Received 7 June 2024, accepted 31 July 2024, date of publication 5 August 2024, date of current version 14 August 2024.

Digital Object Identifier 10.1109/ACCESS.2024.3439106

Voltage Regulation of Auxiliary Power Units for

Electric Vehicle Applications Using Fuzzy
Logic Controller
ABISHEK M , AMUTHA PRABHA N , AND JITENDRA KUMAR GOYAL
School of Electrical Engineering, Vellore Institute of Technology, Vellore, Tamil Nadu 632014, India
Corresponding author: Amutha Prabha N ([email protected])

ABSTRACT This paper proposes a new design methodology to regulate the output voltages of auxiliary
power units in an electric vehicle application. The proposed method includes designing and implementing
a robust fuzzy logic controller to load variation and input voltage fluctuation. The existing output voltage
regulation techniques for auxiliary power units comprise a proportional-integral controller, which fails to
provide good transient performance concerning load variation and input voltage fluctuation. To overcome
the drawback of the existing technique, thereby improving transient performance, a proportional-integral
controller is replaced by a fuzzy logic controller. This paper addresses a detailed analysis of comparing
the proportional-integral and fuzzy logic controllers to segregate and utilize the best among them based
on their performance. In addition, the issue of cross-regulation is also eliminated for the proposed design.
Also, the proposed design involves the use of a single-input triple-output converter of a 100 W prototype
which is designed to furnish three output voltages 24 V, 14.4 V, and 5 V to supply the auxiliary power units
of an electric vehicle. To validate the effectiveness of the proposed approach, simulation is conducted on
the MATLAB/Simulink platform, and the quantitative results of the proposed design, which include output
voltage (Vo1 ) = 24 V, rise time (Tr ) = 0.02 sec, and steady-state error (ess ) = 0, are better than the existing
values: output voltage (Vo1 ) = 23.4 V, rise time (Tr ) = 0.345 sec, and steady-state error (ess ) = 0.6 V.

INDEX TERMS Auxiliary power unit (APU), electric vehicles (EV), fuzzy logic controller (FLC),
proportional integral (PI) controller, single-input dual-output (SIDO) converter, single-input triple-output
(SITO) converter.

NOMENCLATURE Minimum capacitances C1 min , C2 min , C3 min .

Input voltage Vin . Resistors R1 , R2 , R3 .
Input current Iin . Minimum load resistances RL1 min , RL2 min ,
Switches S1 , S2 , S3 . RL3 min .
Duty cycles d1 , d2 , d3 . Output voltages Vo1 , Vo2 , Vo3 .
Duty ratios when switches ON d1ON , d2ON , d3ON . Output reference voltages Vo1 ref , Vo2 ref , and
Duty ratios when switches OFF d1OFF , d2OFF , d3OFF . Vo3 ref .
Maximum duty cycles d1 max , d2 max , d3 max . Ouput currents Io1 , Io2 , Io3 .
Diodes D1 , D2 , D3 . Load resistance RL .
Inductors L1 , L2 , L3 . ESR of the capacitor rc .
Minimum inductances L1 min , L2 min , L3 min . Reference voltage Vref .
Capacitors C1 , C2 , C3 . Proportional gain Kp .
The associate editor coordinating the review of this manuscript and
Integral gain Ki .
approving it for publication was N. Prabaharan .
Switching frequency fs .

2024 The Authors. This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 License.
VOLUME 12, 2024 For more information, see https://creativecommons.org/licenses/by-nc-nd/4.0/ 107583
 Abishek M et al.: Voltage Regulation of APUs for Electric Vehicle Applications Using FLC

