Electrical Engineering

https://doi.org/10.1007/s00202-023-01822-6

ORIGINAL PAPER

Design of a modular converter in hybrid EV charging station

with efficient energy management system
Mohamed Iqubal1 · Paul sathiyan1 · Albert Alexander Stonier2 · Dishore Shunmugham Vanaja3 · Geno Peter4

Received: 16 January 2023 / Accepted: 3 April 2023

© The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature 2023

Abstract
This paper presents the design and development of a modular multiport DC-DC converter for hybrid charging station. The
system is supplied by renewable energy sources (RES) like solar photovoltaic system (SPV), wind energy system (WE) and
fuel cell (FC). The proposed converter has several benefits like current sharing capability, providing lower current ripple and
lower voltage ripple, lower current stress on the switches, lower current rating of switches and using a small-sized inductor.
Due to the unavailability of continuous flow of power from the RES throughout the day, an effective energy management
system using the adaptive neuro-fuzzy inference system (ANFIS) is implemented in the proposed system. ANFIS controller
is implemented to identify and enable the available power from the renewable energy sources to feed the base load as well as
charge the batteries in the hybrid charging station. The ANFIS controller continuously monitors the power availability for the
charging station, thereby feeding minimum power to the base load. The proposed system is simulated and tested using the
MATLAB/SIMULINK software, and a prototype of the same has been developed in the laboratory. From the simulation and
experimental results, it is observed that the proposed converter suits best for integrating hybrid sources and used for charging
the EVs. Improved life expectancy of the battery is obtained using the proposed converter. The proposed converter produces
an efficiency of 97% and is superior to the other existing converters.

Keywords Fuel cell · Fuzzy system · Multiport converter · Solar power · Wind power

1 Introduction emission-free transportation system has been introduced to

reduce the carbon emissions from the conventional automo-
Electric energy systems with renewable energy resources biles. Emission-free transportation system includes electric
have led to the development of economy. The integration vehicles (EVs). The greatest challenge faced by EVs is
of renewable resources with the conventional grid reduces their refueling. Therefore, the need for charging station (CS)
the carbon emission [1]. The automotive transportation sys- increases, which in turn increases the demand on the electri-
tem accounts for about 14% of the total C O 2 emissions. cal grid [3]. To reduce the burden on the grid and to provide
These emissions cause several health issues to mankind as continuous power supply, the charging stations are integrated
well cause negative impact to the environment [2]. Therefore, with the renewable energy systems [4]. Due to the inter-
mittent nature of the renewable energy systems (wind and
B Geno Peter solar), hybrid energy systems have been introduced in charg-
[email protected]
ing stations. Hybrid energy sources (wind, solar and fuel
1 Department of Electrical and Electronics Engineering, cell) comprise two or more renewable energy sources. Suit-
Karunya Institute of Technology and Sciences, Coimbatore, able power electronic circuits paved the way for integration
India of renewable sources with the energy generation systems.
2 School of Electrical Engineering, Vellore Institute of Since majority of the renewable sources are in the form of
Technology, Vellore, Tamil Nadu, India DC, DC-DC converters [5] play a very indispensable part
3 Department of Electrical and Electronics Engineering, Mar in the conversion and controlling processes of renewable
Baselios College of Engineering and Technology, energy systems. Moreover, an electric vehicle’s energy stor-
Thiruvananthapuram, Kerala, India
age system requires a regulated DC power supply. Because of
4 CRISD, School of Engineering and Technology, University of
Technology Sarawak, Sibu, Malaysia

123
 Electrical Engineering

Table 1 Comparison of
conventional multiple port and Parameters Converter
multiport converters conventional multiple Multiport

Common DC−bus Necessary Not necessary

Conversion stages Many stages Few stages
Energy management Complex and slow Simple and fast
Isolation Multiple winding transformer If needed (single or multiple winding transformer)
Control Individual control Centralized control
Device count Need more components Need fewer components

this, the battery charging stations also require a DC-DC con- can be introduced in the market without redesigning them for
verter that is both compact and capable of managing a large different capacities. In the proposed work, the non-isolated
amount of power. The topologies of DC-DC converters can type multiport DC-DC converter is used since it has a reduced
be divided into two categories: isolated [6] and non-isolated switch count. The authors in [11] discussed different topolo-
[7]. Non-isolated topologies were developed with the goal of gies of multiport converters. A buck–boost converter-based
lowering the required quantity of operating switches, thereby MPC is used for a PV-integrated battery energy storage sys-
increasing the efficiency and performance of the system. Iso- tem [12]. Multiport boost converters are integrated for hybrid
lated topologies, on the other hand have certain limitations energy systems in [13] and [14]. Three-port PV-integrated
like leakage inductance, transformer core saturation, thermal battery-connected boost MPC is constructed in [15]. A 1kW
impact, high voltage stress across the switches, a massive MPC topology with energy storage system is developed with
size and high cost are all problems that remain in isolated varied electrical characteristics in [14]. A bidirectional buck-
converters compared to non-isolated converters. The typi- –boost multiport converter with many inputs is proposed in
cally used DC-DC converters are buck [8] (reduces the input [16] for power harvesting. A 4-port boost converter for hybrid
source voltage), boost [9] (enhances the input source volt- system using half bridge topology is proposed in [17], which
age) and buck boost [10] (enhances and reduces the input feeds 3 inputs and one output sources. The authors in [18]
source voltage) converters. In hybrid charging stations, more proposed a three-port boost converter with solar, battery and
number of renewable sources are fed to the DC bus at the an isolated output port. A decoupling network has been intro-
same time. Therefore, multiple converters are required for duced for complete control of the multiport converter.
converting and controlling the power. So, in this case, each One of the main challenges in HCS is to manage and con-
converter requires a separate control and hence leads to com- trol energy flow. The dynamic exchange of energy between
plex control circuitry. Due to these drawbacks, the multiport sources, load, and storage devices in a HCS entails the prob-
DC-DC converters (MPC) are developed and they are also lems of power quality and stability. In order to meet load
categorized as isolated and non-isolated topologies. Table 1 demand, continuous management of energy flow throughout
presents the comparison of conventional multiple converters the system is essential. The aim of this work is to develop
with the multiport converters. From Table 1, it is identified new energy management control algorithm based on artificial
that MPC are simple to use with less control methods and intelligence (AI) methods which can handle unpredictable
reduced conversion stages. Hence, MPCs are mostly used in situations arising in HCS. Better utilization of renewable
hybrid charging stations (HCS). The power rating of MPC energy sources can be achieved through effective energy
should be decided wisely by considering all the aspects, management of the sources and load [19]. The parameters
including the future increase in power demand for charging like energy cost and choice of source combined with maxi-
the EVs. mum utilization of renewable energy can bring down the cost
However, accurate forecasting the future EV load demand of energy [20]. The impact of PV systems on storage ele-
is an error-prone task. Further, operating an overrated MPC is ments and electric vehicles through proper control strategies
neither economically viable nor efficient. Therefore, a mod- in a microgrid is found to yield better results [21]. Auxiliary
ular structure for MPC, where the number of modules can distribution transformers connect electric vehicle charging
be increased as and when the power demand increases, is stations to local loads. In the event of an emergency, the
an optimum solution. A modular MPC has other advantages, charging station’s EMS can be activated to cut off power to
which are as follows: a faulty module can be bypassed in the the vehicle. By keeping an eye on and making adjustments to
event of its failure, and the system can operate at a reduced how the electricity is being used, the EMS can help to keep the
power level, thus increasing the reliability of the system. Fur- system safer. To ensure safe and efficient operation, BMS can
ther, a wide range of MPCs having different power ratings forecast the health of a system [22]. The primary goal of the

