Received 9 December 2022, accepted 1 January 2023, date of publication 4 January 2023, date of current version 11 January 2023.

Digital Object Identifier 10.1109/ACCESS.2023.3234105

A Single Inductor Multi-Port Power Converter

for Electric Vehicle Applications
SHAIK REDDI KHASIM 1 , (Member, IEEE), C. DHANAMJAYULU 1, (Senior Member, IEEE),
AND S. M. MUYEEN 2 , (Senior Member, IEEE)
1 School of Electrical Engineering, Vellore Institute of Technology, Vellore, Tamil Nadu 632014, India
2 Department of Electrical Engineering, Qatar University, Doha, Qatar
Corresponding authors: C. Dhanamjayulu (dhanamjayulu.c@vit.ac.in) and S. M. Muyeen (sm.muyeen@qu.edu.qa)
Open Access funding provided by the Qatar National Library.

ABSTRACT This research presents a non-isolated multi-port power converter feasible to hybridize energy
alternatives in electric vehicles. Due to the hybridization of the various input sources, there are several
advantages in load power distribution in the system. Flexible control of discharging as well as the charging
process concerning the energy sources can be achieved. The developed converter can able to boost the voltage
levels with dual inputs such as a renewable solar PV and the other input as a battery and provides dual outputs
with various voltage levels, which can able to suit the converter fed for several loads like motor drive and the
low-rated loads like lighting and other auxiliary supplies in electric vehicles. Also, as the various voltages
appear at the output, this converter can be interfaced with multilevel inverters fed electric vehicle drivetrain.
The utilization of multilevel inverters reduces the total harmonic distortion and torque ripples in motor drives
in electric vehicles. The proposed converter consists of less number of components making the circuit simple
and cost-effective. With one inductor, two various modes are obtained for charging and discharging states
concerning the energy storage units. A state-space analysis is designed for all the converter operating modes
along with its control design. The proposed multi-port DC-DC converter is designed in MATLAB/Simulink
and tested in a laboratory environment with a hardware setup.

INDEX TERMS Multi-port DC-DC converters, hybrid energy storage system, state-space modeling, electric
vehicles.

I. INTRODUCTION point tracking (MPPT) technique [3], [4]. At the moment, the
Because of the world’s rapidly increasing population and use of solar PV energy systems meets the requirements for
energy demand, which is raising the cost of gas and oil electric vehicle applications. As long as the energy is avail-
and reducing the availability of fossil fuels, electric vehicles able, these provide energy to the required load and charge
(EVs) should be used instead of fossil-fuel cars. As a result, the battery. Hence, there exist many limitations of the solar
interest in developing EVs powered by clean and renew- PV system such as the various irradiance levels causing the
able energy sources to replace fossil-fuel vehicles has grown system less effective in power transfer to the load and the
steadily. Electric vehicles (EVs) are a promising alternative availability is not constant throughout the way. To overcome
for transportation-related applications because they can help these constraints, a solar PV system with a secondary battery
the environment by using renewable energy sources [1], [2]. source is used to provide energy from source to load. The
In the case of electric vehicles, the solar PV system is used hybridised combination of energy storage systems provides
as a clean energy source. Solar PV energy sources use solar continuous power transfer to the load and meets the load’s
energy to generate electricity. The maximum power from high power requirements at high speeds and terrains [5], [6],
the solar panel was extracted using the maximum power [7], [8]. Different voltage ratings are used to interface energy
storage systems such as batteries and supercapacitors. Imple-
The associate editor coordinating the review of this manuscript and menting the system with individual DC-DC converters for
approving it for publication was Alon Kuperman . each source makes the system bulky and complex, increasing

This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
VOLUME 11, 2023 3367
 S. R. Khasim et al.: Single Inductor Multi-Port Power Converter for Electric Vehicle Applications

FIGURE 1. (a) Conventional EV powertrain representing the converter (b) EV powertrain with the proposed converter configuration.

the system’s cost. As a result, a converter with multiple The author offers a three-input converter for the connection
voltage-rated inputs is required to handle multiple energies of batteries, solar cells, and fuel cells [21]. You can charge
and feed them into the system. In general, hybrid energy and discharge the battery from various sources and loads
storage systems are interfaced with a multi-input converter, with proper converter switching. Reference [22] Presents
which comes in a variety of variants depending on the isola- a systematic method for designing non-isolated topologies
tion requirement for non-isolation and isolation type DC-DC using a combination of buck, boost, sepic, and cuk type con-
converters. FIGURE 1 depicts the general layout of the multi- verters. There are two types of converters described: PVSC
input converter-fed electric motor. and PCSC, which are pulsing voltage and current sources.
In an isolated multi-port converter configuration, a high- As a voltage source, PVSC can be connected in series with
frequency transformer is used to produce isolation between an inductor to form a double input converter. Because the
the electric constraints. This allows for efficient isolation and PCSC is a current source, it may be used to build a double-
impedance matching on both sides of the converter. Leakage input converter by connecting it in series with a capacitor.
inductance is used as a storage device in isolated converters to In [20], the energy management strategy among the various
transmit power between both sides of the converter. Isolated inputs like battery, SC and the electric motor is suggested
dc-dc converters frequently include power converters in addi- for use in an electric automobile. Here, instead of employing
tion to the high-frequency transformer. The transformer phase two separate inductors as energy storage components are
shift between the primary and secondary voltage constraints used. When compared to two independent inductors, it is
has been changed to maintain efficient power transfer from claimed that employing the inductors with coupling can able
both converter ports [9], [10], [11]. There are several types of to save 22–26% in volume. In contrast, connected inductors
converters in this category, including half-bridge converters, have a larger volume than a single inductor. This converter
full-bridge converters, and combinational multiport convert- can also regenerate braking energy to the battery and SC.
ers [12], [13], [14]. Because of the transformer, these are Reference [23] proposes transferring load power across input
quite large. Inverters are required in these converters at the sources using a multiport converter with a single inductor.
transformer’s input port, with which dc can be converted to This converter can also be used to transfer power from one
ac supply and ac can be converted to dc using the converters. source to another. In [24], the author introduced a novel zero
As a result, various switches are used in all converter ter- voltage switching DC-DC converter. The layout of conven-
minals, increasing the losses, and the system losses increase tional converters utilized in electric vehicles is represented in
as transformer winding losses increase. These shortcomings FIGURE 2 (a) whereas the proposed converter configuration
are addressed in non-isolated multi-port converters, which are is represented in FIGURE 2 (b).
ideal for electric vehicle applications. Connected inductors, on the other hand, have a larger
Concerning the H-bridge, a non-isolated multiport con- volume than a single inductor. This converter also allows for
verter has been implemented [15]. In practise, various volt- the regeneration of braking energy to the battery and SC.
ages can be obtained by cascading H-bridge configurations Reference [23] suggests transferring load power across input
while taking into account various dc-link voltage levels. sources using a multiport converter with a single inductor.
Because of the converter’s coupling mechanism to multiple This converter can also be used to transfer power between
ports, the negative modes are not investigated [16], [17] sources. References [25] and [26] offer a multi-port con verter
Introduces the concept of a multiphase converter. The energy that may create numerous voltages at its output sections and
stored in the converter’s energy storage sources can be these are the preset values regardless of the variation of the
absorbed as well as delivered to the load. There exists an indi- load power and input voltage interfaced with PV systems.
vidual inductor for each source and hence this is considered a Reference [27] Presents a novel multi-output buck converter
drawback of the converter [18], [19], [20] also cost-efficient control technique that delivers suitable dynamic performance.
three-port converter for EV/HEV applications is proposed. However, this converter is worthless in applications like