Time period t. divider-based duty generator and anti-right half plane zero
Switching time period Ts . method in [6]. Adding auto-tuning increases the complexity
Rise time Tr . of the circuit. Two inductors are integrated into one magnetic
Peak time Tp . core to reduce cost [7]; In this paper, zero voltage switching
Error e(t). (ZVS) turns ON and OFF, possible for only one switch
Change in error 1e(t). among three switches. A non-isolated single-input three-
Delay in error or memory e(t − 1). output high voltage gain converter is explained in [8]; This
Steady-state error ess . converter is a good choice for low-power applications but
Maximum overshoot Mp . has issues with switch inrush current and equitably high
State vector x. current stress. Deadbeat control, accomplished by adding
Input vector u. a current observer to the control system, is an excellent
Output vector y. way for single-input multiple-output (SIMO) converters to
Small perturbations x̂, ŷ, û, d̂. address the cross-regulation problem [9]. Still, the control
Matrices P, Q, R, S. circuitry becomes complex. Step-up and step-down outputs
Matrices for n modes Pn , Qn , Rn , Sn . are obtained by an integrated dual output converter in [10];
Inductor currents IL1 , IL2 , IL3 . energy transfer is possible from the step-up to the step-
Ripple inductor currents 1iL1 , 1iL2 , 1iL3 . down port. The dual-input single-output (DISO) converter
Capacitor voltages VC1 , VC2 , VC3 . is addressed in [11]; when compared to the traditional
Voltage gains MVin1 , MVin2 , MVin3 . single-input dual-output (SIDO) cuk converter, the current
Controllers transfer function Gc1 , Gc2 , Gc3 . stresses of the proposed one are higher. Controllable inductor
Plant transfer function Gvd1 , Gvd2 , and Gvd3 . power flow loops are designed to create multiple power flow
Proportional integral PI. networks [12]; voltage oscillates when the inductor current
Fuzzy logic controller FLC. drops to zero. Combining boost structure with a traditional
Fuzzy inference system FIS. single-ended primary inductor converter (SEPIC) is enforced
Continuous conduction mode CCM. in [13]; Power loss drawn by a coupled inductor is more
Discontinuous conduction mode DCM. than 40% of total power. A single-input multiple-output
Universal serial bus USB. DC-DC converter is used to provide two separate output
Artificial neural networks ANN. voltages for auxiliary power units in electric vehicles [14];
Hybrid energy storage system HESS. This research does not adequately address the impact of
Triangular membership function TMF. efficiency and input voltage fluctuations on the converter’s
Trapezoidal membership function TRMF. performance. Using FLC and Artificial Neural Networks
(ANN) to split the power of the hybrid energy storage
I. INTRODUCTION system (HESS) for electric vehicles reduces battery discharge
Early in the twentieth century, automobiles relied on 6 V during persistently irregular driving activities [15]. Utilising
electrical systems for cranking, lighting, and ignition. The the triangular membership function (TMF) and trapezoidal
low voltage system upgraded from 6 V to 12 V in the membership function (TRMF) in FLC, the system’s output
middle of the 1950s, which now incorporates auxiliary power is improved, offering a more reliable analysis and fuzzy
units and a 12 V battery [1]. Across a DC/DC converter, clustering means for the insulation status analysis of power
the high-voltage onboard battery charges the low-voltage transformer incipient problems [16]. A fuzzy logic controller
battery to power auxiliary power units. Implementing buck is developed specifically for photovoltaic (PV) applications to
operation is optimal for the low-voltage auxiliary power tackle issues such as non-linearity and output fluctuation [17].
application [2]. A multiple-output converter is the best choice Diverse methodologies are employed for the identification
for this application because an electric car has more possible of faults in power electronic converters, primarily targeting
auxiliary units, which means the converter should also offer components such as capacitors and switching transistors.
more outputs. Various topologies were derived under multiple Methods for analysing inductors include monitoring the
port DC-DC converters on voltage-type ports of a buck current flowing through the inductor, injecting voltage
converter [3]. A fly-back converter was formerly used to signals, and employing models such as the Hidden Markov
provide multiple output voltages, but issues such as leakage Model and Mixed Logic Dynamic Model [18]; Identi-
inductance, magnetizing inductance, and air gaps will affect fying deterioration in switching transistors and determin-
the cross-regulation [4]. A single-inductor multiple-output ing capacitor characteristics pose difficulties. To handle
switcher with simultaneous buck, boost, and inverted outputs the non-linearity of a DC-DC three-level boost converter
is obtained in [5]; However, there are still certain errors in (T-LBC), a unique non-linear integral backstepping control
the state space equations and the output voltage computation. technique is introduced. The control gains are enhanced
Single inductor multiple output buck-boost DC converter using genetic algorithms to enhance the tuning of response
with duty cycle and control current predictor uses auto-tuning control [19]. The importance of achieving efficient voltage