123
Electrical Engineering

EMS is to prevent transformer overloading, maximize renew-

able energy use and provide uninterrupted charging. From
the literature, the issues identified in developing a hybrid fast
charging stations are as follows:

• The output power produced by the hybrid energy sources is

fluctuating; therefore, suitable power electronic converters
have to be developed to regulate the power from HES
• Most of the MPC use many device components, which
make the system bulky and very difficult to control
• Poor energy flow management between sources, storage
devices and load

Based on the problems identified, the objectives are

framed.

• To design a modular MPC which delivers an output voltage

with low-output voltage ripples, lesser control complexity
and higher efficiency for the hybrid charging station
• To implement a battery energy storage setup for compen- Fig. 1 Structure of the proposed hybrid charging station
sating the slow response of the renewable sources
• To implement an effective energy management system
which can increase the efficiency of operation of the hybrid
charging station prioritizing the loads, optimizing the use
of storage element and maximizing the use of renewable
energy sources (RES)
• To compare the designed modified topology with the con-
ventional multiport converter and to check its competency
with its counter parts.

This paper is organized as follows: Section II discusses Fig. 2 Basic PV cell

about the proposed system design; Section III presents the
design of PV port; Section IV presents the comparison of the
proposed converter with other converters; Section V presents compensate the load demand when there is no availability of
the energy management of the proposed system; Section VI power from renewable energy sources or batteries.
presents the simulation and experimental results and Section
VII gives the conclusion of the proposed work.
A. PV system modeling

2 Proposed system To analyze a solar photovoltaic cell, its modeling is an essen-

tial part which provides the complete behavior of PV cell. A
The proposed hybrid system along with the intelligent energy photovoltaic system (PV) [23] is made up of solar or PV
management system is presented in Fig. 1. The proposed con- cells. Figure 2 presents the circuit of a basic PV cell. When
verter mainly uses non-conventional energy sources like solar any photon or illumination falls on solar cell, the formation
PV systems, wind generators and fuel cell as the input. The of electron–hole pair takes place. An electric field created
role of energy management system in the proposed work is by photonic energy drives the charges, and hence the flow
to calculate the available loads (EVs) at the particular instant of electrons and holes in opposite area results in a current to
and allocate the power from the renewable sources as per the be flow in the circuit. This current is called a photocurrent,
requirement. Since the renewable energy sources are inter- and hence a DC power can be delivered to external electrical
mittent in nature, the power harvested from these sources circuit. This generated power amount depends upon various
are stored in the battery bank of the charging station. More- factors like solar irradiation intensity and the material of PV
over, the proposed system is integrated with the grid so as to cell to be used. As a PV cell produces a very less amount

123
 Electrical Engineering

of voltage in the range of 0.5 V, various PV cells get con-

nected in series for making the system up to desired level
of voltage. Similarly, for improving the current rating of the
system, several PV cells get connected in parallel connec-
tion. In short, it can be concluded that for proper application
of solar cell in a particular application, a proper series and
parallel combination of cells have to be taken care of as per
the requirement.
By applying Kirchhoff’s law to the above circuit

I  I L − I D − Ish (1)

V + I Rs
Ish  (2)
Rsh

Now Eq. (1) can be altered as:

 qV d
  V + I Rs

I  I L − Io e AkT −1 − V + (3)
Rsh Fig. 3 Structure of PEMFC

Finally, according to the theory of semiconductors, the

nonlinear I-V relationship for the solar PV cell can be Hence, available power in the wind can be written as
expressed as given in Eq. (4)
1
Pwind  ρ Av 3 (7)
 q(V +I Rs )
  V + I Rs
 2
I  I L − Io e AkT −1 − V + (4)
Rsh Since it is not possible to keep the airflow constant in the
wind, all the power in the wind cannot be extracted. In theory,
where I ,I L and Io are cell’s output current, photo-generated
the maximum power that can be harvested from wind is only
current and diode reverse saturation current in amperes,
59.3% and it is known as Betz limit. Hence, available power
respectively.
can be expressed as

B. Wind turbine design

Pout ∞ 0.59 ∗ ρw ∗ As ∗ v 3

Energy from wind [24] is harvested using the turbine-

generator assembly. Wind turbine consists of rotor blades, Pout  Ci ∗ 0.59 ∗ ρw ∗ As ∗ v 3 (8)
nacelle, hub and the tower. The rotor blades are the most
important component of a wind turbine and are designed to where Ci is the power coefficient,v is the wind speed,ρ
harvest maximum power as the wind sweeps across them defines air density and As defines the swept area of the blade.
without damaging the whole unit. The mechanical output  
depends on the intensity as well as the direction of the wind Ci  0.73 151λi − 0.58β − 0.002β 2.14 − 13.2 e−18.4λi
with respect to the turbine. This mechanical power will be
(9)
transformed into electrical power by means of generator. The
generator output is highly dependent on the speed of the wind.
Kinetic energy available in the wind sweeping the rotor where λi  11λ + 0.02β − 0.03β 3 + 1
blade area can be written as
λ − tipspeed ratio; β − Pitchangle
1
K .E wind  mv 2 (5)
2
C. Fuel cell model
where, v is the wind speed,m is the air mass that hits on the
rotor blades and is given by Fuel cell is a renewable source of energy, in which elec-
trochemical reaction converts fuel into electricity. Figure 3
m  ρ Av (6) presents the diagram of proton exchange membrane fuel cell

123
Electrical Engineering

(PEMFC) [25]. In the fuel cell, the following reactions occur

at the anode and the cathode.

