆 𝐵𝑎𝑡 𝑅𝑆 𝐶 𝑅𝑃 𝐶

3
I. Ahmed et al. Journal of Energy Storage 52 (2022) 104813

3.1. Super twisting sliding mode controller

To design a proposed controller for a dual-stage CC-CV converter,

first we define a battery current, 𝑖𝐵𝑎𝑡 which is obtained by applying
KCL,
𝑦
𝑖𝐵𝑎𝑡 = 𝑦1 − 2 (13)
𝐶
Now the error 𝜁1 is define as, the difference between 𝑖𝐵𝑎𝑡 and 𝑖𝐵𝑎𝑡𝑟𝑒𝑓 ,

𝜁1 = 𝑖𝐵𝑎𝑡 − 𝑖𝐵𝑎𝑡𝑟𝑒𝑓 (14)

Fig. 7. Boost operating mode. where 𝑖𝐵𝑎𝑡𝑟𝑒𝑓 is the desired current value for the battery.
Plugging the value of 𝑖𝐵𝑎𝑡 from Eq. (13) into Eq. (14) gives,
𝑦
𝜁1 = 𝑦1 − 2 − 𝑖𝐵𝑎𝑡𝑟𝑒𝑓 (15)
2.2.2.3. Bidirectional mode. The buck and boost mathematical models 𝐶
can be used to create a global model of the system that can be used for Now define the other two errors as:
the controller design. The following switching model for a bidirectional 𝜁2 = 𝑦2 − 𝑦2𝑟𝑒𝑓 (16)
DC–DC converter can be obtained by using Eqs. (1)–(6) as:
𝑑𝑖𝐿 𝜁3 = 𝑦3 − 𝑦3𝑟𝑒𝑓 (17)
𝐿 = −𝑟𝑖𝐿 − 𝑉𝐵𝑎𝑡 + 𝜔𝑉𝐷𝐶 (7)
𝑑𝑡
𝑑𝑉 1 1 where 𝑦2𝑟𝑒𝑓 and 𝑦3𝑟𝑒𝑓 are the desired voltage values across the capaci-
𝐶 𝐵𝑎𝑡 = 𝑖𝐿 𝑉 + 𝑉 (8) tor.
𝑑𝑡 𝑅𝑆 𝐵𝑎𝑡 𝑅𝑆 𝐶
𝑑𝑉 Taking the time derivative of Eqs. (15)–(17), where the values of
1 1 1
𝐶𝐵𝑎𝑡 𝑐 = 𝑉 − 𝑉 − 𝑉 (9) 𝑖̇ 𝐵𝑎𝑡𝑟𝑒𝑓 , 𝑦̇ 2𝑟𝑒𝑓 and 𝑦̇ 3𝑟𝑒𝑓 are equal to zero
𝑑𝑡 𝑅𝑆 𝐵𝑎𝑡 𝑅𝑆 𝐶 𝑅𝑃 𝐶
𝑦̇2
where 𝜁̇1 = 𝑦̇1 − (18)
𝐶
𝜔 = (𝑄𝑢1 + (1 − 𝑄)(1 − 𝑢2 ))
𝜁̇2 = 𝑦̇2 (19)
𝑄 is a parameter which depends upon the operating modes of the
DC–DC converter, 𝜁̇3 = 𝑦̇3 (20)
• If 𝑄 = 1 (Buck mode) Now choose a sliding surface 𝜎 that allows the system’s states to follow
• If 𝑄 = 0 (Boost mode) their corresponding trajectories as:

The following equations represent the average mathematical modeling 𝜎 = 𝑎1 𝜁1 + 𝑎2 𝜁2 + 𝑎3 𝜁3 (21)