3368 VOLUME 11, 2023

S. R. Khasim et al.: Single Inductor Multi-Port Power Converter for Electric Vehicle Applications

electric cars, where various input energy excitations like analogous circuit depicted in Fig. 2.
solar and a battery are utilized. This can be solved by incor-   
q(VPV + RS IPV )
 
porating multi-port converters. Reference [28] Describes IPV = IP − I0 exp −1
NS ATK
a non-isolated high step-up multi-port converter and its 
(VPV + RS IPV )
performance evaluation using a variety of parameters. − (1)
References [29] and [30] evaluate the MPPT of a solar PV NS RP
converter under partial shading conditions. Another flaw in The series and parallel resistances are Rs and Rp, respec-
the proposed converter is its inability to transmit energy tively, and the count of PV cells connected in both series and
across input sources. parallel respectively are Ns and Np . K represents the Boltz-
A multi-port non-isolated converter is suggested in this mann constant (1.3806503102 J/K), T gives the temperature,
work, which is based on the mixture design of multiple inputs A is the semiconductor device’s ideality factor, and Iph is
and outputs of the converter. In comparison to previous sce- the obtained current, which is dependent on the respective
narios, the suggested converter contains fewer components. irradiation and nominal temperature stated in equation (2)
This converter can regulate the flow of electricity between 
G

sources and loads. Furthermore, the suggested converter IP = [ISOC−STC + Ki (T − TSTC )] − (2)
GSTC
includes many outputs, each of which can have a different
voltage level. ISOC−STC is a short circuit current (SCC) under basic operat-
The remainder of the paper is presented as follows. ing and testing conditions (STC), Ki represents the coefficient
Section II provides the modeling of solar PV and MPPT. of SCC, G (in W/m2 ) provides the respective irradiation of the
Section III explains the construction and modes of opera- solar cell, GSTC (1000W/m2 ) provides the irradiation at STC,
tion of the developed converter. The dynamic modeling is and TSTC (25◦ C) is the cell temperature at STC.
shown in Section IV. Whereas the control architecture of ISOC−STC + Ki (T − TSTC )
 
the converter is discussed in section V. Both simulation and IO = (3)
exp [(VOCV −STC + KV (T − TSCC ) /AVth )]
experimental results are discussed in Sections VI and the
article is concluded in Section VII. where VVOCV−STC denotes the open-circuited voltage (OCV)
provided at STC, Kv denotes the OCV coefficient and Vth
II. MODELING OF PV ARRAY denotes the cell’s thermal equivalent voltage and these values
The model of a solar PV cell plays a prominent role in under- may be found on the module’s datasheet.
standing a solar PV system. FIGURE 1 shows the developed  
qVPV
 
VPV

circuit in its entirety, which includes solar PV, a type of dual PPV = VPV × NP IPh − IO exp −
NS AKT NS
input multi-port DC-DC converter. Analogous circuits with
(4)
I-V and P-V aspects, irradiance and temperature effects, and
partial shading circumstances are all models that may be Fig. 3 depicts the I-V/P-V properties of a solar cell. Repre-
used to describe this. Photonic energy is denoted by photo, sentation of a graph shows the operating point of a solar PV
while voltaic energy is denoted by voltaic, meaning that pho- is inherently unstable, varying constantly from zero to OCV.
tonic energy drawn from the sun is transformed into electric The maximum power point is a single point in this process
energy [31]. Solar modules are made up of various modules that produces maximum power based on the specified design
with respective solar cells. Reference [32] Utilizes P-N semi- of the solar module at varying irradiance and temperatures
conductor diodes. The output of developed solar PV tends to (MPP). VMPP , and IMPP are the relative voltages and currents
alter as temperature and ambient conditions change [33]. As a at this moment, as indicated in FIGURE 3.
result, consider the following aspects when modeling a solar
PV system. The internal resistance Rs and Rp are linked in
series and parallel to the diode in a solar cell, which is known
as an analogous circuit depicted in FIGURE 2.

FIGURE 2. A solar cell’s equivalent circuit. FIGURE 3. Solar Cell I-V Characteristics.

The internal resistance Rs and Rp are linked in series and Irradiance, temperature, and the quantity of series-
parallel to the diode in a solar cell, which is known as an connected strings and parallel-connected strings all influence

VOLUME 11, 2023 3369

 S. R. Khasim et al.: Single Inductor Multi-Port Power Converter for Electric Vehicle Applications

the voltage and current generated by the PV array. As a result, B. MPPT WITH ARTIFICIAL NEURAL NETWORK
it is necessary to make an informed decision on the sort of This article uses an ANN to track the greatest PowerPoint.
solar panel to choose. 1Soltech 1STH-215-P panel is consid- A three-layer ANN is employed to get to MPP, as illustrated
ered from available data of the solar modules in MATLAB in FIGURE 4.
with 2 parallel strings and 2 series modules connected for The ANN’s input variables are temperature T and irradi-
one string in this work. Table 1 lists the characteristics of the ance G, whereas the output variable is MPP voltage Vmpp .
chosen solar panel, as well as the readings for one parallel- Some data must be acquired as input as well as the output
connected string and one series module at 1000 W/m2 irradi- variables for training. The neuron weights in distinct levels
◦
ance and a temperature of 250 C . are therefore calculated as a consequence. The solar PV data
can be collected and programmed in MATLAB. ANN may
A. MPPT CONTROLLER be taught in several different methods. The backpropagation
The maximum power varies based on the environment’s tem- approach is used for training the error of ANN in this article.
perature and irradiance. MPPT with a real-time implemen-
tation is an essential component of any solar module system
since the maximum accessible energy of solar cells fluctuates
concerning the weather. The developed MPPT approaches
may be divided into three classes in the technical literature 10:
Techniques concerning artificial intelligence, methods con-
cerning direct methods, and ways based on indirect methods.
The MPP is found using direct procedures, also known as
genuine searching methods, which include repeatedly per-
turbing the PV array’s operational point. In this section,
the MPPT schemes widely utilized are Hill Climbing HC,
Perturb and Observe P&O and Incremental Conductance INC FIGURE 4. The structure of a neural network.
in PV systems. The P&O approach comprises analyzing the
operating voltage of PV to achieve the MPP. The respective Once the training of the ANN is completed the neuron
duty cycle of the developed converter is affected by hill- weights are to be specified such as the Vmpp is the output
climbing methods, such as the P&O scheme [14]. Although of the ANN for the input values of T and G. Using the V-I
these technologies are characterized by their simplicity, they characteristic of the modeled PV, the current at the MPPT
are limited to low-rated applications due to their steady state. Impp can now be calculated. As a result, multiplying Vmpp
The steady-state oscillation can be reduced utilization of and Impp yields maximum power _Pmax _.
the incremental conductance technique can decrease steady- The PV and the MPPT tracker shown in Fig. 4 are made up
state oscillation since the plot representing the power versus of a converter and an ANN-based control unit. The following
voltage at the MPP is zero. ANN and indirect approaches have equation is used to manage a chopper with a particular Vmpp
been developed to increase the dynamic behavior and per- and Impp duty cycle at any given time:
formance of the MPPT. Artificial intelligence techniques that s
focus on the nonlinear features of PV arrays provide a speedy Vmpp Io
but computationally challenging solution to the MPPT issue. D=1− × (5)
Impp V0
The indirect methods function by calculating the array’s MPP
based on the output characteristics. The open-circuit voltage
OCV and the short-circuit current SCC approaches can be III. PROPOSED DC-DC CONVERTER
used to compute the MPP. The specifications of the solar PV A multi-port converter is developed with a single inductor
module are represented in TABLE 1. four switches S1 , S2 , S3 , S4 and four diodes D1 , D2 , D3 , D4
along with two capacitors C1 and C2 . The converter is excited
TABLE 1. Parameters of 215W solar panel.
with two input energy storage sources such as a solar PV input
Vpv and battery input Vbat . As one input energy source cannot
able to feed the power requirement in an electric vehicle, the
requirement of two energy sources is essential. The proposed
double input and double output type converter are shown in
FIGURE 5 [23]. The power fed to the developed converter
by using the two input energy sources has been analysed
using the load resistances R1 and R2 . The effective switching
of the converter is controlled by the power flow among the
input energy sources and load resistances by enhancing the