Reaction at Anode 2H2 → 4H + + 4e− (10)

AtCathode O2 + 4H+ + 4e− → 2H2 O + Heat (11)

1
Overall Reaction H2 + O2 → H2 O + Heat (12)
2
Equations (13) and (14) are used to model the PEMFC:
   
Ik 1
Vfc  E oc − N Aln × I − RI f c (13)
Ic s 3d + 1

where E oc  K c E n
    −G 
z Fk PH2 +PO2
Ic  Rh e RT
(14)
A RT
zα F

where, V f c is the cell voltage of fuel cell (V ), E oc represents Fig. 4 Proposed multiport converter
the open circuit voltage (V ), Ic defines the exchange current,
A defines the Tafel slope, K c signifies the voltage constant
converter are operated with 50% lesser duty cycle for elim-
at nominal operating condition.
ination of input and output current ripples. Table 2 presents
the operating modes of the basic module (PV port), and Fig. 5
D. Multiport converter
presents the switching waveforms of the PV port. Figure 6
presents the operating modes of the PV port.
A multiport converter is mostly employed for a hybrid charg-
Figure 6 a presents the first mode of operation, and its
ing station. The surplus power generated by the hybrid
corresponding waveform is shown in Fig. 5, in which the
sources can be delivered to the grid using a multiport con-
switch S p is triggered. On conduction, the inductor current
verter. Multiport converters allow electric vehicles to be
i L i p starts increasing and stores energy from the PV system.
charged directly from a DC supply. In order to develop a well-
Meanwhile, the current in the output inductor i L op dis-
coordinated MPC, which is capable of integrating hybrid
charges via the diode D p . This makes the capacitor voltage
sources with DC microgrid, detailed technical requirements
VC  to increase. The output inductor L op feeds the required
of each port of the converter should be specified. The main p
current to the DC bus.
specifications are the operating voltage for individual ports
and their power handling capability.
VL i p (t)  Vpv  VD p ; VL i p (t)  Vpv − VC p  VS p
Figure 4 depicts the design of a modular multi-input
Landsman converter. The suggested converter topology Vpv
L i p (t)  + i L i p (t0 )  i S p (t); i L i p (t)
works in the discontinuous conduction mode, which elim- Lip
inates the need for reverse recovery losses in fast diodes Vpv − VC p
and provides inherent zero-current switching. The suggested  (t − t0 ) + i L i p (t0 )  i D p (t)
L i p
converter consists of six switching devices, speedy recov-
VL op (t)  VC p − Vbus ; VL op (t)  −Vbus ; VS p  VD p  0
ery diodes and filters. For simplicity and understanding the
modes of operation, the basic or individual port is consid- (15)
ered (PV port). The PV port consists of two switches S p
and S p ,two diodes D p and D p ,two capacitors C p and C p , Ripple current of the inductor is presented as
two input inductors L i p and L i p and two output inductors
Iri pple (t)  I L i p (t) − I L i p (t)
L op and L op . Phase-shifted pulse width modulation method 
(PSPWM) [26] is used for triggering the switches of the t1 Vpv t1 Vpv − V p t1 V p 
 −  (16)
proposed converter. Moreover, the switches of the proposed Lip L i p Lk

123
 Electrical Engineering

Table 2 Operating modes of the converter

Mode T ime period Sp S p Dp D p Cp C p Lip L i p

I 0 ≤ t ≤ t1 ✓ × × ✓ ↓ ↑ ↑ ↓
II t1 ≤ t ≤ t2 ✓ × × × ↓ ↓ ↑ ∀
III t2 ≤ t ≤ t3 × × ✓ × ↑ ↓ ↓ ∀
IV t3 ≤ t ≤ t4 × ✓ ✓ × ↑ ↓ ↓ ↑
V t4 ≤ t ≤ t5 × ✓ × × ↓ ↓ ∀ ↑
VI t5 ≤ t ≤ t6 × × × ✓ ↓ ↑ ∀ ↓

−O N , −O F F, ↑ −Charging, ↓ −Discharging, ∀ − Dischargedcompletely

completely. The voltages across the capacitors C p and C p

discharge its respective voltages through the output inductors
L op and L op .

VL i p (t)  Vpv  VD p ; VL i p (t)  0; VS p  VD p  0

Vpv
i L i p (t)  + i L i p (t0 )  i s p (t); i L i p (t)  0
Lip
VL op (t)  Vpv − VC p − Vbus ; VL op (t)  Vpv − VC p − Vbus
(17)

Figure 6c presents the third mode of operation, and its

corresponding waveform is shown in Fig. 5, in which the
switches S p and S p are in off condition and the diode D p
conducts. The diode D p makes the inductor L i p to discharge
Fig. 5 Waveforms of the PV port
its stored energy, meanwhile the voltage across the capacitor
C p increases as shown in Fig. 5. Moreover, from Fig. 5, it is
where observed that the input inductor L i p is still in DCM, whereas
the voltage across the capacitor C p discharges and feeds the
output current via the output inductance L op .
L k  L i p  L i p

From the above equation, it has been observed that the

input current ripple depends on the voltage across the capac- VL i p (t)  Vpv − VC p  VS p ; VL i p (t)  0
itor C p and the inductance and the turn on duration of the Vpv − VC p
switch (t1 ). Therefore, with the help of PSPWM [26], a i L i p (t)  + i L i p (t0 )  i D p (t); i L i p (t)  0
Lip
reduced duty cycle is provided to the switch S p so that the VL op (t)  −Vbus ; VL op (t)  Vpv − VC p − Vbus ; VS p  VD p  0
switching duration is lessened, which thereby produces a (18)
lower ripple current and lower current stress in the converter
system.
Figure 6b presents the second mode of operation, and
its corresponding waveform is shown in Fig. 5, in which Figure 6d presents the fourth mode of operation, and its
the diode D p stops conduction because of the presence of corresponding waveform is shown in Fig. 5, in which the
reverse polarity voltage across the capacitor C p . Moreover, switch S p is turned on. On conduction, the inductor current
the switch S p is still in conduction state and the input induc- i L  starts increasing and stores energy from the PV system.
ip
tor (L i p ) still continues to store charge from the PV system. The capacitor C p starts to discharge its voltage and supplies
The input inductor (L i p ) enters the discontinuous conduc- the current to the DC bus via the inductor L op Meanwhile,
tion mode during this time duration and discharges its voltage the energy stored in the inductor L i p is still in discharging

123
Electrical Engineering

condition.