107584 VOLUME 12, 2024

Abishek M et al.: Voltage Regulation of APUs for Electric Vehicle Applications Using FLC

regulation and stability in hybrid electric vehicles (HEVs)

across various load conditions is highlighted in [20]; The
proposed controller, based on fuzzy logic, is demonstrated to
outperform existing control methods in terms of accurately
tracking reference values, minimising overshoot, achieving
rapid convergence, ensuring stability, and regulating the volt-
age of the DC bus bar. The disturbance/uncertainty estimation
and attenuation (DUEA) framework is the preferred method
for improving disturbance rejection in DC/DC converters,
providing less control effort and enhanced capabilities.
DUEA approaches such as disturbance observer-based
control (DOBC) and active disturbance rejection control
(ADRC) are highly effective in reducing disturbances and
improving control performance [21]. Adaptive neuro-fuzzy
inference system (ANFIS) control approach provides notable
benefits such as reducing the overall amount of fuzzy
rules and computational workload, achieving rapid controller
dynamics, demonstrating robustness across a wide range of
load variations, and operating without the requirement of a
mathematical model of the system [22]. SIDO converter is
discussed in [23]; in that study, a PI controller is utilized
to track the desired output voltages under a closed-loop
control technique. The drawback of this technique is that
when the converter is subjected to disturbances like load
variation and supply voltage fluctuation, it behaves turbulent
for some instances and fails to provide good transient
performance. Considering the problem statements from the
literature study, this paper examines the complications faced
by the converter in delivering the indispensable power
for auxiliaries in an electric vehicle and sorts out the
issue. FIGURE 1. (a). Block diagram of the existing single-input double-output
converter with PI controller, (b). Block diagram of single-input
The novel contributions of the paper are as follows: triple-output converter with FLC.

• Number of outputs are increased for existing converter

by adding a converter in a parallel configuration, leading
to extra load utilization.
II. REVISITING EXISTING TECHNIQUES
• Implementing fuzzy logic controller replacing PI con-
The existing technique shows that the converter can produce
troller, which provides good transient performance
two different output voltages for applying auxiliary power
concerning load variation and input voltage fluctuation.
units in an electric vehicle. Also, the PI controller is
• Finally, the quantitative analysis is effectuated for PI
used to track the desired output voltages under closed-loop
and FLC under transient performance with the Simulink
conditions. As a result, it brings the required voltages of 24 V
software.
and 14.4 V, as shown in Figure 1(a).
The structure of the paper is organized as follows: Although the existing system can get the required output
Section II discusses the issues with voltage regulation voltages when it is subjected to load variation or input voltage
in the current SIDO converter, which operates under the fluctuation, it behaves turbulent for some instances and fails
PI controller, as well as the introduction of the SITO to produce good transient performance. This arises due to
converter with FLC. Section III discusses the complete power imbalance while varying the load, and this issue can
analysis of a single input triple output (SITO) converter. be resolved using an effective controller. Keeping this point
Section IV performs small-signal modeling for the SITO in view, in this paper, FLC is substituted for the PI controller
converter. Section V tabulates parameter specifications based to handle this problem effectively. FLC governs a single-
on load requirements. Section VI discusses the various types input triple-output converter to get the desired output voltages
of controllers and the rules used in designing their gain of 24 V, 14.4 V and 5 V by the application of the windscreen
values. Section VII presents the simulation results for better wiper motor, air conditioning system and USB of an electric
understanding. Section VIII provides the conclusion to this vehicle, as shown in Figure 1(b). The SITO converter is
work. Finally, the limitations and future scope are highlighted enforced to engage in source variation, load variation under
in Section IX. open loop and closed loop condition. The performance of

VOLUME 12, 2024 107585

 Abishek M et al.: Voltage Regulation of APUs for Electric Vehicle Applications Using FLC

β2 k3
 
L2 −β2 t
PI and FLC are compared, and the top-notch is chosen for VC2 (t) = Vin − e cosω2 t + ω2 k4
auxiliary power unit application. 2C2 R2
ω2 k3
 