of a DC–DC converter, which is obtained by using Eqs. (7)–(9).
where 𝑎1 , 𝑎2 and 𝑎3 are all positive constants.
−𝑟 1 𝑉 The derivative of Eq. (21) with respect to time gives,
𝑦̇ 1 = 𝑦 − 𝑦 + 𝑢 𝐷𝐶 + 𝑑(𝑡) (10)
𝐿 1 𝐿 2 𝐿
1 1 1 𝜎̇ = 𝑎1 𝜁̇1 + 𝑎2 𝜁̇2 + 𝑎3 𝜁̇3 (22)
𝑦̇ 2 = 𝑦1 − 𝑦2 + 𝑦 (11)
𝐶 𝑅𝑆 𝐶 𝑅𝑆 𝐶 3 For the stability analysis, the Lyapunov candidate function is consid-
1 1 1 ered as:
𝑦̇ 3 = 𝑦 − 𝑦 − 𝑦 (12)
𝑅𝑆 𝐶𝐵𝑎𝑡 2 𝑅𝑆 𝐶𝐵𝑎𝑡 3 𝑅𝑃 𝐶𝐵𝑎𝑡 3
1
𝑉 = 𝜎2 (23)
where 𝑦̇ 1 represents the average inductor current 𝑖𝐿 , 𝑦̇ 2 and 𝑦̇ 3 denote 2
the average output voltage 𝑉𝐵𝑎𝑡 of the converter and the average volt- Taking the time derivative of Eq. (23) yields:
age 𝑉𝐶 of the battery capacitor respectively. The bounded disturbance
𝑑(𝑡) has been added in the system, whose bounds are as follow: 𝑉̇ = 𝜎 𝜎̇ (24)

Substituting the value 𝜎̇ from Eq. (22) into Eq. (24), we get:
𝑀1 ≤ 𝑑(𝑡) ≤ 𝑀2
𝑉̇ = 𝜎(𝑎1 𝜁̇ 1 + 𝑎2 𝜁̇ 2 + 𝑎3 𝜁̇ 3 ) (25)
where 𝑀1 and 𝑀2 are the disturbance’s lower and upper limits, respec-
tively. Physically, the disturbance can be observed as switching spikes Now substituting the values of 𝜁̇ 1 , 𝜁̇ 2 and 𝜁̇ 3 from Eqs. (18)–(20)
in the converter or any transient significant enough to test the proposed respectively, in Eq. (25), we have:
controllers. ( )
𝑦̇
𝑉̇ = 𝜎 𝑎1 (𝑦̇ 1 − 2 ) + 𝑎2 𝑦̇ 2 + 𝑎3 𝑦̇ 3 (26)
𝐶
3. Controller design The following constraints are added to make the Lyapunov function’s
derivative is negative definite:
The dual-stage constant current-constant voltage (CC-CV) charge 𝑦̇2
process is performed during the G2V mode, while during the V2G 𝑎1 (𝑦̇ 1 − ) + 𝑎2 𝑦̇ 2 + 𝑎3 𝑦̇ 3 = − 𝛼1 |𝜎|0.5 𝑠𝑖𝑔𝑛(𝜎)−
𝐶
mode, the only CC process with a negative current reference is per- 𝑡

formed. The robust nonlinear controllers have to be designed to achieve 𝛼2 𝑠𝑖𝑔𝑛(𝜎)𝑑𝜏 (27)
∫0
the following objectives:
Substituting the value of 𝑦̇ 1 from Eq. (10) into Eq. (27), we have:
(
• to ensure constant current and voltage regulation during G2V and 𝑡
−𝑟 1
−𝛼1 |𝜎|0.5 𝑠𝑖𝑔𝑛(𝜎) − 𝛼2 𝑠𝑖𝑔𝑛(𝜎)𝑑𝜏 =𝑎1 𝑦 − 𝑦 +
V2G modes, ∫0 𝐿 1 𝐿 2
• to charge and discharge the battery safely during G2V and V2G )
𝑉 𝑦̇
modes respectively. 𝑢 𝐷𝐶 + 𝑑(𝑡) − 2
𝐿 𝐶

4
I. Ahmed et al. Journal of Energy Storage 52 (2022) 104813

+ 𝑎2 𝑦̇ 2 + 𝑎3 𝑦̇ 3 (28) Taking the time derivative of Eq. (41) yields:

Now the overall controller 𝑢𝑠𝑡𝑠𝑚𝑐 can be obtained from Eq. (28) as: 𝑉̇ 1 = 𝜁1 𝜁̇ 1 + 𝜇𝜂𝜁1 (42)
(
−𝑎1 𝑟𝑦1 𝑎1 𝑦2 𝑎 𝑦̇ Substituting the value of 𝜁̇1 from Eq. (37) into Eq. (42) yields:
𝑢𝑠𝑡𝑠𝑚𝑐 =𝑍 − + 𝑑(𝑡) − 1 2 + 𝑎2 𝑦̇ 2
𝐿 𝐿 𝐶 ( )
) −𝑟 1 𝑉 𝑦̇
𝑡 𝑉̇ 1 = 𝜁1 𝑦1 − 𝑦2 + 𝑢 𝐷𝐶 − 2 + 𝜇𝜂 (43)
+ 𝑎3 𝑦̇ 3 + 𝛼1 |𝜎|0.5 𝑠𝑖𝑔𝑛(𝜎) + 𝛼2 𝑠𝑖𝑔𝑛(𝜎)𝑑𝜏 (29) 𝐿 𝐿 𝐿 𝐶
∫0
For the stability of the system, put:
where
−𝑟 1 𝑉 𝑦̇
−𝐿 𝑦 − 𝑥 + 𝑢 𝐷𝐶 − 2 + 𝜇𝜂 = −𝑎1 𝜁1 (44)
𝑍=
𝑎1 𝑉𝐷𝐶 𝐿 1 𝐿 2 𝐿 𝐶
where 𝑎1 is a positive constant. Using Eqs. (43) and (44) we have:
For an equivalent control, put 𝑑(𝑡) = 0 in eq (29). The corresponding 𝑢𝑒𝑞
and switching 𝑢𝑠𝑤 controllers can be derived from Eq. (29) as follows: 𝑉̇ 1 = −𝑎1 𝜁12 (45)
( )
−𝑎1 𝑟𝑦1 𝑎1 𝑦2 𝑎1 𝑦̇ 2 Eq. (45) denotes that 𝑉̇ 1 is negative definite. The virtual control law
𝑢𝑒𝑞 = 𝑍 − − + 𝑎2 𝑦̇ 2 + 𝑎3 𝑦̇ 3 (30)
𝐿 𝐿 𝐶 𝑦2 = 𝜐 can be derived from Eq. (44) as:
and ( )
−𝑟𝑦1 𝑢𝑉𝐷𝐶 𝑦̇
( 𝑡 ) 𝜐=𝐿 + − 2 + 𝜇𝜂 + 𝑎1 𝜁1 (46)
𝐿 𝐿 𝐶
𝑢𝑠𝑤 = 𝑍 𝛼1 |𝜎|0.5 𝑠𝑖𝑔𝑛(𝜎) + 𝛼2 𝑠𝑖𝑔𝑛(𝜎)𝑑𝜏 (31)
∫0
Now define another error 𝜁2 for tracking the state 𝑦2 to 𝜐 as,
𝛼1 , 𝛼2 , 𝜆𝑚𝑖𝑛 , 𝜆𝑚𝑎𝑥 and 𝜉 are the design parameters which are given
in [30] as: 𝜁2 = 𝑦 2 − 𝜐 (47)
𝜉 Using the value of 𝑦2 from Eq. (47) in Eq. (37) gives:
𝛼2 > + 𝑀2 , (32)
𝜆 𝑚𝑖𝑛 −𝑟𝑦1 𝜁2 𝑦̇ 𝑢𝑉
𝜐
4𝜉𝜆𝑚𝑎𝑥 (𝛼2 + 𝜉) 𝜁̇ 1 = − − − 2 + 𝐷𝐶 (48)
𝐿 𝐿 𝐿 𝐶 𝐿
𝛼12 ≥ + 𝑀1 , (33)
𝜆2𝑚𝑖𝑛 𝜆𝑚𝑖𝑛 (𝑝2 − 𝜉) Now substituting the value of 𝜐 from Eq. (46) into Eq. (48), as:
with conditions, 𝜁2
𝜁̇ 1 = −𝜇𝜂 − 𝑎1 𝜁1 − (49)
| 𝑑 𝜎̇ 𝑑 𝜎̇ | 𝐿
𝜉 > || + [𝑓 (𝑥, 𝑡) + 𝑏(𝑡)𝑢(𝑡) + 𝑑(𝑡)]||, (34)
| 𝑑𝑡 𝑑𝑥 | Plugging the value of 𝜁̇ 1 from Eq. (49) into Eq. (42), we get 𝑉̇ 1 as:
where 𝑓 (𝑥, 𝑡) is the system matrix and 𝑏(𝑡) is the input matrix. 𝜁1 𝜁2
𝑉̇ 1 = −𝑎1 𝜁12 − (50)
| 𝑑 𝜎̇ | 𝐿