VL i p (t)  Vpv − VC p  VS p ; VL i p (t)  Vpv

Vpv − VC p
i L i p (t)  + i L i p (t0 )  i D p (t); i L i p (t)  0
Lip
VL op (t)  −Vbus ; VL op (t)  VC p − Vbus ; VS p  VD p  0
(19)

Figure 6e presents the fifth mode of operation, and its

corresponding waveform is shown in Fig. 5, in which still
switch S p is in on condition of charging the inductor L i p
Meanwhile, the energy in the inductor L i p is completely dis-
charged and now the diode D p gets deactivated.The voltages
across the capacitors C p and C p discharges its respective
voltages through the output inductors L op and L op .

VL i p (t)&  0; VL i p (t)  Vpv  VD p (20)

Vpv
i L i p (t)&  0; i L i p (t)  + i L i p (t0 )  i S p (t) (21)
L i p

VL op (t)  Vpv − VC p − Vbus ; (22)

VL op (t)  Vpv − VC p − Vbus ; VS p  VD p  0 (23)

Figure 6f presents the sixth mode of operation and its

corresponding waveform is shown in Fig. 5. In this mode,
the switches are in off condition and the diode D p is reverse
bias, only the input inductor L i p starts releasing its stored
energy as the diode D p is forward biased. Therefore, the
capacitor C p starts charging via the diode D p and output
inductor L op .

VL i p (t)  0; VL i p (t)  Vpv − VC p  VS p (24)

Vpv − VC p
i L i p (t)  0; i L i p (t)  + i L i p (t0 )  i D p (t)
L i p
(25)

VL op (t)  Vpv − VC p − Vbus (26)

VL op (t)  −Vbus ; VS p  VD p  0 (27)

3 Design of the PV port

The design is considered for 1kW PV port. The PV panel

used in this work supplies a DC voltage of 200V . Usually,
the DC voltage in the prototype is produced from an AC
Fig. 6 Operating modes of the PV port (mode I—mode VI) supply voltage of 230V followed by a diode bridge rectifier.

123
 Electrical Engineering

Therefore, in that case, the DC voltage will be calculated  2.03mH (33)

using Eq. 29.
√
2
The DC link capacitor id designed using Eq. 34.
2Vs
VDC 
π I DCbus
√ Co 
2
2 ∗ 230 2ωVDCbus
  198V (28)
π 2.63

2 × 314 × .02 × 380
However, for designing the PV port, let’s consider the
 551.4μF ∼  600μF (34)
input voltage as 200V . The DC link voltage of the system
is maintained as 380V . Therefore, the duty cycle (δ) for the
switches can be calculated using Eq. 30. To avoid unwanted harmonics from the PV system a LC
filter is introduced and its design is presented in Eqs. 35 and
VDCbus 36.
δ
V pv + VDCbus
I pk
380V Cf  tan θ
  0.65 (29) ωV pk
200V + 380V 
P ∗ V pv
Therefore, the range of duty cycle is selected as.   tanθ
ω*V pv
δmax  0.65 and δmin  0.55. 1000 ∗ 200
The input inductance L i p is calculated using Eq. 31.  ∗ tan10
314 ∗ 200
δmin (1 − δmin )V pv 554.14n F (35)
Lip  (30)
2 ∗ n ∗ I DCbus ∗ f s
The capacitance value is chosen as 440n F
where f s is the switching frequency,I DCbus presents the DC
bus current, n denotes the modular phases 1
Lf 
4π 2 f c2 C f
0.55(1 − 0.55)200 1
Lip   
380 ∗ 20000
2 ∗ 2 ∗ 1000 4 × π 2 × 50002 × 440 × 10−9
 235μH (31)  2.302mH ∼  2.5mH (36)

Since the value of L i p should be lower than L i p for current

waveshaping L i p is considered half of L i p . Therefore, L i p
is considered with a value 115μH 4 Comparison of the proposed multiport
The capacitor C p and C p are calculated by considering converter with similar converter
10% of the voltage ripple and is presented in Eq. 32. topologies
VDCbus ∗ δmax
C p  C p    Table 3 presents the comparison of efficiency of the devel-
2 ∗ n ∗ f s ∗ R L ∗ V
2
c
oped converter with the conventional converter [27]. It is
380 ∗ 0.65 observed that the developed converter has a very few losses
  2 when compared with the conventional system. Table 4 gives
2 ∗ 2 ∗ 20000 ∗ 1000
380
∗ 50 the information regarding the comparison of the developed
 0.43μF ∼
 0.5μF (32) converter with similar converters in terms of component
count, port count and efficiency. From Table 4, it is observed
The output inductors L op and L op are calculated using that the efficiency of the proposed converter is 97% which
Eq. 33. is higher than the topologies presented in [28–30] but its
slightly lower when compared with the topology in [31].
δmax I DCbus The efficiency of the proposed converter is obtained by
L op, op    considering the parameters of the experimental setup. The
16n f s2 C p, p i DCbus
2 parameters and the efficiency calculation with their respec-
0.65 ∗ 2.63 tive results are discussed in the result section. The proposed
  0.263
16 ∗ 2 ∗ (20000)2 ∗ 0.5 ∗ 10−6 ∗ 2
converter uses only fewer switches for its operation and has