L2 −β2 t
− e sinω2 t −β2 k4 + (4)
III. SINGLE INPUT TRIPLE OUTPUT CONVERTER 2C2 R2
The SITO converter topology is designed to furnish three Vin
iL3 (t) = + e−β3 t [k5 cosω3 t + k6 sinω3 t] (5)
different output voltages for low-voltage applications in R3
β3 k5
 
electric vehicles. Figure 2(a) depicts the circuit topology L3 −β3 t
of the SITO converter. It consists of active and passive VC3 (t) = Vin − e cosω3 t + ω3 k6
2C3 R3
components, namely, switches (S1 , S2 , and S3 ), diodes (D1 , L3 −β3 t

ω3 k5

D2 , and D3 ), inductors (L1 , L2 , and L3 ), capacitors (C1 , C2 , − e sinω3 t −β3 k6 + (6)
2C3 R3
and C3 ) and resistive loads (R1 , R2 , and R3 ) respectively.
The output voltages generated from the SITO converter are 2) SWITCHING STATE 2
(Vo1 , Vo2 , and Vo3 ) based on the duty cycles (d1 , d2 , and
Switching state 2 denotes switches (S1 , S2 and S3 ) are
d3 ). The generated output voltages are lesser than the input
turned OFF simultaneously. During this interval, the supply
voltage (Vin ), which ensures that the given circuit works in
voltage will be disconnected, and the energy stored inside the
buck operation. The benefits of the SITO converter are as
inductors (L1 , L2 and L3 ) will be discharged to the resistive
follows:
loads (R1 , R2 and R3 ) individually. At this condition, the
• Simple circuit configuration to generate multiple output diodes (D1 , D2 and D3 ) function as freewheeling diodes
voltages. and operate in forward-biased conditions. Thus, the flow of
• There are no constraints on inductor current and duty currents through the inductors, capacitors, resistive loads, and
cycle variation. diodes are sketched in Figure 2(c).
• It circumvents the issue of cross-regulation. The inductor current and capacitor voltage equations are
• Circuit configuration can be extended to N-outputs. given as,
The analysis of SITO converter is demonstrated under both
Continuous Conduction Mode (CCM) and Discontinuous iL1 (t) = e−β1 t [k7 cosω1 t + k8 sinω1 t] (7)
Conduction Mode (DCM) of operation as follows: VC1 (t) = −L1 e −β1 t
(−β1 k7 + ω1 k8 ) cosω1 t
− L1 e −β1 t
(ω1 k7 − β1 k8 ) sinω1 t (8)
A. CONTINUOUS CONDUCTION MODE −β2 t
iL2 (t) = e [k9 cosω2 t + k10 sinω2 t] (9)
The CCM mode involves two switching states; switching
state 1 performs the conduction of switches; during this VC2 (t) = −L2 e −β2 t
(−β2 k9 + ω2 k10 ) cosω2 t
occasion, the supply voltage (Vin ) feeds the resistive loads, − L2 e −β2 t
(ω2 k9 − β2 k10 ) sinω2 t (10)
and inductors will get energised. Switching state 2 illustrates iL3 (t) =e −β3 t
[k11 cosω3 t + k12 sinω3 t] (11)
that the energy stored in the inductor will be discharged to the
load through freewheeling diodes. VC3 (t) = −L3 e −β3 t
(−β3 k11 + ω3 k12 ) cosω3 t
− L3 e −β3 t
(ω3 k11 − β3 k12 ) sinω3 t (12)
1) SWITCHING STATE 1
Switching state 1 denotes that the switches (S1 , S2 , and S3 ) whereas, kj is a constant ∀ j = 1,2,. . . 12
are turned ON simultaneously. During this interval, inductors
(L1 , L2 , and L3 ) are energized by the input voltage (Vin ) and 1
dispensed to the resistive loads (R1 , R2 , and R3 ) indepen- β1 = ;
2R1 C1
dently. In this instance, the diodes (D1 , D2 , and D3 ) execute 1
in reverse-biased conditions. The current flow through β2 = ;
2R2 C2
the switches, inductors, capacitors and resistive loads are
1
sketched in Figure 2(b). The inductor current and capacitor β3 = (13)
voltage equations are given as, 2R3 C3
v !
u
1u 1 4
Vin ω1 = t − ;
iL1 (t) = + e−β1 t [k1 cosω1 t + k2 sinω1 t] (1) 2 R21 C12 L1 C1
R1 v
β1 k1
 