0 ≤ 𝜆𝑚𝑖𝑛 ≤ | | ≤ 𝜆𝑚𝑎𝑥 . (35)
| 𝑑𝑢 | Taking the time derivative of Eq. (47) as:
Now Eq. (26) can be simplified as:
𝜁̇2 = 𝑦̇ 2 − 𝜐̇ (51)
𝑡
1.5
𝑉̇ = −𝛼1 |𝜎| 𝑠𝑖𝑔𝑛(𝜎) − 𝜎𝛼2 𝑠𝑖𝑔𝑛(𝜎)𝑑𝜏 (36) Taking the time derivative of Eq. (46) gives:
∫0
𝑦̈2
where the values of parameters 𝛼1 and 𝛼2 meet the requirements of 𝜐̇ = −𝑟𝑦̇1 + 𝑢𝑉
̇ 𝐷𝐶 − 𝐿 + 𝐿𝜇𝜁1 + 𝐿𝑎1 𝜁̇1 (52)
inequalities (32) and (33) respectively. It is evident from Eq. (36) that 𝐶
𝑉̇ is negative definite, ensuring the stability of the system. Plugging the value of 𝜁̇1 from Eq. (49) into Eq. (52), we get 𝜐̇ as:
𝑦̈2 𝜁
3.2. Integral backstepping sliding mode controller 𝜐̇ = −𝑟𝑦̇1 + 𝑢𝑉
̇ 𝐷𝐶 − 𝐿 + 𝐿𝜇𝜁1 + 𝐿𝑎1 (−𝜇𝜂 − 𝑎1 𝜁1 − 2 ) (53)
𝐶 𝐿
Now the cumulative Lyapunov candidate function 𝑉11 is considered as:
In this section, an integral backstepping sliding mode controller has
been designed. The equivalent controller is obtained by adding integral
1 2
action to the error term. A switching term would be added to make it 𝑉11 = 𝑉1 + 𝜁 (54)
2 2
robust against external disturbances.
Taking the derivative of Eq. (54) with respect to time gives:
Substituting the value of 𝑦̇ 1 from Eq. (10) into Eq. (18) yields:
−𝑟 1 𝑉 𝑦̇ 𝑉̇ 11 = 𝑉̇ 1 + 𝜁2 𝜁̇ 2 (55)
𝜁̇ 1 = 𝑦 − 𝑦 + 𝑢 𝐷𝐶 − 2 (37)
𝐿 1 𝐿 2 𝐿 𝐶
Substituting the value of 𝑉̇ 1 from Eq. (50) in Eq. (55) gives:
Now introduce the integrator term 𝜂 as: ( )
𝜁
𝑡 𝑉̇ 11 = −𝑎1 𝜁12 + 𝜁2 𝜁̇ 2 − 1 (56)
𝜂= (𝑖𝐵𝑎𝑡 − 𝑖𝐵𝑎𝑡𝑟𝑒𝑓 )𝑑𝑡 (38) 𝐿
∫0
Now by substituting
Taking the time derivative of Eq. (38) gives:
𝜁1
𝜁̇ 2 − = −𝑎2 𝜁2 (57)
𝜂̇ = 𝑖𝐵𝑎𝑡 − 𝑖𝐵𝑎𝑡𝑟𝑒𝑓 (39) 𝐿
where 𝑎2 is a positive constant.
From Eqs. (14) and (39), we get:
The stability of the system is ensured by using Eqs. (56) and (57)
𝜂̇ = 𝜁1 (40) as:

For the stability of the system, the Lyapunov candidate function is 𝑉̇ 11 = −𝑎1 𝜁12 − 𝑎2 𝜁22 (58)
considered as:
Plugging the value of 𝜐̇ from Eq. (53) in Eq. (51), we get 𝜁̇ 2 as:
1 𝜇 (
𝑉1 = 𝜁12 + 𝜂 2 (41)
2 2 𝑦̈
̇ 𝐷𝐶 − 𝐿 2 + 𝐿𝜇𝜁1 +
𝜁̇ 2 =𝑦̇ 2 − −𝑟𝑦̇1 + 𝑢𝑉
where 𝜇 is a positive constant. 𝐶

5
I. Ahmed et al. Journal of Energy Storage 52 (2022) 104813

Table 1 Table 3
Circuit components and values. Comparison between the dynamic performances of proposed controllers.
Parameters Values Response ST-SMC IBS-SMC
Capacitor (𝐶) 700 μF Rise time (s) 1.8003 1.8200
Inductance (𝐿) 5 mH Settling time (s) 2.2439 2.2862
Resistance (𝑟) 0, 1 Ω Overshoot 0.0014 0.0092
Battery Capacitor (𝐶𝐵𝑎𝑡 ) 500 F Undershoot 0 0
Series resistance (𝑅𝑆 ) 0, 0.06 Ω Peak time (s) 2.4972 2.4580
Parallel resistance (𝑅𝑃 ) 1 KΩ Peak value 230.0032 230.0211
Switching frequency (𝑓𝑠 ) 20 kHz Steady state error 0.0032 0.0211
𝑉𝐷𝐶 400 V

Table 2
Parametric values of the controllers.
Controller Parameters Values
ST-SMC 𝑎1 100
𝑎2 800
𝑎3 1500
𝛼1 15
𝛼2 0.89
IBS-SMC 𝑎1 500
𝑎2 100
𝑎3 750
𝜇 50
𝑘 0.17

( )
𝜁 ) Fig. 8. PWM signal diagram.
𝐿𝑎1 −𝜇𝜂 − 𝑎1 𝜁1 − 2 (59)
𝐿
From Eqs. (57) and (59), we get 𝑢̇ 𝑒𝑞 as:
( (
1 𝑦̈
𝑢̇ 𝑒𝑞 = ̇ 𝐷𝐶 − 𝐿 2 + 𝐿𝜇𝜁1 +
𝑦̇2 − −𝑟𝑦̇1 + 𝑢𝑉
𝑉𝐷𝐶 𝐶
) )
( 𝜁2 ) 𝜁1
𝐿𝑎1 −𝜇𝜂 − 𝑎1 𝜁1 − − + 𝑎2 𝜁2 (60)
𝐿 𝐿

The switching term is introduced to make the controller robust as:

𝑢𝑠 = −𝑘𝑠𝑖𝑔𝑛(𝜎) (61)

where 𝑘 is a positive constant. The overall control 𝑢 is given as:

𝑢 = 𝑢𝑒𝑞 + 𝑢𝑠 (62)

where

𝑡
𝑢𝑒𝑞 = 𝑢̇ 𝑒𝑞 𝑑𝑡 (63)
∫0
Now using the values of 𝑢𝑒𝑞 from Eq. (63) and 𝑢𝑠 from Eq. (61) in
Eq. (62), we get IBS-SMC control 𝑢𝐼𝐵𝑆−𝑆𝑀𝐶 law as:
( ( Fig. 9. Step.
𝑡 𝑦̈
1
𝑢= ̇ 𝐷𝐶 − 𝐿 2 + 𝐿𝜇𝜁1 +
𝑦̇2 − −𝑟𝑦̇1 + 𝑢𝑉
∫0 𝑉𝐷𝐶 𝐶
) )
( 𝜁 ) 𝜁 4.1. Simulations for G2V mode
𝐿𝑎1 −𝜇𝜂 − 𝑎1 𝜁1 − 2 − 1 + 𝑎2 𝜁2 𝑑𝑡 − 𝑘𝑠𝑖𝑔𝑛(𝜎) (64)
𝐿 𝐿
The objective of this section is to show the behavior of the designed
controllers in G2V mode in the presence of disturbances. The reference
4. Simulations and results
value for the battery voltage 𝑉𝐵𝑎𝑡 is taken as 230 V, and Fig. 10 shows
the satisfactory performance of the designed controllers in both CC and
The proposed controllers have been validated using the MAT- CV stages. However, it can be seen that in the case of IBS-SMC there
LAB/Simulink environment. The system parameters are listed in Ta- is a considerable amount of chattering, while in the case of ST-SMC it
ble 1, while the controller gains are given in Table 2. The generation is negligible. Similarly, Fig. 11 shows the tracking of battery current to
of the required control signals depends upon the operating modes as the reference value of 32 A. It can be observed that in the CC stage,
shown in Fig. 8. As the designed controllers are robust against external IBS-SMC shows a considerable steady state error along with chattering
disturbances, a step type disturbance has been added to the system due to the effect of disturbances present in the system, whereas ST-SMC
prior to the simulations for their comparison in both G2V and V2G shows better tracking of the required reference with a reduced amount
modes. The step disturbance as shown in Fig. 9 has been added to the of chattering. The inner voltage of a battery (𝑉𝐶 ) represents its state of
state 𝑥1 of the system. charge and is shown in Fig. 12.