123
Electrical Engineering

Table 3 Comparison of
efficiency of the developed Loss in components Modular Landsman Conventional Landsman
converter with the conventional converter converter[27]
converter
Power switches 2 ∗ 4.156W  1 ∗ 14.165W  14.165W
8.312W
Diodes D p , D p 2 ∗ 7.411W  14.822W 2 ∗ 23.17W  46.34W
Passive
components L p , L p 2 ∗ 2.517W  5.034W 1 ∗ 12.137W  1.137W
C p , C p 2∗.0245 W+1∗0.237W  1 ∗ 2.054W  2.054W
0.286W
Total losses 28.454W 63.696W
Target efficiency 97.29% 85.05%
Experimental
efficiency 97%

Table 4 Comparison of the

developed multiport converter [Ref] Required Total number of Efficiency
with similar topologies components Ports (η)
Switches Diodes Inductor Capacitor

[28] 12 1 6 1 3 –
[29] 6 2 2 1 3 –
[30] 3 3 3 3 3 –
[31] 2 2 2 1 3 93%
[32] 3 4 1 2 3 94.8%
[33] 2 2 2 3 2 95.12%
[34] 4 3 5 5 3 97.5%
Proposed 6 6 6 4 3 97%

a very less switching losses when compared with the topolo- pulses for various components of the system in accordance
gies in [28, 29]. The topology presented in [28] cannot be with the variation in the load in order to manage the flow of
extended with more number of ports, but the proposed topol- energy in the HCS.
ogy can be extended to many number of ports. Apart from
these advantages, the proposed converter outperforms well
E. Generation of training dataset
when it is operated using the soft switching methods. The
use of soft switching methods will completely remove the
input current ripples. Figure 7 presents the graph related to To effectively control the HCS with the suggested ANFIS
efficiency and output power of the proposed converter with method, a good training dataset is necessary. Based on the
conventional and similar topologies. power availability from hybrid sources, the load demand and
the state of charge (SOC) of the battery, the following evalu-
ation procedure is used to construct the energy management
training dataset. Equation 37 provides the total generated
5 Energy management system power PHgen (t) .

An ANFIS controller is used for energy management in the

PHgen (t)  PR E S(t) + PBat (t) (37)
proposed HCS. The development of this algorithm was done
with the intention of improving the power transfer capabil-
ities between the source side and the load side. The inputs where PR E S (t)  Ppv (t) + Pwind (t) + PFC (t).
of renewable energy generation, load power and the state- Ppv (t) implies PV power,Pwind (t) implies wind power,
of-charge (SOC) of the battery are used to train the ANFIS PFC (t) implies fuel cell power andPBat (t) implies battery
network. The suggested technique creates effective control power.

123
 Electrical Engineering

Fig. 7 Comparison of efficiency vs output power for a proposed con-

verter vs conventional converter b proposed converter vs similar Fig. 8 a ANFIS for the proposed system b Takagi–Sugeno method
topologies

F. Adaptive neuro-fuzzy inference system (ANFIS)

The reference power (PHgen r e f (t)) required to meet the
load demand at time ‘t’ is presented in Eq. 38.
Normally, the ANFIS is a layered structure, which is well
ref
PHgen (t)  PE V _load (t) (38) illustrated in Fig. 8a. It embraces five functional layers such
as fuzzification, product, normalization, defuzzification and
Similarly, the reference PV, wind, fuel cell and battery output. Figure 8b shows the fuzzy reasoning of the ANFIS.
power are calculated using the equations from (39–41). A common rule set for the first-order Takagi–Sugeno inter-
ference system with three inputs is described in Eqs. (44),
ref
Ppv (t)  Ppv_mppt (t) (39) (45) and (46)

ref
Pwind (t)  Pwind_mppt (t) (40) Rule1 : If X is A1 , Y is B1 and Z is C1 then f 1
 a1 X + b1 Y + c1 Z + d1 (44)
ref
PFC (t)  PFC_mppt (t) (41)
Rule2 : If X is A2 , Y is B2 and Z is C2 then f 2
ref ref  a2 X + b2 Y + c2 Z + d2 (45)
Pbat (t)  PHgen (t) − PR E S (t) (42)

Using the above equations, the the training dataset of the Rule3 : If X is A3 , Y is B3 and Z C3 then f 3
ANFIS is developed and is presented in Eq. 43.  a3 X + b3 Y + c3 Z + d3 (46)
⎡ ⎤ ⎡ ⎤
PR E S (0), PEV_ load (0), S OCbat (0) PB (0) where, X , Y , Z are the input parameters,a1 ,a2 ,a3 ,b1 ,b2 , b3, .
⎢ ⎥ ⎢ ⎥
⎢ PR E S (1), PEV_ load (1), S OCbat (1) ⎥ ⎢ PB (1) ⎥ c1 , c2 , c3 , d1 d2 , d3, represent the linear
⎢ ⎥ ⎢ ⎥
⎢ .. ⎥  ⎢ .. ⎥ (43) parameters,A1 ,A2 ,A3 ,B1 ,B2 , B3, C1 , C2 , andC3 char-
⎣ . ⎦ ⎣ . ⎦
acterize the nonlinear parameters of the ANFIS network.
PR E S (t), PEV_ load (t), S OCbat (t) PB (t) The activation weights relating to the fuzzy rules are

123
Electrical Engineering

calculated as Table 5 Reference power parameters calculation

Power source Power control

W  μ Ai (X ) · μ Bi (Y ) · μCi (Z ) (47)
DC-link PDCr e f (t)  VDClink ∗ I DCr e f
Ppv_mppt (t)
The final rule is obtained as PV Ppv r e f (t)  I pv 
Pwind_mppt (t)
Wind Pwind r e f (t)  Iwind
f i  ai X + bi Y + ci Z + di , i  1, 2 . . . (48) Battery Ibatter y r e f (t)  Pbat r e f (t)
Vbatter y
ref
Fuel cell ref PFC (t)
I FC (t) 
Therefore, the output of ANFIS is presented as y and is VFC

presented in Eq. 49.

 The logical operations are done in this layer, and it is repre-

 Wi f i
y Wi f i   , i  1, 2 . . . (49) sented by π . The output of this layer is presented in Eq. 51.
Wi

R L2, n1 to 27  Wn1 to 27  μAi (PR E S )

where, Wi represents the activation strength of each rule. The
∗μB j (Pload ) ∗ μCk (S OCbat ), (51)
ANFIS layer framework for the proposed work is depicted
i  1, 2, 3; j  1, 2, 3; k  1, 2, 3
in Fig. 7a.