L1 −β1 t u !
VC1 (t) = Vin − e cosω1 t + ω1 k2 1u 1 4
2C1 R1 ω2 = t − ;
2 R22 C22 L2 C2
ω1 k 1
 
L1 −β1 t
− e sinω1 t −β1 k2 + (2) v
u !
2C1 R1 1u 1 4
Vin ω3 = t − (14)
iL2 (t) = + e−β2 t [k3 cosω2 t + k4 sinω2 t] (3) 2 R23 C32 L3 C3
R2
107586 VOLUME 12, 2024
Abishek M et al.: Voltage Regulation of APUs for Electric Vehicle Applications Using FLC

FIGURE 2. (a). Single-input triple-output DC-DC converter, (b). Switches are turned ON and the diodes are maintained in reverse-biased condition.
(c). Switches are turned OFF and the diodes are maintained in forward-biased condition, (d). Theoretical waveform of the SITO converter under CCM.

The corresponding output voltage equations for the proposed The theoretical waveform of the SITO converter under
converter are, continuous conduction mode of operation is portrayed in
Figure 2(d). It encompasses the input voltage (Vin ), output
Vo1 = d1 ∗ Vin ; voltages (Vo1 ,Vo2 , and Vo3 ), and inductor currents (iL1 , iL2 ,
Vo2 = d2 ∗ Vin ; and iL3 ). The output voltage remains constant for the given
Vo3 = d3 ∗ Vin (15) input voltage; the inductor current charges when the switch

VOLUME 12, 2024 107587

 Abishek M et al.: Voltage Regulation of APUs for Electric Vehicle Applications Using FLC

is under ON condition and discharges when the switch The duty ratio d1ON is given as,
is OFF. s
2fL1 MVin1
d1ON = (22)
B. DISCONTINUOUS CONDUCTION MODE Ro1 (1 − MVin1 )
During DCM, the switching states of the SITO converter are 2fL1
similar to that of the SITO converter under CCM. But, the d1ON = 1 − (23)
R1
only difference is that the inductor current drops to zero in 2fL1 Io1
switching state two until the subsequent gate pulses are given d1ON = 1 − (24)
Vo1
to the switches. The boundary conditions between CCM and
DCM are given as, The boundary condition between CCM and DCM is given as,
I > 1IL for CCM
s
2fL1 MV2in1
I < 1IL for DCM (16) MVin1 = DB =
Ro1 (1 − MVin1 )
(25)

For DCM in buck converter, Equation 19 is rearranged as follows,

K < Kcrit (17) 2fL1
M 2 + MVin1 − 1 = 0 (26)
where K = 2L/RTs as given in [24]. d1OFF R1 Vin1
The switching states of SITO converter under DCM of The voltage gain obtained by solving the equation 23 under
operation are given as follows, DCM is given as,
1) SWITCHING STATE 1 2
MVin1 = r (27)
Switching state 1 involves the conduction of switches (S1 , 1+ 1+ 8fL1
S2 , and S3 ) and diodes (D1 , D2 , and D3 ) in reverse-biased d12 R1
ON
conditions. The inductor gets charged during this condition,
The duty ratio d1OFF is given as,
and power is delivered to the loads (R1 , R2 , and R3 ), as shown
in Figure 3(a).
 