3370 VOLUME 11, 2023

S. R. Khasim et al.: Single Inductor Multi-Port Power Converter for Electric Vehicle Applications

In the low power requirement mode, when the required

current is drawn is less and the battery is getting charged, it is
feasible to operate the converter in discontinuous conduction
mode (DCM). Hence, this model is analyzed in section IV.
The analysis is done in such a way that each input source
is analyzed separately. The converter in one mode can be
considered as a single input operation. There exist two major
operations of the converter such as battery discharging mode
and charging mode which are represented in detail are repre-
sented below.
FIGURE 5. Proposed multi-port converter with two inputs and two
outputs [23]. A. BATTERY DISCHARGING MODE
During this mode, the loads are supplied by two input power
sources Vpv and Vbat . The switches S1 , S3 , S4 are active and
feed the power to the load and the switch S2 is inactive. Every
input voltages. The output voltages can be changed concern- switch has its operation in process of power delivered from
ing the multilevel inverter input voltages. source to load. S1 adjusts the current through the battery to
With the interfacing solar PV energy source with a bat- a required value by altering the current through the inductor.
tery, the proposed converter is well suited for its robust- The switch S3 keeps the total output voltage VT = Vpv + Vbat
ness. The power feeding from the respective sources along at a constant level. The voltage output Vpv is also controlled
with the load utilizing the proposed circuit components is by the switch S4 . The adjustment of VT and Vpv affects the
represented in FIGURE 5. The load resistances R1 and output voltage Vbat . FIGURE 7 represents inductor voltage
R2 can be modeled concerning the equivalent resistance of and current waveforms, as well as the gate signals of switches.
the motor and the multilevel inverter input voltages. The There exist four different operation modes in a single switch-
proposed converter is capable of interfacing with various ing period:
multilevel inverters. The multilevel inverters require asym-
metrical voltage source requirements which can be fed with 1) SWITCHING CONDITION 1 (0 < t < D3 T )
the proposed converter. The four switches can able to regulate During this operation, the switches S1 and S3 are active. Due
the flow of power among the sources to the output voltages to the diodes D1 and D2 being in reverse bias when S1 is
and makes the voltages more flexible with the multilevel switched on, switch S4 is inactive. Diode D0 is in reverse bias
inverter. since S3 is ON. FIGURE 6(a) represents the equivalent model
In the proposed converter the source Vpv delivers power of the developed converter. Here, Vbat charges the inductor L,
to Vbat and charges the battery but Vbat sources do not feed which causes the current through the inductor to rise. The
the power to source Vpv . In the case of electric vehicle capacitors C1 and C2 can able to deliver energy to the load
applications, solar PV cannot be charged from the battery and resistances R1 and R2 by discharging. The respective equa-
the battery is charged from an external source or solar PV tions of the inductor and capacitors are represented below in
source. In this article, solar PV can be utilized as a prominent equation (6).
source of the system and the battery is the energy storage
diL
system device. Concerning the power utilization of the battery L = Vpv
and state of charge, there exist two operating modes of the dt
dV1 V1
developed converter. For all modes of the converter, one C1 =−
switch is inactive by activating the three switches. Both input dt R1
sources are active and feed the power at high load power dV2 V2
C2 =− (6)
requirements. During this mode of operation, switches S1 , S3 , dt R2
and S4 are active and feed the power to the load and switch
S2 is inactive. In the case when the load power requirement 2) SWITCHING CONDITION 2 (D3 T < t < D1 T )
is less, the Vbat will be charged from the source Vpv and During this operation, switch S1 remains active while switch
the only source Vpv supplies the energy to the load. In this S3 is inactive. Switch S4 remains off when S1 is switched on
condition, the switches S1 , S2 , and S4 become an active state because the respective diodes D1 and D2 are in reverse bias.
and switch S3 is in an off state. As the output voltage is The equivalent circuit model of the developed converter is
dependent on the current obtained from the converter, power represented in FIGURE 6(b). During this mode, Vpv charges
should be balanced from the input powers to the load and the the inductor L and increases the current through the inductor.
ripple current drawn needs to be reduced. Hence for these Additionally, the respective capacitors C1 and C2 are getting
requirements, the steady-state and dynamic activity has been discharged supplying the energy stored to the R1 and R2 load
analyzed in the continuous conduction mode (CCM). resistances respectively. The equations of the capacitors and

VOLUME 11, 2023 3371

 S. R. Khasim et al.: Single Inductor Multi-Port Power Converter for Electric Vehicle Applications

FIGURE 6. Battery discharging modes: (a) Mode-1 (b) Mode-2 (c) Mode-3 (d) Mode-4.

dV1 V1
C1 =−
dt R1
dV2 V2
C2 =− (7)
dt R2
3) SWITCHING CONDITION 3 (D1 T < t < D4 T )
During this operation, switch S1 is inactive, while switch S3
is still turned off with the switch S4 is active. The diode
D2 is in reverse bias. FIGURE 6(c) represents the analogous
circuit model of the developed converter. The inductor gets
discharged by storing the energy which can be distributed to
the load with C1 and R1 leading to a decrease in inductor
current. In this condition, the capacitor C1 is getting charged,
while C2 is discharged, supplying the energy stored to the
load resistance R2 . The equations of the capacitors and induc-
tor during this operation specifically are represented below in
equation (8).
diL
L = Vpv − V1
dt
dV1 V1
C1 = iL −
dt R1
dV2 V2
C2 =− (8)
dt R2
4) SWITCHING CONDITION 4 (D4 T < t < T )
During this operation, all switches become inactive. As a
FIGURE 7. Typical steady-state waveform while discharging a battery.
result, diode D2 is in forwarding bias. The inductor L gets
discharged by storing the energy supplied to the capacitors C1
inductors during this operation specifically are represented and C2 with the R1 and R2 load resistances. The capacitors
below in equation (7). C1 and C2 are getting charged. The analogous circuit model
diL of the developed converter is represented in FIGURE 6(d).
L = Vpv
dt The equations of the capacitors and inductors during this

3372 VOLUME 11, 2023

S. R. Khasim et al.: Single Inductor Multi-Port Power Converter for Electric Vehicle Applications

operation specifically are represented below in equation (9).

diL
L = Vpv − (V 1 + V2 )
dt
dV1 V1
C1 = iL −
dt R1
dV2 V2
C2 = iL − (9)
dt R2

B. BATTERY CHARGING MODE

During the mode of operation, Vpv feeds loads by providing
power to Vbat . This situation appears in the case where the
load-rated power is less and the battery gets charged. During
this condition, switches S1 , S2 and S4 are operational, while
switch S3 is inactive. This mode, like the converter in the
above mode of operation, takes into account a set duty for
each switch. The total output voltage, VT = Vpv + Vbat to be
adjusted to the appropriate level by turning on S1 . Switch S2
keeps the battery charging current (Ibat ) at a required level.
The voltage output Vpv is likewise accessed by the switch S4 .
The modification of VT and Vpv controls the voltage out-
put Vbat . FIGURE 8 represents the gate switching signals of
switches as well as the waveforms representing the voltage
and current of the inductor. As per the distinct switch states,
the four alternative operating modes exist in one switching
period addressed below.
FIGURE 8. The typical steady-state waveform in the charging mode of a
1) SWITCHING CONDITION 1 (0 < t < D1 T ) battery.
During this mode of operation, the switch S1 is active causing
S2 and S4 to be in reverse bias and inactive. Furthermore,
because the diode D2 is reverse-biased, it does not conduct. below in equation (11).
FIGURE 9(a) represents the analogous circuit model of the diL
developed converter operating in this mode. Vpv charges L = Vpv − Vbat
dt
inductor L increasing the inductor current. Where the capac- dV1 V1
itors of C1 and C2 get discharged by supplying the energy C1 =−
dt R1
stored to R1 and R2 load resistances. The equations of the dV2 V2
capacitors and inductors during this operation specifically are C2 =− (11)
dt R2
represented below in equation (10).
diL 3) SWITCHING CONDITION 3 (D2 T < t < D4 T )
L = Vpv
dt During this mode, the switch S1 and switch S2 become inac-
dV1 V1 tive and the switch S4 is switched on in this state. The diode
C1 =−
dt R1 D2 is reverse bias. The circuit of the developed converter is
dV2 V2 depicted in FIGURE 9(c). The Inductor gets discharged and
C2 =− (10)
dt R2 distributes its energy stored to C1 and R1 , resulting in a drop
in inductor current. The capacitors C1 and C2 are charged
2) SWITCHING CONDITION 2 (D1 T < t < D2 T )
with a capacitor C2 is discharged, delivering its energy stored
to the R2 load resistance. The equations of the capacitors and
During this mode, the switch S1 becomes inactive, while the
inductors during this operation specifically are represented
switch S2 is turned on. As the diodes D1 and D2 are in reverse
below in equation (12).
bias, S4 is inactive. FIGURE 9(b) represents the circuit of
the developed converter. Since the inductor current falls over diL
L = Vpv − V1
this period, and inductor distributes its energy stored to the dt
battery Vbat . The capacitors C1 and C2 are discharged in dV1 V1
C1 = iL −
this mode, delivering the energy stored to R1 and R2 load dt R1
resistances respectively. The equations of the capacitors and dV2 V2
C2 =− (12)
inductors during this operation specifically are represented dt R2

VOLUME 11, 2023 3373

 S. R. Khasim et al.: Single Inductor Multi-Port Power Converter for Electric Vehicle Applications

FIGURE 9. Battery charging modes: (a) Mode-1 (b) Mode-2 (c) Mode-3 (d) Mode-4.