Normalization stage
Fuzzification stage

The normalized firing strengths for fuzzy rules are calculated

The generated power PR E S , PE V _Load and S OCbat are con- in this layer. The normalization is carried out using Eq. 52.
sidered as the inputs for fuzzification layer. A1 ,A2 ,A3 ,B1 ,B2 ,
B3, C1 , C2 , andC 3 constitute the linguistic labels of fuzzy  Wn
theory such as ‘small,’ ‘medium’ and ‘large’ for associated Wn  R L3, n  , n  1, 2, . . . . . . .27
W1 + W2 + . . . . . . .. + W27
with this layer. Three membership functions are assigned for (52)
each input parameter. Bell-shaped membership functions are
employed which can be expressed generally as where, the normalized outputs are RL3,n or Wn ’.

1 Defuzzification stage
R L1, i   2qi , i  1, 2, . . . ..9 (50)
xi − pi
1+ ri
This section is accountable for running an ideological func-
tion that provides an output membership function based on
where pi, qi and ri (referred as premise parameters) repre-
the defined fuzzy rules. Equation 53 provides the results from
sent the center value, width and slope of input parameters,
the defuzzification stage.
respectively.
The obtained outputs are
 Wn
Wn f n  R L4, n 
W1 + W2 + . . . .. + W27
R L1, i  μAi (PG ), i  1, 2, 3 [an PR E S + bn Pload + cn S OCbat + dn ]; n  1, 2, . . . 27
R L1, j  μB j−3 (PL ), j  4, 5, 6 (53)
R L1, k  μCk−6 (S OCbat ), k  7, 8, 9
where RL4,n or Wn ’ fn are the outputs of the defuzzification
where, R L1, i R L1, j and R L1, k represent the output of the layer.
fuzzy layer and μAi (PG ), μB j−3 (PL ) and μCk−6 (S OCbat )
characterize the membership function of the input parameters Final stage
of the fuzzy layer.
The ’THEN’ fuzzy rule is represented by the output stage.It
Logical stage is possible to tally up the signals coming in, and that total

123
 Electrical Engineering

Fig. 9 Flowchart presenting the Implementation of ANFIS for Energy

Management in HCS

  Fig. 10 Solar photovoltaic outputs under case I a irradiance b temper-

is presented as Wn f n .The final output is calculated using
ature c power
Eq. 54.
27

27
 Wn f n the past from the energy sources and how much power is
n1
L5 y Wn f n   27
(54) needed at the given instant in time. If the power produced
n1 n1 Wn
is greater than the amount needed to run the EV loads, then
Once the ANFIS has been properly trained, it can be the surplus can be used to charge batteries. However, if the
entrusted with the authority to control the HCS’s energy amount of power generated is less than the load requirement,
management. Flowchart of the ANFIS training and testing the necessary amount of power must be drawn from the stor-
procedure is depicted in Fig. 8. In response to changes in age systems.
load, the suggested technique estimates the sources’ required
reference power. To do this, the suggested approach needs to
know how much power was generated at a given instant in

123
Electrical Engineering

Fig. 12 Fuel cell outputs under case I a power b pressure

Fig. 11 Wind system outputs under case I a wind speed b wind power

6 Results

G. Simulation results

To verify the dynamic performance of the proposed con-

troller, the simulation of the proposed system is carried out
in MATLAB 2021b version. The hybrid charging station is
modeled by considering a 100 − kW solar PV system. A Fig. 13 Load curves for case I
200 − kW wind turbine system run by permanent magnet
synchronous generator (PMSG). The PEMFC-type fuel cell
is considered with a rating of 100kW .The station batteries of Case-I: Abundant availability of power from renewable
100kW is chosen, which can provide 300kW h of power and a energy source.
total load of 80kW with a fixed load of 40kW are considered.
The reference parameters for the ANFIS-based energy man- In the first case, the reference temperature, irradiation and
agement system are calculated from the equations presented output power of the PV system are obtained from the past
in Table 5. data and compared with the existing data. Likewise, the ref-
Two cases are considered for the proposed system (Fig. 9). erence value of the other renewable sources like wind and
fuel cell systems is obtained. Figure 10 presents the com-
parative graph of reference and measured values related to

123
 Electrical Engineering

Fig. 14 Case I operation of ANFIS for compensating the load demand

with the power from renewable sources

Fig. 16 Solar photovoltaic outputs under case I a irradiance b temper-

ature c power

power. From Fig. 11, it is identified that the instantaneous

wind speed and output power obtained from the wind energy
system are very close to the reference value; hence, it is con-
firmed that there is abundant availability of wind energy in
the HCS. Likewise, the electrolyzer power and hydrogen gas
pressure have been obtained for the proposed system and are
presented in Fig. 12. Figure 13 presents the reference and
actual load curves. The role of the ANFIS controller is to
Fig. 15 a Battery power b battery SoC for case I manage the power and activate the respective switches of the
ports to satisfy the load demand. The controller mostly works
by considering the difference in the power from renewable
the PV system in terms of irradiance, temperature and output systems and the load demand. Figure 14 presents the out-
power. From Fig. 10, it is identified that the instantaneous put waveforms of the reference power, available power and
output obtained from the PV system is very close to the ref- load demand. In this case, it is observed that the difference
erence value; hence, it is confirmed that there is abundant in power from renewable sources is positive; therefore, the
availability of PV source in the HCS. Figure 11 presents the station battery is being charged from the renewable sources.
comparative graph of reference and measured values related Figure 15 presents the battery power and battery SoC for case
to the wind energy system in terms of wind speed and wind 1.

123
Electrical Engineering

Fig. 19 Load curves for case II

Fig. 17 Wind system outputs under case II a wind speed b wind power

Fig. 20 Case II operation of ANFIS for compensating the load demand

with the power from renewable sources

Case-II: Poor availability of power from renewable

energy source.