1
d1OFF = d1ON −1 (28)
MVin1
2) SWITCHING STATE 2 v
u !
Switching state 2 illustrates the energy stored in the inductors u 8fL1
d1OFF = t 1+ 2 −1 (29)
(L1 , L2 , and L3 ) delivered to the loads (R1 , R2 , and R3 ). During d1ON R1
this condition, the switches (S1 , S2 , and S3 ) are turned OFF,
and diodes (D1 , D2 , and D3 ) are maintained in a forward- The same procedure is to be followed to derive the
biased condition, as shown in Figure 3(b). Also, the switching expressions for output voltages (Vo2 , Vo3 ) output currents
state 2 performs when the switches (S1 , S2 , and S3 ) are (Io2 , Io3 ) ripple inductor currents (1iL2 , 1iL3 ) duty ratios
turned OFF, and the diodes (D1 , D2 , and D3 ) are maintained (d2 , d3 ) and voltage gains (MVin2 , MVin3 ) of load 2 and load
in reverse-biased condition. The inductor current gradually 3 respectively refer [25] to derive the equations.
drops to zero until the subsequent pulses are given to the The theoretical waveform of the SITO converter under
switches, as shown in Figure 3(c). When a gate signal is discontinuous conduction mode of operation is portrayed
applied to the switches, the switching state one operation is in Figure 3(d). It consists of the input voltage (Vin ),
executed for the given converter. constant output voltages (Vo1 , Vo2 , and Vo3 ), and charging
The volt-sec balance equation to obtain the output voltage and discharging inductor currents (iL1 , iL2 , and iL3 ) based on
expression for load one is given as, the ON and OFF condition of the switches respectively. When
the switch is OFF, the inductor current gradually drops to
(Vin − Vo1 ) d1ON Ts = Vo1 d1ON Ts (18)
zero and remains the same for some instances. Suddenly, the
The output current equation for load one is given as, inductor starts charging when the gate pulse is given to the
1 Ts
Z
(d1ON + d1OFF ) 1iL1 switches (S1 , S2 , and S3 ) respectively.
Io1 = iL1 dt = (19)
T 0 2
IV. SMALL SIGNAL MODELLING
The ripple inductor current is given as, The dynamic equations of the state-space form are given as,
(Vin − Vo1 ) d1ON Vo d1 (1 − MVin1 )
1iL1 = = 1 ON (20) dx
= Px + Qu;
L1 fL1 MVin dt
The output current for load 1 is rearranged as, y = Rx + Su (30)
Vo1 d12 (1 − MVin1 ) whereas, x denotes state vector, u denotes input vector, and y
Io1 = (21)
2fL1 MV2in1 denotes output vector respectively.

107588 VOLUME 12, 2024

Abishek M et al.: Voltage Regulation of APUs for Electric Vehicle Applications Using FLC

FIGURE 3. (a). Switches are turned ON and the diodes are maintained in reverse-biased condition, (b). Switches are turned OFF and the diodes are
maintained in forward-biased condition, (c). Switches are turned OFF and the diodes are maintained in reverse-biased condition, (d). Theoretical
waveform of the SITO converter under DCM.

The formulas given above are valid for a single switching y = dR1 x + (1 − d)R2 x + dS1 u + (1 − d)S2 u (32)
cycle. The state and output equations for n mode of operation
are written as, where, n = 1, 2 based on the mode of operation.
ẋ = Pn x + Qn u; To linearize the preceding state equation, perturbations in
voltage, current, and duty ratio are applied and expressed as,
y = Rn x + Sn u (31)
The expression for state equations is given as, X = X + x̂; Y = Y + ŷ; U = U + û; d = D + d̂
ẋ = dP1 x + (1 − d)P2 x + dQ1 u + (1 − d)Q2 ; (33)

VOLUME 12, 2024 107589

 Abishek M et al.: Voltage Regulation of APUs for Electric Vehicle Applications Using FLC

whereas X includes a perturbation in the state vector x̂,

Y includes a perturbation in the output vector ŷ, U includes a
perturbation in the input vector û, and d includes a pertur-
bation in the duty cycle d̂, respectively. For the suggested
converter, the average model provided in state-space form is
as follows,
   
iL1 iL1
 iL   iL 
 2  2
d  iL3  = P +  iL3  + QVin
  
(34)
dt VC1 
 
VC1 
 
VC2  VC2 
VC3 VC3
− ( 1−d
 1

0 0 0 L1 ) 0 0
 0
 0 0 0 ( 1−d 2
L2 ) 0  

 0 1−d3 
0 0 0 0 ( L3  )  FIGURE 4. Closed loop diagram of the SITO converter.
P= ( 1−d1 ) −1
 C1 0 0 R1 C1 0 0  
 0 ( 1−d 2
) 0 0 −1
0
 
C2 R2 C2 
1−d3 −1 motor, air conditioner, and USB are considered for design
0 0 ( C3 ) 0 0 R3 C3 purposes. The required voltages for the above loads are 24 V,
1
 
14.4 V, and 5 V, and their corresponding load currents are
 L1  4.1 A, 3 A, and 2.4 A, respectively.
 
 d2  The formulas for designing inductors are:
   
L 
 2 0 0 0 1 0 0 RL1 min
Q=  d3  ; R = 0 0 0 0 1 0 (35) L1 min = ;
  