4) SWITCHING CONDITION 4 (D4 T < t < T ) there are several dynamic models. As a result, distinct con-
During this mode, all the switches become inactive. As a trollers must be created for each mode of operation.
result, the diode D2 is in forward bias. The circuit of the
developed converter is depicted in FIGURE 9(d). The induc- 1) BATTERY DISCHARGING MODE
tor gets discharged in this condition and the stored energy is The basis for optimal controller design is the small-signal
delivered to the capacitors C1 and C2 as well as R1 and R2 model. An effective model will be especially useful for such
load resistances. The charging of the capacitors C1 and C2 a multi-port converter to achieve closed-loop-based control
can be made. The equations of the capacitors and inductor importantly for optimizing the dynamics of the converter. The
during this operation specifically are represented below in multiport converter, unlike ordinary consisting of single input
equation (13). and output converters of a high-order type system, represents
derivation of transfer functions is time-consuming; as a result,
diL obtaining values of the respective zeros and poles for the
L = Vpv − (V 1 + V2 ) analysis is a challenging task. However, the dynamics of
dt
dV1 V1 the plant can be expressed as a matrix, which necessitates
C1 = iL − the use of computer software to draw bode plot graphs of
dt R1
dV2 V2 various transfer functions. The duty ratios, state variables
C2 = iL − (13) and input energy excitations have two components dc values
dt R2
(X, D, V) and the respective perturbations (x, d, v), according
to the small-signal modeling approach [30]. The perturba-
IV. DESIGNING OF THE PROPOSED CONVERTER
tions are expected to be minor and do not fluctuate much
A. DYNAMIC MODELLING
over a single switching time. Hence the developed converter
The suggested converter can able to be controlled by switches
has the following equation, where the state variables are
S1 , S2 , S3 and S4 as indicated in the previous section. Each
inductor current iL (t) and the respective capacitor voltages
switch has a distinct function. Output voltages and battery
V1 (t) and V2 (t). The small-signal equations are derived by
charging or discharging current can be adjusted by properly
substituting (14) into (6)–(9), applying making to four state
regulating switch duty cycles. The converter’s closed-loop
equations multiplied by the values of the respective duty cycle
controller must first be designed using a dynamic model.
ignoring second-order terms:
As previously stated, the developed converter has two major
modes of operation: while the battery discharge, two inputs
iL (t) = IL + IˆL (t)
supply energy to load and in the case of charging the battery
the source Vpv not only delivers load-rated power but it V1 (t) = V1 + Vˆ1 (t)
charges Vbat . For each of the converter’s operation modes, V2 (t) = V2 + Vˆ2 (t)

3374 VOLUME 11, 2023

S. R. Khasim et al.: Single Inductor Multi-Port Power Converter for Electric Vehicle Applications

d1 (t) = D1 + dˆ1 (t) cycles of the switches are obtained with steady-state equa-
d2 (t) = D2 + dˆ2 (t) tions represented as follows:
d3 (t) D3 + dˆ3 (t)
    
= V1 Vbat − Vpv V2 D1 V1 + V2 − Vpv
d4 (t) = D4 + dˆ4 (t) (14)
 R1 Ib V1 0   D3  =  R1 Ib 
0 V2 R2 Ib D4 R2 Ib
d IˆL (t)
L = (V bat − Vpv )dˆ3 (t) (20)
dt
ˆ (t) + (1 − D3 )Vˆpv (t)
+ D̂3 Vbat As a result of the above matrix equation, the duty cycles of
− (1 − D1 )Vˆ1 (t) the respective switches are calculated and presented in the A,
B, C, and D matrices. The three control variables d1 (t), d3 (t),
+ (D4 − 1)Vˆ2 (t) + V1 dˆ1 (t) + V2 dˆ4 (t)
and d4 (t) in system small-signal models (t) the state variables.
d Vˆ1 (t) Vˆ1 (t) The small-signal model yields the following transfer function
C1 = −IL dˆ1 (t) + (1 − D1 )IˆL (t) − )
dt R1 matrix for the converter:
d Vˆ2 (t) Vˆ2 (t)
C2 = −IL dˆ4 (t) + (1 − D4 )IˆL (t) − (15) G = C (SI − A)−1 B + D (21)
dt R2
As a result, the system is described in a matrix by utilizing where
an approach of a state-space type model, with state variables y = Gu (22)
iL (t), Vpv (t) and Vbat (t). The state-space model is as follows:
The control variables count is indicated by the rank of the
dX
= AX + BU transfer function matrix, G in this study is 3 depending on the
dt number of control variables and (17).
Y = CX + DU (16)     
y1 g11 g12 g13 u1
where X denotes the state variables, U denotes the control  y2  =  g21 g22 g23   u2  (23)
inputs d1 (t), d3 (t), and d4 (t), and Y denotes the system y3 g31 g32 g33 u3
outputs Vpv (t), VT (t) and ib (t). As a result, the matrices like
X, Y and U take the below type: where gij denotes the transfer function among yi and uj , and
y and u are the respective system output and input vectors.
IˆL (t) Vˆ1 (t) dˆ4 (t)
     
As a result, there are three transfer functions:
X =  Vˆ1 (t)  , Y =  VˆT (t)  , U =  dˆ3 (t)  . (17)
Vˆ2 (t) Iˆb (t) dˆ1 (t) Vˆ1 (s)
= g11
d4 (s)
Using (15) and state equations (16) to fill in the A, B, C, and
VˆT (s)
D matrices, the following result is obtained: = g22
d3 (s)
 (D1 −1) (D4 −1) 
0 L L Iˆb (s)
 (1−D )
1 1
 = g23 (24)
A =  C1
 − R1 C1 0   d1 (s)
(1−D4 ) 1
0 − R2 C2
 C2 (V −V )
 2) BATTERY CHARGING MODE
V2 V1 bat pv
 L L L The switches S1 , S2 and S4 become on state, while switch
− CIL1

B= 0 0 
 S3 is inactive, during this mode of operation, the developed
− IL 0 0 converter, in which Vpv supplies energy to loads and Vbat .
 C1    The first small-signal model can be created by putting (14)
0 1 0 0 0 0 into (10)–(13) by applying the analysis to four state equations
C = 0 1 1, D = 0 0 0  (18) respectively multiplied by the values of the corresponding
D3 0 0 0 IL 0 duty cycles by then ignoring the second-order terms, resulting
where Vpv and Vbat are the respective voltages of the given in small-signal equations represented below:
input energy sources. V1 and V2 are output voltages as well. d IˆL (t)
Except for the duty cycle of switches D1 , D3 , D4 and the L = Vˆpv (t) + Vbat dˆ1 (t)
inductor current IL , all parameters in A, B, C, and D matrices dt
+ (D1 − D2 )Vbat ˆ (t) + (V 1 − Vbat )dˆ2 (t)
are known. The following equation can be used to calculate
the inductor current: + V2 dˆ4 (t) + (D2 − 1)Vˆ1 (t) − (1 − D4 )Vˆ2 (t)
Ib d V1 (t)
ˆ Vˆ1 (t)
IL = (19) C1 = −IL dˆ2 (t) + (1 − D1 )IˆL (t) −
D3 dt R1
where Ib is the current drawn by the battery D1 , D3 , and D4 d V2 (t)
ˆ V2 (t)
ˆ
C2 = −IL dˆ4 (t) + (1 − D4 )IˆL (t) − (25)
are the only unknown parameters of stated matrices. The duty dt R2

VOLUME 11, 2023 3375

 S. R. Khasim et al.: Single Inductor Multi-Port Power Converter for Electric Vehicle Applications

FIGURE 10. The control scheme of the developed converter (a) Discharging mode (b) Charging mode.