6
I. Ahmed et al. Journal of Energy Storage 52 (2022) 104813

Fig. 10. Comparative analysis of Battery voltage 𝑉𝐵𝑎𝑡 during G2V mode.

Fig. 13. Battery current 𝑖𝐵𝑎𝑡 during discharging (V2G mode).

Fig. 14. Battery voltage 𝑉𝐵𝑎𝑡 during discharging (V2G mode).

Fig. 11. Battery current 𝑖𝐵𝑎𝑡 during G2V mode.

4.2. Simulations for V2G mode

In this section, the performance of both the designed controllers has

been compared in V2G mode. Fig. 13 shows the discharging of the bat-
tery using only the CC stage with a negative reference of 10 A. Both of
the designed controllers are able to track the desired reference, but the
amount of chattering present in the case of IBS-SMC is considerable as
compared to that of ST-SMC. Further, the large undershoot at delayed
convergence of IBS-SMC makes it less robust as the system is already
under the effect of disturbance. Similarly, Figs. 14 and 15 show the 𝑉𝐵𝑎𝑡
and 𝑉𝐶 of the battery by both the designed controllers respectively. It
is clear from the figures that SoC of battery is rapidly decaying in case
of IBS-SMC while in case of ST-SMC the fast draining of battery is being
preserved.

4.3. Comparison between proposed controllers

The designed controllers have also been evaluated and compared in

terms of their dynamic response, as shown numerically in Table. (3).
Fig. 12. Battery inner voltage 𝑉𝐶 during G2V.
The settling time for ST-SMC is less than IBS-SMC, which shows its fast

7
I. Ahmed et al. Journal of Energy Storage 52 (2022) 104813

Fig. 15. Battery inner voltage 𝑉𝐶 during discharging (V2G mode). Fig. 17. Battery voltage 𝑉𝐵𝑎𝑡 during G2V (HIL).

Fig. 18. Battery current 𝑖𝐵𝑎𝑡 during G2V (HIL).

Fig. 16. HIL setup.

convergence to the desired values. The magnitude of overshoots in the

case of IBS-SMC is relatively larger than that of ST-SMC which demon-
strates the good transient response of ST-SMC. Further, the steady state
error for ST-SMC is also small as compared to IBS-SMC, which indicates
tight control in the presence of disturbances. In a nutshell, ST-SMC
outperforms IBS-SMC in almost every aspect of dynamic response and
shows more robustness than IBS-SMC. All these factors therefore make
ST-SMC a good choice for robust control of the proposed system (see
Table 3).

4.4. Hardware in the loop simulations

Fig. 19. Battery current 𝑖𝐵𝑎𝑡 during V2G (HIL).
Hardware-in-Loop experiments have also been carried out for the
experimental verification of the designed ST-SMC controller. A dual
core Delfino F283769D microprocessor is used for validating the con-
troller in both G2V and V2G modes. Fig. 16 represents the setup for the Fig. 17 shows the charging process of the battery in the CC-CV stage.
HIL experiment. The result shows that ST-SMC performs satisfactorily under the effect