In the second case, the variation in the climatic conditions

causes variation in the output power of the renewable energy
system. Due to poor availability of power from renewable
sources, the ANFIS controller chose to charge the station bat-
tery with the available power from the renewable resources
to satisfy the base load demand. If there is zero availabil-
ity of power from the renewable sources, then the ANFIS
controller activates the circuit breaker switch to charge the
station battery from the grid. Figure 16 presents the com-
parative graph of reference and measured values related
to the PV system in terms of irradiance, temperature and
output power under varying climatic conditions (spring sea-
son). Figure 17 presents the comparative graph of reference
Fig. 18 Fuel cell outputs under case II a power b pressure
and measured values related to the wind energy system in
terms of wind speed and wind power under varying climatic

123
 Electrical Engineering

Fig. 23 Load power supplied for case II

Fig. 21 a Battery power b battery SoC for case II

Fig. 24 Prototype of the proposed converter

Table 6 Parameters of prototype

Fuel cell emulator

Output voltage 48 V
Output current 15 A
Rated power 720 W
PV simulator
Output voltage 48 V
Output current 5A
Rated power 240 W
Wind emulator
Output voltage 48 V
Fig. 22 Case II operation of ANFIS for compensating the base load Output current 12.5 A
demand with the power from renewable sources and battery power Rated power 600 W

condition. Figure 18 presents the load curves for varying cli- From Fig. 22, it is identified that the ANFIS controller allo-
matic conditions. Figure 19 presents the output waveforms cates the renewable power as well as the battery power to the
of the reference power, available power and load demand. base load based on its availability. The load allocation for
Figure 20 presents the battery power and battery SoC for case II is presented in Fig. 23.
case 2. Figure 21 presents the output waveforms of the avail-
able power from renewable sources and the battery power.

123
Electrical Engineering

Fig. 25 Experimental output voltage and output current waveforms of Fig. 26 Experimental output voltage and output current waveforms of
fuel cell system solar PV system

H. Experimental results

the three ports is 720 W, 600 W and 240 W, respectively.

The prototype of the proposed converter is designed in the LA55-P and LV-25P are the current and voltage sensors used
laboratory, and the experimental results are rendered from the to measure the instantaneous current and voltage at the out-
prototype. Figure 24 presents the hardware of the proposed put. In the prototype, six MOSFET IR540 switches are used
system. The input to the proposed system is fed from three with six ES1B fast recovery diodes. In the prototype, the
transformers with respective diodes. From the diodes, the DC energy management code is programmed into the microcon-
power obtained is passed through the respective capacitors troller and the respective sensors are used to predict the actual
for removing the harmonic content and fed to the individual power and then activate the corresponding switches for feed-
converter ports. Tektronix MSO 2014 is used to obtain the ing power to the base load. Figure 25 presents the output
waveforms of the proposed system. Tektronix TCP 300 cur- voltage and output current of the fuel cell system. Figure 26
rent probe is used to measure the current across the load. The presents the output voltage and output current of the solar
heart of the charging station is the multiport converter, which PV system. Figure 27 presents the output voltage and output
is developed in the laboratory. Table 6 presents the parame- current of the wind energy system. Figure 28 presents the
ters of the proposed system. Finally, the total power fed to output voltage and output current of the MPC.

123
 Electrical Engineering

Fig. 27 Experimental output voltage and output current waveforms of Fig. 28 Experimental output voltage and output current waveforms
Wind Power system from the multiport converter

The total power obtained by the three sources is 1560 W.

The power obtained from the converter is 1512. Therefore,
the efficiency of the system is obtained as 97%. are also performed in the hardware prototype. From the test
results, it is evident that the proposed converter can act as a
high step-up converter for hybrid renewable energy applica-
7 Conclusion tions.

In this work, hybrid renewable energy-based EV charging Author contributions Project administration, Investigation, M.I.;
station is developed using the sources such as solar PV, wind Resources, Writing–review & editing, P.S.; Validation, Resources,
A.A.S.; Validation, D.S.V;Writing–original draft, G.P
energy and the fuel cell. The major contribution of the work
is focused on designing a multiport converter, which is capa- Funding No funding was provided for completing this work.
ble of providing a greater output voltage for charging the
battery banks of the charging station. A flexible control tech-
nique is proposed using the ANFIS that improves the usage
Declarations
of PV, wind power and fuel cell and manages the power flow
in the charging station. To determine the scheme’s feasibil- Conflict of interest All the authors declare that we have no conflict of
ity, detailed simulation studies are performed. Investigations interest.