State-space equations can be used to express these equations. The small signal converter model yields the following transfer
The state variables in this operation mode, which is compara- function matrix for the converter:
ble to the discharging mode of a battery with iL (t), V1 (t), and
V2 (t). The following are examples of state variables, input G = C (SI − A)−1 B + D (30)
and output matrices:
where
IˆL (t) Vˆ1 (t) dˆ4 (t)
     
y = Gu (31)
X =  Vˆ1 (t)  , y =  VˆT (t)  , u =  dˆ1 (t)  . (26)
Vˆ2 (t) Iˆb (t) dˆ2 (t) The three transfer functions are represented as follows, where
where VT (t) = V1 (t) + V2 (t). Hence for this mode, there y and u are the respective output and input vectors of the
exists the matrices A, B, C and D represented below: system.
(D2 −1) (D4 −1) Vˆ1 (s)
 
0
 (1−D ) − L1 L
d4 (s)
= g11
A =  C1 2 R1 C1 0  
(1−D4 )
0 1
− R2 C2 VˆT (s)
 C2 (V −V )  d1 (s)
= g22
Vbat 1 bat V2
 L L
− CIL1
L
Iˆb (s)
B= 0 0   = g33 (32)
0 0 − CIL1 d2 (s)
It’s also worth noting that in the charging of the battery, in the
   
0 1 0 0 0 0
C = 0 1 1,D =  0 0 0 situation where the load power and battery current become
D2 − D1 0 0 −I L IL 0 low, the converter may switch to discontinuous conduction
(27) mode. The inductor current must be less than its current ripple
for converters in DCM. The equations for the developed
Except for the duty cycle of switches D1 , D2 , D4 and the dc converter while charging the battery can be determined by
value of inductor current IL , all parameters in the aforemen- making the voltage across the inductor and current through
tioned matrices are known. The following equation applies to the capacitor at the switching period:
the inductor current:
Ib Vbat D1 + (V1 − Vbat )D2 + V2 D4 = V1 + V2 − Vpv
IL = (28) V1
D2 − D1 −IL D2 + IL =
where Ib is the desired battery current value. D1 , D2 and D4 R1
are the unknown parameters of stated matrices. Steady-state V2
−IL D4 + IL = (33)
equations are used to calculate the duty cycle of switches. R2
In battery charging mode, the inductor current ripple can also
    
Vbat V1 − Vbat V2 D1 V1 + V2 − Vpv
 −V1 V1 +R1 Ib 0   D2  =  R1 Ib  be written as
−V2 V2 R2 Ib D4 R2 Ib Vpv D1 Ts
1IL = (34)
(29) L

3376 VOLUME 11, 2023

S. R. Khasim et al.: Single Inductor Multi-Port Power Converter for Electric Vehicle Applications

As a result, if the following conditions are met, the converter TABLE 2. Parameters of the converter.
will work in DCM from (28), (33), and (34).
V2 Vpv D1 Ts
IL = <
R1 (1 − D1 ) L
V2 Vpv D1 Ts
IL = <
R2 (1 − D4 ) L
Ib Vpv D1 Ts
IL = < (35)
(D2 − D1 ) L
V1 Lfs
R1 >
D1 (1 − D2 )Vpv seen from the produced bode plots that the phase of the
V2 Lfs crossover frequency of g11 is –179.99◦ . As a result, the
R2 >
D1 (1 − D4 )Vpv phase margin becomes insufficient, and the closed-loop sys-
D1 (D2 − D1 )Vpv tem is inherently not stable. As a result, a lead compensator
Ib > (36) increases the phase margin and improves the stability of the
Lfs
system.
In reality, (36) verifies that when the battery charging current
is low with light loading of output, the proposed converter S +Z
K (s) = K (39)
will operate in DCM. S +P
K = 2.9, Z = 906.05 and P = 7641.6 in this equation.
V. CONTROLLER DESIGN Bode graphs after adjustment are given in FIGURE 11(b).
The two operating modes and the respective transfer func- The system’s stability has been improved by using a lead
tions are initiated in the above sections. Because of the compensator. In addition, the transfer function of g22 is (34).
distinct transfer functions in both operating modes of the Bode graphs for g22 are shown in FIGURE 12(a) and 12(b).
converter, a separate controller must be constructed for each Bode diagrams reveal that the g22 phase margin is insuffi-
model. Control block schematics of the converter in the two cient, necessitating the use of a lead compensator. As a result,
modes of operation are illustrated in FIGURE 10. R1 = R2 = a lead compensator is based on the utilization of the following
35 in battery discharging mode and R1 = R2 = 70 in the transfer function.
mode of charging the battery.
S + 2098.9
which (35) can be rewritten as in (37) and (38), shown at Klead (s) = 2.8960 (40)
the bottom of the page. S + 16982
However, utilizing the lead compensator lowers the dc gain
A. CONTROLLER FOR BATTERY DISCHARGING MODE which increases the steady-state error. A lag compensator is
There exist three transfer functions in this way as explained in to be incorporated into the given system to solve the problem.
Section IV. To construct the system compensators, computer To avoid the lag compensator’s effect on the lead compen-
software must obtain frequency-domain analysis for every sator, the lag compensator zero is set ten times lower than the
transfer function. System compensators provide the appropri- crossover frequency.
ate steady-state error and the respective phase margin, as well As a result, the lag transfer function is as follows:
as excellent stability and bandwidth. The transfer function of
S +Z
g11 is (37) given at the bottom of the page using the A, B, C, Klag (s) = K (41)
and D matrices shown in section IV and the transfer function S +P
of the respective matrix G. where K = 1, P = 58.345, Z= 583.45. In FIGURE 12(b),
The open-loop system of the bode diagram of g11 bode plots after the application of compensation by using the
for the converter in its simulation parameters is repre- lead-lag compensators are represented. Switch S1 controls the
sented in Table 2, designed using MATLAB software. current of source 2 (battery), as previously stated. In actuality,
FIGURE 11(a) and 11(b) depict bode diagrams. It can be S1 handles inductor current to set the battery current to the

   
V2 (1−D1 ) V2 (1−D1 ) IL (1−D1 )(D4 −1)
V1 (s)
ˆ LC1 S + LR2 C 1 C2 − LC1 C2
g11 = = 2 R R C +(D −1)2 R R C
(37)
d4 (s)
     
ˆ 3 R1 C 1 +R2 C 2 2 L+(1−D 1 ) 1 2 2 4 1 2 1 R1 (1−D1 )2 +(D4 −1)2 R2
S + R1 R C C2 S + LR1 R2 C 1 C2 S + LR1 R2 C 1 C2
2 1     
Vbat −Vpv ) 1−D1 D4 −1 1−D1 D4 −1 Vbat −Vpv )
VˆT (s) L C1 − C2 S + R2 C C2 − R1 C1 C2
1 L
g22 = = (38)
dˆ3 (s) L+(1−D1 )2 R1 R2 C2 +(D4 −1)2 R1 R2 C 1
     2 2

R1 C 1 +R2 C2
3
S + 2
S +
R1 R2 C 1 C2 LR1 R2 C 1 C2 S + R1 (1−D1 ) +(D4 −1) R2
LR1 R2 C 1 C2

VOLUME 11, 2023 3377

 S. R. Khasim et al.: Single Inductor Multi-Port Power Converter for Electric Vehicle Applications

FIGURE 11. Simulations of the bode plots of g11 (s) (a) Before compensation (b) After applying the compensator.

FIGURE 12. Simulations of the bode plots of g22 (s) (a) Before compensation (b) After applying the compensator.

required value, as illustrated in (42), as shown at the bottom B. CONTROLLER DESIGN FOR BATTERY CHARGING MODE
of the page. During this mode of operation, the switches S1 , S2 and S4 are
FIGURE 13(a) shows the g33 bode plot. As can be seen in turned on and S3 is in an off state during this as described
the diagram, the phase margin is ample, resulting in a stable in the previous section. Switch S4 is utilized to regulate
system with a low dc gain. As a result, the steady-state error the voltage output V1 . FIGURE 14 (a) shows the g11 bode
is extremely valuable. A lag compensator is used to reduce diagram. A lead compensator is needed due to a lack of phase
steady-state inaccuracy. margin. As a result, the following transfer function is used to
S + 200 construct the lead compensator:
Klag (s) = (43)
S +5
Bode graphs after applying the lag compensator are presented S + 482.7
Klead (s) = 2.9 (44)
in FIGURE 13(b). S + 4065.8

  h   i  
D3 V 1 IL (1−D1 )D3
iˆb (s)
V1 D3
L S 2 + D3LV1 R11C + R21C2 + IL (1−D LC 1
1 )D3
S + LR R C
1 2 1 2 C + LR C C
2 1 2
1
g33 = = (42)
dˆ1 (s) L+(1−D1 )2 R1 R2 C2 +(D4 −1)2 R1 R2 C 1
     
3 R1 C 1 +R2 C2 2 R1 (1−D1 )2 +(D4 −1)2 R2
S + R1 R C C2 S + LR1 R C C2 S + LR1 R C C2
2 1 2 1 2 1

3378 VOLUME 11, 2023

S. R. Khasim et al.: Single Inductor Multi-Port Power Converter for Electric Vehicle Applications

FIGURE 13. Simulations of the bode plots of g33 (s) (a) Before compensation (b) After applying the compensator.