8
I. Ahmed et al. Journal of Energy Storage 52 (2022) 104813

References

[1] H. El Fadil, F.Z. Belhaj, A. Rachid, F. Giri, T. Ahmed-Ali, Nonlinear modeling and
observer for supercapacitors in electric vehicle applications, IFAC-Papers online
50 (1) (2017) 1898–1903.
[2] J.G. Pinto, et al., Bidirectional battery charger with grid-to-vehicle, vehicle-to-
grid and vehicle-to-home technologies, in: Industrial Electronics Society, IECON
2013-39th Annual Conference of the IEEE, 2013, pp. 5934–5939.
[3] T. Izumi, M. Hirota, K. Hatanaka, Y. Isoyama, Bidirectional charging unit for
vehicle-to-X (V2X) power flow, SEI Tech. Rev. (79) (2014) 39–42.
[4] A. Rachid, H. El Fadil, F.Z. Belhaj, K. Gaouzi, F. Giri, Lyapunov- based control
of single-phase ac-dc power converter for BEV charger, in: Proceeding of the
3rd IEEE International Conference on Electrical and Information Technologies,
ICEIT, Rabat, Morocco, 2017, pp. 1–5, http://dx.doi.org/10.1109/EITech.2017.
8255296.
[5] A. Rachid, H. El Fadil, K. Gaouzi, F.Z. Belhaj, Output feedback control of
bidirectional dc-dc power converter for BEV charger, in: Proceeding of the
4th International Conference on Automation, Control Engineering and Computer
Science, ACECS, Tangier, Morocco, vol. 19, 2017, pp. 46–51.
[6] J.G. Vítor Monteiro, João Luiz Afonso Pinto, Operation modes for the electric
vehicle in smart grids and smart homes: Present and proposed modes, IEEE Trans.
Veh. Technol. 65 (3) (2015) 1007–1020.
[7] Liwen Pan, Chengning Zhang, An integrated multifunctional bidirectional AC/DC
and DC/DC converter for electric vehicles applications, Energies 9 (7) (2016)
Fig. 20. Battery voltage 𝑉𝐵𝑎𝑡 during V2G (HIL). 493.
[8] Mauricio Restrepo, et al., Modeling and testing of a bidirectional smart charger
for distribution system EV integration, IEEE Trans. Smart Grid 9 (1) (2016)
152–162.
of disturbances and tracks the desired trajectory in real time scenarios. [9] M. Karuppiah, P. Dineshkumar, K. Karthikumar, Design a electric vehicle charger
based sepic topology with PI controller, in: 2020 IEEE International Confer-
Similarly, Fig. 18 shows the performance of the ST-SMC for charging
ence on Advances and Developments in Electrical and Electronics Engineering,
current and the internal voltage of the battery (representing the SoC). ICADEE, pp. 1–5.
The performance of ST-SMC in V2G mode, where the discharging [10] S. Devi Vidhya, M. Balaji, Hybrid fuzzy PI controlled multi-input DC/DC
current of the battery and the reduction in SoC are shown in Figs. 19 converter for electric vehicle application, Automatika 61 (1) (2020) 79–91.
and 20 respectively. [11] Khairy Sayed, Hossam A. Gabbar, Gabbar ’electric vehicle to power grid inte-
gration using three-phase three-level AC/DC converter and PI-fuzzy controller,
The HIL results are very much in line with the simulated results Energies 9 (7) (2016) 532.
and thus show the applicability of the designed ST-SMC for real time [12] Chengqi She, et al., Battery state of health estimation based on incremental ca-
scenarios. pacity analysis method: synthesizing from cell-level test to real-world application,
IEEE J. Emerg. Sel. Top. Power Electron. (2021).
[13] Cong Wang, et al., A vehicle rollover evaluation system based on enabling state
5. Conclusion and parameter estimation, IEEE Trans. Ind. Inf. 17 (6) (2020) 4003–4013.
[14] Zhenpo Wang, et al., A data-driven method for battery charging capacity
abnormality diagnosis in electric vehicle applications, IEEE Trans. Transp.
In this study, a supertwisting sliding mode-based controller is pro-
Electrification (2021).
posed for the control of BEV chargers in both G2V and V2G operations. [15] Aziz Rachid, et al., Nonlinear output feedback control of V2G single phase
IBS-SMC has also been designed for the control of a bi-directional on-board BEV charger, Asian J. Control 22 (5) (2020) 1848–1859.
power converter in a charger for comparison purposes. IBS-SMC shows [16] I. Ahmed, I. Ahmad, S. Ahmed, H.M.M. Adil, Robust nonlinear control of battery