0 0 0 0 0 1 1 − d1 min
 L3 
 
0 RL2 min
L2 min = ;
 
0 1 − d2 min
RL3 min
0 L3 min = (40)
1 − d3 min
The formula for deriving output voltages and duty cycles is
given in [26], The formulas for designing capacitors are:
d1 max
V̂o1 (s) = GVd1 d̂1 (s); C1 min = ;
2 r c fs
V̂o2 (s) = GVd2 d̂2 (s); d2 max
C2 min = ;
V̂o3 (s) = GVd3 d̂3 (s) (36) 2 r c fs
d3 max
The converter’s transfer function model is given in [27], C3 min = (41)
2 r c fs
V̂o1 (s) Vin The formulas for designing resistors are:
= sL1
(37)
d̂1 (s) s2 L 1 C1 + R1 +1 Vo1 Vo Vo
R1 = ; R2 = 2 ; R3 = 3 (42)
V̂o2 (s) Vin Io1 Io2 Io3
= sL2
(38)
d̂2 (s) s2 L2 C2 + R2 +1 where, RL denotes load resistance and rc denotes ESR of
V̂o3 (s) Vin the capacitor. The parameters are designed with the ripple
= sL3
(39) in inductor current and ripple in output voltage in mind for
d̂3 (s) s2 L 3 C3 + +1
R3 safety concerns on loads, and the values are displayed in
The closed loop diagram for the SITO converter consists Table 1. The switching frequency (fs ) is 10 kHz since IGBT
of output voltages Vo1 , Vo2 , and Vo3 reference voltages switches operate only under 20 kHz and are recommended
Vo1 ref , Vo2 ref , and Vo3 ref transfer function model of the due to less switching loss. The input voltage is supplied from
controllers Gc1 (s), Gc2 (s), and Gc3 (s) and transfer function a battery source of 48 V, widely used in electric vehicle power
model of plants Gvd1 (s), Gvd2 (s), and Gvd3 (s) respectively as trains.
shown in the Figure 4.
VI. DESIGN OF CONTROLLERS
V. PARAMETER SPECIFICATIONS The closed-loop controller design aims to guarantee that
The parameters are designed to apply auxiliary power loads in the output voltage (Vo ) tracks the reference voltage (Vref ).
electric vehicles. Low voltage loads such as windscreen wiper In this paper, the PI controller and FLC are deployed to

107590 VOLUME 12, 2024

Abishek M et al.: Voltage Regulation of APUs for Electric Vehicle Applications Using FLC

TABLE 1. Specifications for components.

TABLE 2. Specification for PI controller.

get the desired output voltages from the system. The design
and working principles of both controllers are discussed in
upcoming sub-sections.

A. PI CONTROLLER
The PI controller is commonly utilized for feedback control
in buck converters due to its simple implementation and
cost-effectiveness. The equation governing the PI controller,
as given in [28] is written as,
Z τ
u(t) = Kp e(t) + Ki e(τ ) dτ (43)
0
According to the above equation 43, u(t) is the control
signal, e(t) is the error signal, Kp is the proportional gain, FIGURE 5. (a). Block diagram of PI controller (b). Block diagram of FLC.