FIGURE 14. Simulations of the bode plot of g11 (s) (a) Before compensation (b) After applying the compensator.

A lag-type compensator is introduced to the proposed system the g33 bode diagram. The phase margin becomes sufficient
to reduce steady-state error: and hence there will be no need for stabilizing the compen-
S + 140 sator, as shown in the diagram. Only a lag compensator is
Klag (s) = (45) employed to reduce steady-state inaccuracy.
S + 14
S + 267.4
The bode plot after the application of the lead-lag compen- Klag (s) = (48)
S + 2.67
sator is illustrated in FIGURE 14(b). Switch S1 is also utilized
to regulate the total output voltage VT in this operation mode. The bode diagram after compensation is given in
FIGURE 15(a) and 15(b) show the g22 bode diagram. FIGURE 16(b).
Because g22 phase margin is insufficient, the compensator The inductor and capacitor values can be obtained using
utilized is represented below the following relations:
(V in + Vb ) × d2
S + 585.4 L = (49)
Klead (s) = 2.9 (46) 2 × 1V01 × fs
S + 4937
V01 × d1
In addition, a lag compensator is used to reduce steady- C1 = C2 = (50)
2 × 1V01 × fs × Z01
state error:
S + 170 VI. RESULTS AND DISCUSSION
Klag (s) = (47)
S + 17 A. SIMULATION RESULTS
The bode diagram after compensation is given in The performance of the developed converter is analyzed and
FIGURE 15(b). In addition, switch S2 is utilized to regulate designed using MATLAB software. The simulation param-
the battery charging current. FIGURE 16(a) and 16(b) show eters of the developed converter are shown in TABLE 2.

VOLUME 11, 2023 3379

 S. R. Khasim et al.: Single Inductor Multi-Port Power Converter for Electric Vehicle Applications

FIGURE 15. Simulations of the bode plot of g22 (s) (a) Before compensation (b) After applying the compensator.

FIGURE 16. Simulations of the bode plot of g33 (s) (a) Before compensation (b) After applying the compensator.

Vpv = 35V, Vbat = 48V are the input voltage sources. The
battery model is utilized as an input source in simulation 2.
The converter’s output voltages should be regulated at
V1ref = 80V and V2ref = 40V. Hence, the total output voltage
should be able to regulate at VTref = 120V. In addition, for
battery discharge and charging modes, battery current should
be regulated at Ibref = 3A and Ibref = 0.9A, respectively. For
the battery draining and charging modes R1 = R2 = 35 and FIGURE 17. Output voltages in battery discharging mode.
R1 = R2 = 70 are the load resistances respectively.
Initially, the converter is in the discharging battery mode,
with the switches S1 , S3 , and S4 active, as shown in previous
sections. Each switch is controlled by a compensator, which
is built in Section V. FIGURE 17 depicts the output voltages
V1 and VT. Regulating the output voltage represented yields
the required values. FIGURE 18 depicts the battery cur-
rent. The battery current is compared to the reference value.
FIGURE 19 illustrates the inductor current. At a specific time
(t = 0.01s), the battery source’s reference current is abruptly
FIGURE 18. Output voltages in battery discharging mode.
increased from 3A to 4A. At a specific time (t=0.07s), the
load power is abruptly increased.
Load resistances are modified to R1 = R2 = 17.5 at
that moment. The respective change in the battery current is voltages in response to load variations. In reality, the load
represented in FIGURE 20. As can be seen, controllers did an power can be fed among the input sources by managing the
excellent job of regulating the battery current and the output battery current. In the charging mode of the battery, source 1

3380 VOLUME 11, 2023

S. R. Khasim et al.: Single Inductor Multi-Port Power Converter for Electric Vehicle Applications

output port voltages V1 and VT. The reference values track

the output voltages, as shown in this diagram. FIGURE 22
depicts the battery current after it has been adjusted to the
desired value. The battery current is negative, indicating that
the battery is being charged. The battery’s reference current
is rapidly reduced from −0.9 A to −2 A at a specific time
(t = 6.5 s), as shown in FIGURE 23. Furthermore, at a pre-
determined time (t = 13s), the output load power is rapidly
FIGURE 19. Inductor current in discharging mode.
increased. In this case, the respective output voltages and
battery current. The converter is well controlled by developed
compensators, as can be seen in this diagram. It’s important
to note that every converter has a particular operating range.

B. EXPERIMENTAL RESULTS
The converter developed is designed in MATLAB/Simulink
and tested in a laboratory environment with a hardware setup
represented in FIGURE 31. Two separate power sources
FIGURE 20. Change of battery reference current. were used as inputs. In the setup, input sources are a dc
power supply and a 48V battery consisting of four series
12V, 7.5Ah lead-acid batteries. The experimental specifica-
tions are shown in TABLE 2. The power supply is provided
with a voltage of 35V the represents a solar PV voltage.
The output voltages reference values are V1ref = 80V and
V2ref = 40 V respectively, and there use two resistive loads
utilizing a dSPACE RTI1104 controller. The analysis is done
in two separate converter operation modes. The following
are detailed descriptions of the testing results for battery dis-
FIGURE 21. Output voltages in battery charging mode. charge as well as charging modes. At times before transition
time, the converter is in battery discharging mode and after
the transition time, the converter is in battery charging mode.
It is obvious that against the operation mode changing and
change of the loads the output voltages have been regulated
to the desired values. The transition voltages from charging
to discharging are represented in FIGURE 27.

1) BATTERY DISCHARGING MODE

The battery’s reference current and output port voltages are
FIGURE 22. Battery voltages at desired value. defined in this mode as Ibref =3A, V1ref = 80V, and V2ref =
40V, with R1 =R2 =35 (less than 35). The output port
voltages of the converter are displayed in FIGURE 24 under
these conditions. The output port voltage V1 and the total
overall output port voltage VT = V1 + V2 are displayed in this
diagram. FIGURE 25 depicts the current fed from the battery
and the inductor current. The current drawn by the battery
is 3A. The correct distribution of power among the input
energy sources can be achieved by managing the battery
FIGURE 23. Change of battery reference current. current. Ppv = 150W, Pbat = 143W, P1 = 189W, and P2 =
47W in this mode.

supplies power to source 2 in addition to the loads (battery). 2) BATTERY CHARGING MODE
Switches S1 , S2 and S4 are active in this mode. In this mode, The load power is modest in this mode. As a result, source 1
the required output voltages are V1ref = 80V, V2ref = 40V can both deliver loads by charging the battery. The output
and the current reference of battery charging is Ibref = 0.9A, port voltages and the battery charging current is defined as
similar to the battery discharge mode. Ibref = 0.9A, V1ref = 80V, and V2ref = 40V, respectively,
In this case, the compensators created during the charging as well as R1 =R2 =70 (less than 70). The respective
of the battery are used. FIGURE 21 depicts the respective output port voltages of the developed converter are displayed

VOLUME 11, 2023 3381

 S. R. Khasim et al.: Single Inductor Multi-Port Power Converter for Electric Vehicle Applications

FIGURE 28. Battery charging current.

FIGURE 24. Output voltages in battery discharging mode.

FIGURE 29. The battery current is discharging and the charging mode.

FIGURE 25. Drawn current from battery and inductor current in

discharging mode.

FIGURE 30. Variation of inductor current from battery discharging to

charging.

FIGURE 26. Solar PV voltage and current.

FIGURE 27. Output voltages of a converter.