electric vehicle charger in grid to vehicle applications, J. Energy Storage 42
good dynamic performance, but it exhibits the chattering effect, which
(2021) 103039.
causes heat and power losses in the system. In order to resolve this [17] H.K. Khalil, Nonlinear Systems, third ed., Prentice-Hall, Upper Saddle River, NJ,
issue, a supertwisting algorithm-based controller is designed which re- USA, 2002.
duces the chattering effect and exhibits better dynamic performance as [18] Shahzad Ahmed, et al., Conditioned-based robust nonlinear control of plug-in
hybrid electric vehicle with saturated control actions, J. Energy Storage 43
compared to IBS-SMC. The simulation results show the fast convergence
(2021) 103201.
and reduced steady error in the response of ST-SMC as compared to IBS- [19] Siew-Chong Tan, Yuk-Ming Lai, K. Tse Chi, General design issues 400 of sliding-
SMC. HIL-based simulations have been performed where the proposed mode controllers in DC–DC converters, IEEE Trans. 401 Ind. Electron. 55 (3)
controller tracks the desired currents and voltage in the presence of (2008) 1160–1174.
external disturbances with negligible chattering effect, which proves [20] J.J.E. Slotine, W. Li, Applied Nonlinear Control, Vol. 199, Prentice-Hall,
EnglewoodCliffs, NJ, 1991.
the robustness of the proposed controller. [21] Hafiz Mian Muhammad Adil, Shahzad Ahmed, Iftikhar Ahmad, Control of
MagLev system using supertwisting and integral backstepping sliding mode
CRediT authorship contribution statement algorithm, IEEE Access 8 (2020) 51352–51362.
[22] Shahzad Ahmed, et al., Supertwisting sliding mode algorithm based nonlinear
MPPT control for a solar PV system with artificial neural networks based
Ijaz Ahmed: Conceptualization, Methodology, Software, Formal reference generation, Energies 13 (14) (2020) 3695.
analysis, Writing – original draft, Visualization. Hafiz Mian Muham- [23] V. Monteiro, et al., Assessment of a battery charger for electric vehicles
with reactive power control, in: IECON 2012-38th Annual Conference on IEEE
mad Adil: Resources, Investigation, Visualization. Shahzad Ahmed:
Industrial Electronics Society, 2012, pp. 5142–5147.
Software, Investigation, Writing – original draft. Iftikhar Ahmad: Su- [24] J. Gallardo-Lozano, E. Romero-Cadaval, V. Miñambres-Marcos, D. Vinnikov,
pervision, Data accusation, Writing – review & editing, Validation. T. Jalakas, H. Hõimoja, Grid reactive power compensation by using electric
Zubair Rehman: Co-supervision, Visualization. vehicles, in: 2014 Electric Power Quality and Supply Reliability Conference, PQ,
2014, pp. 19–24.
[25] M. Yilmaz, P.T. Krein, Review of battery charger topologies, charging power
Declaration of competing interest levels, and infrastructure for plug-in electric and hybrid vehicles, IEEE Trans.
Power Electron. 28 (5) (2013) 2151–2169.
[26] Abdelouahad Tahri, Hassan El Fadil, Josep M. Guerrero, Fouad Giri, F.Z. Chaoui,
The authors declare that they have no known competing finan-
Modeling and nonlinear control of electric power stage in hybrid electric
cial interests or personal relationships that could have appeared to vehicle, in: 2014 IEEE Conference on Control Applications, CCA, IEEE, 2014,
influence the work reported in this paper. pp. 641–646.

9
I. Ahmed et al. Journal of Energy Storage 52 (2022) 104813

[27] Gaizka Saldaña, et al., Analysis of the current electric battery models for electric [29] Akhil Garg, et al., Design and analysis of capacity models for Lithium-ion battery,
vehicle simulation, Energies 12 (14) (2019) 2750. Measurement 120 (2018) 114–120.
[28] L. Calearo, A. Thingvad, M. Marinelli, Modeling of battery electric vehicles for [30] Cristian Kunusch, et al., Sliding mode strategy for PEM fuel cells stacks breathing
degradation studies, in: 2019 54th International Universities Power Engineer- control using a super-twisting algorithm, IEEE Trans. Control Syst. Technol. 17
ing Conference, UPEC, 2019, pp. 1–6, http://dx.doi.org/10.1109/UPEC.2019. (1) (2008) 167–174.
8893474.

10