123
Electrical Engineering

Ethical approval This article does not contain any studies with human 17. Surulivel N, Debnath D, Chakraborty C (2023) A novel Bi-
participants or animals performed by any of the authors. directional four-port DC-DC converter suitable for bipolar DC solar
household integration. IEEE Trans Power Electron
18. Udayakumar AK, Raghavan RRV, Houran MA, Elavarasan
RM, Kalavathy AN, Hossain E (2023) Three-port Bi-directional
References DC–DC converter with solar PV system fed BLDC motor drive
using FPGA. Energies 16(2):624
1. Shunmugham Vanaja D, Stonier AA (2020) A novel PV fed asym- 19. Zandrazavi SF, Guzman CP, Pozos AT, Quiros-Tortos J, Franco
metric multilevel inverter with reduced THD for a grid-connected JF (2022) Stochastic multi-objective optimal energy management
system. Int Trans Electric Energy Syst 30(4):e12267 of grid-connected unbalanced microgrids with renewable energy
2. Jose PS, Jose P, Wessley G, Rajalakshmy P (2022) Environmental generation and plug-in electric vehicles. Energy 241:122884
impact of electric vehicles. In: E-Mobility (pp. 31–42). Springer, 20. Liao J, Yang Y, Luo R, Luo C, Fu C (2022) Energy management
Cham. strategy of four-wheel drive SUV electric-hydraulic hybrid (EHH)
3. Fang X, Wang Y, Dong W, Yang Q, Sun S (2023) Optimal energy power system based on optimal instantaneous efficiency. Sustain
management of multiple electricity-hydrogen integrated charging Energy Technol Assess 52:102139
stations. Energy 262:125624 21. Tran MK, Panchal S, Khang TD, Panchal K, Fraser R, Fowler M
4. Parashar K, Verma K, Gawre SK (2023) Power quality analysis of (2022) Concept review of a cloud-based smart battery management
grid connected solar powered EV charging station: a review. Recent system for lithium-ion batteries: feasibility, logistics, and function-
Adv Power Electron Drives: Select Proceed EPREC 2022:259–271 ality. Batteries 8(2):19
5. Elkeiy MA, Abdelaziz YN, Hamad MS, Abdel-Khalik AS, Abdel- 22. Vishnu Murthy K, Sabareeshwaran K, Abirami S, Bharani Prakash
rahem M (2023) Multiport DC-DC converter with differential T (2022) Overview of battery management systems in electric vehi-
power processing for fast EV charging stations. Sustainability cles. In: E-Mobility (pp. 137–147). Springer, Cham.
15(4):3026 23. Ra N, Ghosh A, Bhattacharjee A (2023) IoT-based smart energy
6. Assadi SA, Gong Z, Coelho N, Zaman MS, Trescases O (2023) management for solar vanadium redox flow battery powered
Modular multi-port electric-vehicle DC fast-charge station assisted switchable building glazing satisfying the HVAC system of EV
by a dynamically reconfigurable stationary battery. IEEE Transac- charging stations. Energy Convers Manage 281:116851
tions on Power Electronics. 24. Marinescu C (2023) Progress in the development and implementa-
7. Masike L, Gitau MN (2023) A modular circuit synthesis oriented tion of residential EV charging stations based on renewable energy
modelling approach for non-isolated DC-DC converters in CCM. sources. Energies 16(1):179
Energies 16(3):1047 25. Enescu FM, Birleanu FG, Raboaca MS, Raceanu M, Bizon N,
8. Dwivedi R, Singh S, Singh B Design of 50kW multilevel AC/DC Thounthong P (2023) Electric vehicle charging station based on
converter fed two phase parallel interleaved DC-DC converter for photovoltaic energy with or without the support of a fuel cell-
fast EV charging stations. electrolyzer unit. Energies 16(2):762
9. Saadaoui A, Ouassaid M, Maaroufi M (2023) Overview of inte- 26. Zhou X, Choy WJ, Alcaide AM, Wang S, Guenter S, Leon JI,
gration of power electronic topologies and advanced control Buticchi G (2023) Common DC-link capacitor harmonic current
techniques of ultra-fast EV charging stations in standalone micro- minimization for cascaded converters by optimized phase-shift
grids. Energies 16(3):1031 modulation. Energies 16(5):2098
10. Bai,Y (2023) Design and implementation of silicon-carbide-based 27. Akhtar MF, Raihan SRS, Rahim NA, Akhtar MN, Abu Bakar
four-switch buck-boost DCDC converter for DC Microgrid Appli- E (2023) Recent developments in DC-DC converter topologies
cations (Doctoral dissertation, Virginia Tech). for light electric vehicle charging: a critical review. Appl Sci
11. Fan J, Wu K, Chen A, Wang M, Xie X, Fan J (2022) Fingerprint- 13(3):1676
like NiCoP electrode material with rapid charging/discharging 28. Wu H, Zhang J, Xing Y (2014) A family of multiport buck–boost
performance under large current density. Journal of Alloys and converters based on DC-link-inductors (DLIs). IEEE Trans Power
Compounds, 163866. Electron 30(2):735–746
12. Huang Z, Zhou D, Wang L, Shen Z, Li Y (2023) A review of single- 29. Hintz A, Prasanna UR, Rajashekara K (2014) Novel modular
stage multiport inverters for multisource applications. IEEE Tran multiple-input bidirectional DC–DC power converter (MIPC) for
Power Electron HEV/FCV application. IEEE Trans Ind Electron 62(5):3163–3172
13. Pradhan SK, Sekaran S, Chakraborty I (2023) A novel non-isolated 30. Akar F, Tavlasoglu Y, Ugur E, Vural B, Aksoy I (2015) A bidirec-
high gain multiport DC-DC converter for integrating fuel cell/solar tional nonisolated multi-input DC–DC converter for hybrid energy
PV and battery energy storage system. Energy Sour, Part A: Recov- storage systems in electric vehicles. IEEE Trans Veh Technol
ery, Utilization, and Environ Eff 45(1):1994–2012 65(10):7944–7955. https://doi.org/10.1109/tvt.2015.2500683
14. Sun Q, Xie H, Liu X, Niu F, Gan C (2023) Multiport PV-assisted 31. Hu R, Zeng J, Liu J, Yang J (2018) Double-input DC-DC converter
electric-drive-reconstructed bidirectional charger with G2V and for applications with wide-input-voltage-ranges. J Power Electron
V2G/V2L functions for SRM drive based EV application. IEEE 18(6):1619–1626
J Emerg Selected Topics in Power Electron. 32. Karthikeyan B, Sundararaju K, Palanisamy R (2021) A variable
15. Gao S, Zhu Q (2023) Wide input range non-isolated three-port con- step size fuzzy logic controller based maximum power point track-
verters for stand-alone PV storage power systems. J Power Electron ing controller for proton exchange membrane fuel cell powered
23(2):191–201 resonant pulse width modulation high step up converter with mul-
16. Jalilzadeh T, Rostami N (2023) New multi-operational multi-port ticarrier sinusoidal pulse width modulation inverter fed induction
DC–DC converter with bidirectional capability. IET Renew Power motor. Int Trans Electr Energy Syst. https://doi.org/10.1002/2050-
Gen 7038.13093

123
 Electrical Engineering

33. Aljarajreh H, Lu DDC, Siwakoti YP, Tse CK, See KW (2021) Publisher’s Note Springer Nature remains neutral with regard to juris-
Synthesis and analysis of three-port DC/DC converters with two dictional claims in published maps and institutional affiliations.
bidirectional ports based on power flow graph technique. Energies
14(18):5751. https://doi.org/10.3390/en14185751 Springer Nature or its licensor (e.g. a society or other partner) holds
34. Mumtaz F, Yahaya NZ, Meraj ST, Kannan R, Singh BSM, Ibrahim exclusive rights to this article under a publishing agreement with the
O (2020) Multi-input multi-output DC-DC converter network for author(s) or other rightsholder(s); author self-archiving of the accepted
hybrid renewable energy applications. In 2020 International Con- manuscript version of this article is solely governed by the terms of such
ference on Innovation and Intelligence for Informatics, Computing publishing agreement and applicable law.
and Technologies (3ICT) (pp. 1–6). IEEE

123