FIGURE 31. Experimental setup.
in FIGURE 27 under this circumstance. FIGURE 28 depicts
the battery charging current. The current of source 1 equals
the current of the inductor. Ppv = 214W, Pbat = 51W, P1 = from battery discharge to the battery charging mode is exam-
96W, and P2 = 25W in this mode. The suggested converter’s ined. The developed converter was initially made to run in
operation battery discharge mode with the increase of load resistances

3382 VOLUME 11, 2023

S. R. Khasim et al.: Single Inductor Multi-Port Power Converter for Electric Vehicle Applications

TABLE 3. Component comparison with existing topologies.

FIGURE 32. Comparison of various existing topologies with the proposed DIDO converter (a) NS, (b) ND, (c) NL (d) NC (e) CC/Port.

over time. Load resistances in discharge mode are R1 = battery has been modified from 3A to − 0.9A in battery dis-
R2 = 35, while load resistances in charging mode are charging mode. Hence, the current through the battery during
R1 = R2 = 70. In addition, the reference current of the charging is displayed in FIGURE 28 under such conditions.

VOLUME 11, 2023 3383

 S. R. Khasim et al.: Single Inductor Multi-Port Power Converter for Electric Vehicle Applications

The converter is in battery draining mode before the transition [2] P.-J. Liu, Y.-K. Lo, H.-J. Chiu, and Y.-J. Emery Chen, ‘‘Dual-current pump
time and battery charging mode after the transition time. module for transient improvement of step-down DC–DC converters,’’
IEEE Trans. Power Electron., vol. 24, no. 4, pp. 985–990, Apr. 2009, doi:
The respective output port voltages have been controlled to 10.1109/TPEL.2008.2010322.
the desired values in the face of shifting operation modes [3] J. Macaulay and Z. Zhou, ‘‘A fuzzy logical-based variable step size
and changing loads. By shifting the mode of operation from P&O MPPT algorithm for photovoltaic system,’’ Energies, vol. 11, no. 6,
p. 1340, May 2018, doi: 10.3390/en11061340.
charging to discharging, identical results can be produced. [4] M. Dhananjaya, D. Ponuru, T. S. Babu, B. Aljafari, and H. H. Alhelou,
The battery current is displayed in FIGURE 29 as the mode ‘‘A new multi-output DC–DC converter for electric vehicle application,’’
changes from charging to discharging. The solar PV input IEEE Access, vol. 10, pp. 19072–19082, 2022.
[5] L. Wang, E. G. Collins, and H. Li, ‘‘Optimal design and real-
voltage and respective current are represented in FIGURE 26. time control for energy management in electric vehicles,’’ IEEE
The variation of inductor current from battery discharging to Trans. Veh. Technol., vol. 60, no. 4, pp. 1419–1429, May 2011, doi:
charging is shown in FIGURE 30. 10.1109/TVT.2011.2122272.
The proposed topology with passive elements and switches [6] M. Zandi, A. Payman, J. P. Martin, S. Pierfederici, B. Davat, and
F. Meibody-Tabar, ‘‘Energy management of a fuel cell/supercapacitor/
is compared to other similar converter topologies in Table 3. battery power source for electric vehicular applications,’’ IEEE
The results show that the suggested converters are compa- Trans. Veh. Technol., vol. 60, no. 2, pp. 433–443, Nov. 2011, doi:
rable to that other converters. Furthermore, the suggested 10.1109/TVT.2010.2091433.
[7] K. Suresh, C. Bharatiraja, N. Chellammal, M. Tariq, R. K. Chakrabortty,
converter provides advantages such as a modular structure M. J. Ryan, and B. Alamri, ‘‘A multifunctional non-isolated dual input-dual
with fewer components and the integration of a variety of output converter for electric vehicle applications,’’ IEEE Access, vol. 9,
input sources with varying voltage-current characteristics. pp. 64445–64460, 2021.
[8] S. R. Khasim and C. Dhanamjayulu, ‘‘Design and implementation of
This converter can be used in electric vehicle power circuits. asymmetrical multilevel inverter with reduced components and low
The proposed converter, in particular, can be used in a regen- voltage stress,’’ IEEE Access, vol. 10, pp. 3495–3511, 2022, doi:
erative mode. 10.1109/ACCESS.2022.3140354.
[9] A. Ajami, H. Ardi, and A. Farakhor, ‘‘A novel high step-up DC/DC
converter based on integrating coupled inductor and switched-
VII. CONCLUSION capacitor techniques for renewable energy applications,’’ IEEE Trans.
Power Electron., vol. 30, no. 8, pp. 4255–4263, Aug. 2015, doi:
This research proposes a multi-port converter with a com- 10.1109/TPEL.2014.2360495.
bined structure for hybridizing power sources in the appli- [10] M. R. Banaei, H. Ardi, R. Alizadeh, and A. Farakhor, ‘‘Non-isolated multi-
cations like electric vehicles. There is only one inductor in input-single-output DC/DC converter for photovoltaic power generation
systems,’’ IET Power Electron., vol. 7, no. 11, pp. 2806–2816, Nov. 2014,
the proposed converter. The suggested converter can able to
doi: 10.1049/iet-pel.2013.0977.
transfer energy among the various energies like solar PV [11] O. Hegazy, R. Barrero, J. Van Mierlo, P. Lataire, N. Omar, and
and ESSs such as a battery. The power supply and ESS are T. Coosemans, ‘‘An advanced power electronics interface for electric
referred to as solar PV source and battery respectively. The vehicles applications,’’ IEEE Trans. Power Electron., vol. 28, no. 12,
pp. 5508–5521, Dec. 2013, doi: 10.1109/TPEL.2013.2256469.
developed converter can also be used as a single-input, multi- [12] C. Dhanamjayulu, S. Padmanaban, V. K. Ramachandaramurthy,
output device. Multiple outputs with varying voltage levels J. B. Holm-Nielsen, and F. Blaabjerg, ‘‘Design and implementation
are possible. In the battery discharge mode, input sources sup- of multilevel inverters for electric vehicles,’’ IEEE Access, vol. 9,
pp. 317–338, 2021.
ply power to the output, whereas in the case of charging mode, [13] Y. Cao and J. A. Abu Qahouq, ‘‘Evaluation of bi-directional single-inductor
out of the two inputs, one input source not only delivers power multi-input battery system with state-of-charge balancing control,’’ IET
to loads but also provides to the battery. Transfer function Power Electron., vol. 11, no. 13, pp. 2140–2150, Nov. 2018.
[14] S. R. Khasim, D. C, S. Padmanaban, J. B. Holm-Nielsen, and M. Mitolo,
matrices for each mode are computed individually, and the ‘‘A novel asymmetrical 21-level inverter for solar PV energy system with
respective compensators utilized for closed-loop converter reduced switch count,’’ IEEE Access, vol. 9, pp. 11761–11775, 2021, doi:
control are built. The battery current and the output voltages 10.1109/ACCESS.2021.3051039.
are regulated to desired values under various scenarios such [15] R. Ahmadi and M. Ferdowsi, ‘‘Double-input converters based on H-bridge
cells: Derivation, small-signal modeling, and power sharing analysis,’’
as a quick rise in load power and a sudden shift in the battery IEEE Trans. Circuits Syst. I, Reg. Papers, vol. 59, no. 4, pp. 875–888,
reference current. Multilevel inverters can be connected to Apr. 2012, doi: 10.1109/TCSI.2011.2169910.
outputs with various voltage levels. Induction motor torque [16] L. Solero, A. Lidozzi, and J. A. Pomilio, ‘‘Design of multiple-input power
converter for hybrid vehicles,’’ IEEE Trans. Power Electron., vol. 20, no. 5,
ripple is reduced when multilayer inverters are used in electric pp. 1007–1016, Sep. 2005, doi: 10.1109/TPEL.2005.854020.
automobiles. However, electric vehicles have at least two [17] B. Faridpak, M. Farrokhifar, M. Nasiri, A. Alahyari, and N. Sadoogi,
separate dc voltages, one for cabin lighting and the other for ‘‘Developing a super-lift luo-converter with integration of buck converters
for electric vehicle applications,’’ CSEE J. Power Energy Syst., vol. 7, no. 4,
the motor. Multilevel inverters are also beneficial in grid con- pp. 811–820, 2020.
nectivity of renewable energy resources like solar PV. Finally, [18] S. R. Khasim and D. C, ‘‘Design of multiple input DC–DC converters for
the operation of this converter is tested in the laboratory using interfacing solar/fuel cell/battery electric vehicles,’’ in Proc. Innov. Power
Adv. Comput. Technol. (i-PACT), Nov. 2021, pp. 1–7.
a hardware setup.
[19] C. Dhanamjayulu, S. R. Khasim, S. Padmanaban, G. Arunkumar,
J. B. Holm-Nielsen, and F. Blaabjerg, ‘‘Design and implementation of
multilevel inverters for fuel cell energy conversion system,’’ IEEE Access,
REFERENCES
vol. 8, pp. 183690–183707, 2020, doi: 10.1109/access.2020.3029153.
[1] S. R. Khasim and C. Dhanamjayulu, ‘‘Selection parameters and syn- [20] K. Suresh, N. Chellammal, C. Bharatiraja, P. Sanjeevikumar, F. Blaabjerg,
thesis of multi-input converters for electric vehicles: An overview,’’ and J. B. H. Nielsen, ‘‘Cost-efficient nonisolated three-port DC–DC con-
Renew. Sustain. Energy Rev., vol. 141, May 2021, Art. no. 110804, doi: verter for EV/HEV applications with energy storage,’’ Int. Trans. Electr.
10.1016/j.rser.2021.110804. Energy Syst., vol. 29, no. 10, p. e12088, Oct. 2019.

3384 VOLUME 11, 2023

S. R. Khasim et al.: Single Inductor Multi-Port Power Converter for Electric Vehicle Applications

[21] C. Dhanamjayulu, D. Prasad, S. Padmanaban, P. K. Maroti, C. DHANAMJAYULU (Senior Member, IEEE)

J. B. Holm-Nielsen, and F. Blaabjerg, ‘‘Design and implementation received the B.Tech. degree in electronics and
of seventeen level inverter with reduced components,’’ IEEE Access, communication engineering from JNTU Univer-
vol. 9, pp. 16746–16760, 2021, doi: 10.1109/ACCESS.2021.3054001. sity, Hyderabad, India, the M.Tech. degree in con-
[22] Y.-C. Liu and Y.-M. Chen, ‘‘A systematic approach to synthesizing multi- trol and instrumentation systems from the IIT
input DC–DC converters,’’ IEEE Trans. Power Electron., vol. 24, no. 1, Madras, Chennai, India, and the Ph.D. degree
pp. 116–127, Jan. 2009, doi: 10.1109/TPEL.2008.2009170. in power electronics from the Vellore Institute
[23] H. Wu, K. Sun, S. Ding, and Y. Xing, ‘‘Topology derivation of non-
of Technology, Vellore, India. He was invited
isolated three-port DC–DC converters from DIC and DOC,’’ IEEE
as a Visiting Researcher with the Department of
Trans. Power Electron., vol. 28, no. 7, pp. 3297–3307, Jul. 2013, doi:
10.1109/TPEL.2012.2221746. Energy Technology, Aalborg University, Esbjerg,
[24] B. N. Alajmi, M. I. Marei, I. Abdelsalam, and N. A. Ahmed, ‘‘Multi- Denmark, funded by the Danida Mobility Grant, Ministry of Foreign Affairs
phase interleaved converter based on cascaded non-inverting buck-boost of Denmark on Denmark’s International Development Cooperation. He was
converter,’’ IEEE Access, vol. 10, pp. 42497–42506, 2022. a Postdoctoral Fellow at the Department of Energy Technology, Aalborg Uni-
[25] B. N. Alajmi, M. I. Marei, I. Abdelsalam, and M. F. AlHajri, ‘‘Analysis versity, from October 2019 to January 2021. Since 2010, he has been a Senior
and design of a multi-port DC–DC converter for interfacing PV systems,’’ Assistant Professor with the Vellore Institute of Technology. His research
Energies, vol. 14, no. 7, p. 1943, Apr. 2021. interests include multilevel inverters, power converters, active power filters,
[26] K. Shreya, H. Srivastava, P. Kumar, G. Ramanathan, P. Madhavan, and power quality, grid-connected systems, smart grids, electric vehicle, electric
C. Bharatiraja, ‘‘CUK converter fed resonant LLC converter based electric spring, and tuning of memory elements and controller parameters using
bike fast charger for efficient CC/CV charging solution,’’ J. Appl. Sci. Eng., softswitching techniques for power converters, average modeling, steady-
vol. 24, no. 3, pp. 331–338, 2021. state modeling, and small-signal modeling stability analysis of the converters
[27] B. Axelrod, Y. Berkovich, and A. Ioinovici, ‘‘Continuous-time digital con- and inverters. He is an Academic Editor of the Journal International Transac-
troller,’’ in Proc. Int. Symp. Circuits Syst. (ISCAS), vol. 3, no. 2, Jun. 2003, tions on Electrical Energy Systems (Wiley-Hindawi). He is also an Academic
pp. 564–573.
Editor of the Journal Mathematical Problems in Engineering (Hindawi).
[28] S. Athikkal, B. Chokkalingam, S. I. Ganesan, B. Lehman, and
T. B. Lazzarin, ‘‘Performance evaluation of a dual-input hybrid step-up
DC–DC converter,’’ IEEE Trans. Ind. Appl., vol. 58, no. 3, pp. 3769–3782,
Jun. 2022.
[29] F. Kardan, R. Alizadeh, and M. R. Banaei, ‘‘A new three input DC/DC
converter for hybrid PV/FC/Battery applications,’’ IEEE J. Emerg. Sel.
Topics Power Electron., vol. 5, no. 4, pp. 1771–1778, Dec. 2017.
[30] M. AL-Emam, M. I. Marei, and W. El-khattam, ‘‘A maximum power point
tracking technique for PV under partial shading condition,’’ in Proc. 8th
IEEE India Int. Conf. Power Electron. (IICPE), Dec. 2018, pp. 1–6.
[31] P. Mohseni, S. H. Hosseini, M. Sabahi, T. Jalilzadeh, and M. Maalandish, S. M. MUYEEN (Senior Member, IEEE) received
‘‘A new high step-up multi-input multi-output DC–DC converter,’’ IEEE the B.Sc.Eng. degree from the Rajshahi Uni-
Trans. Ind. Electron., vol. 66, no. 7, pp. 5197–5208, Sep. 2018. versity of Engineering and Technology (RUET),
[32] N. Asim, K. Sopian, S. Ahmadi, K. Saeedfar, M. A. Alghoul, O. Saadatian, Bangladesh, formerly known as the Rajshahi Insti-
and S. H. Zaidi, ‘‘A review on the role of materials science in solar cells,’’ tute of Technology, in 2000, and the M.Eng. and
Renew. Sustain. Energy Rev., vol. 16, no. 8, pp. 5834–5847, 2012, doi: Ph.D. degrees from the Kitami Institute of Tech-
10.1016/j.rser.2012.06.004. nology, Japan, in 2005 and 2008, respectively,
[33] A. Amir, A. Amir, J. Selvaraj, and N. A. Rahim, ‘‘Study of the all in electrical and electronic engineering. He is
MPP tracking algorithms: Focusing the numerical method techniques,’’
a Full Professor with the Electrical Engineering
Renew. Sustain. Energy Rev., vol. 62, pp. 350–371, Sep. 2016, doi:
Department, Qatar University. He is a keynote
10.1016/j.rser.2016.04.039.
speaker and an invited speaker at many international conferences, work-
shops, and universities. He has published more than 350 articles in different
SHAIK REDDI KHASIM (Member, IEEE) journals and international conferences. He has published seven books as
received the B.Tech. degree (Hons.) in electrical an author or editor. His research interests include power system stability
engineering and the M.Tech. degree (Hons.) from and control, electrical machine, FACTS, energy storage systems (ESS),
JNTUA, Andhra Pradesh, India, in 2012 and 2015, renewable energy, and HVDC systems. He is a Senior Member of Chartered
respectively. He is currently pursuing the Ph.D. Professional Engineers, Australia, and a fellow of Engineers Australia. He is
degree in power electronics with the Vellore Insti- serving as the Editor/an Associate Editor for many prestigious journals from
tute of Technology, Vellore, India. His research IEEE, IET, and other publishers, including IEEE TRANSACTIONS ON ENERGY
interests include multilevel inverters, power con- CONVERSION, IEEE POWER ENGINEERING LETTERS, IET Renewable Power Gen-
verters, and electric vehicles. eration, and IET Generation, Transmission and Distribution.

VOLUME 11, 2023 3385