and Ki is the integral gain. However, the proportional gain

Kp , which reduces the rising time, will not minimize the
steady-state error. Despite the fact that the integral gain system behavior and communicating control rules in plain
Ki reduces the steady-state error, it may also intensify the language are prerequisites for FLC. It mitigates the impact
abrupt response. Figure 5(a) displays the PI controller’s block of disturbances like changes in supply voltage and load
diagram. The gain values for the PI controller are given variation. It works on a Fuzzy Inference System (FIS), which
in [14]. The Ziegler-Nichols thumb rule selects the gain consists of three stages, namely,
values, and tuning these gains provides the required output. • Fuzzification – Real-world measurements such as
The Ziegler-Nichols approach is widely used for tuning PI voltage and current are precise, so it is necessary to
controllers because it is easy to implement and can be applied convert them from precise to imprecise (fuzzy) and vice
to many systems. While it offers a satisfactory initial position, versa. Fuzzification refers to the conversion of crisp
it may result in exceeding the desired target. With the same or precise physical values into fuzzy values. In reality,
technique, gain values for converter stages 1, 2, and 3 are this is a mapping procedure that uses membership
calculated. The gain values are displayed in Table 2. functions to translate a crisp value to a value in a
Although PI controllers are reliable, they lack agility. They fuzzy subset. Various membership functions, including
may demonstrate gradual adaptations and sometimes brief triangular, gaussian, trapezoidal, singleton, sigmoid,
overreactions before reaching a stable state. Achieving opti- etc.,. are available. In this work, a triangular membership
mal performance necessitates meticulous tuning. Alternative function is used.
control systems may be more appropriate for highly dynamic • Rule evaluation – The designer establishes fuzzy rules
processes or anomalies in voltage regulation and transient based on the specified requirements. These rules estab-
analysis. lish a connection between the membership functions of
the input and output variables. The construction of these
B. FUZZY LOGIC CONTROLLER (FLC) rules is based on the IF - THEN statement.
This section addresses the method of designing a fuzzy • Defuzzification – For a specific crisp input, numerous
logic controller. FLC is a robust controller which provides fuzzy rules are enabled. The inference engine produces
good transient performance in case of uncertainties such as aggregated fuzzy sets based on the activated rules.
load variation and input voltage fluctuation. Understanding We require a single crisp value from the aggregated

VOLUME 12, 2024 107591

 Abishek M et al.: Voltage Regulation of APUs for Electric Vehicle Applications Using FLC

TABLE 3. Fuzzy rule base. VII. SIMULATION RESULTS

The simulation is carried out in the MATLAB/ Simulink
platform under both open-loop and closed-loop operations
of the SITO converter. Under open-loop and closed-loop
conditions, variations of source and loads are carried out
to extract the robustness of the controller operating under
transient conditions to maintain a constant voltage. The
required rated output voltages for auxiliary power loads such
as windscreen wipers, air conditioners, and USBs are 24 V,
14.4 V, and 5 V, respectively. The SITO converter operates
fuzzy sets to utilize as the control signal for the under both open-loop and closed-loop conditions to obtain
plant. The method of defuzzification involves combining the required output voltages. The FLC and PI controller are
several fuzzy sets into a single crisp value. De- utilized to perform the closed-loop operation. Among them,
fuzzification techniques include the center of gravity the effective controller is selected for auxiliary power unit
(COG), weighted average, mean of maximum, and application.
others. This work uses the centroid for de-fuzzification.
The FLC assigned here has two inputs and one output. The A. OPEN-LOOP CONTROL OF SITO CONVERTER
inputs are combination of error e(t) and rate of change in error Under open-loop conditions, the converter provides three
1e(t) gives the output duty cycle (d) as shown in Figure 5(b). different voltages: Vo1 = 23.44 V, Vo2 = 14.5 V, and Vo3 =
The corresponding equations for inputs are: 4.458 V respectively, as shown in Figure 7(a). There is
e(t) = Vo − Voref a difference in the required load voltage and the obtained
voltage. It is due to the impedance in the circuit, which
1e(t) = e(t) − e(t − 1) (44) arises from the passive elements such as resistor, inductor,
where, Vo denotes output voltage, Voref denotes reference and capacitor.
voltage and e(t − 1) denotes delay in error or memory.
The interaction of input membership functions namely error 1) OPEN-LOOP CONTROL OF SITO CONVERTER UNDER
e(t) and change in error 1e(t), based on the rules framed VARIABLE SOURCE AND CONSTANT LOADS CONDITION
by the designer, produces the crisp outputs of FLC, known Under open-loop condition, when the SITO converter is
as the gain value (or) duty cycle. Figure 6(a) and 6(b) depicts exposed to supply voltage fluctuation for constant load
the input membership functions, error e(t) and change in error application, there is a voltage and current transient occurs
1e(t), Figure 6(c) depicts the output membership function (d) when the source voltage varies from 40 V to 48 V at time
and the fuzzy rules are depicted in Figure 6(d). The selection intervals t = 0.25 s and 48 V to 60 V at t = 0.5 s and
of membership values and rules is framed by being aware of again it drops to 40 V at t = 0.75 s respectively as shown
efficiency, steady state error, overshoot, and ripples. The FLC in Figure 7(b). The output voltages and currents obtained are
is designed based on the rule mentioned below: unstable and varying in nature. This variable output voltage
• If Error e(t) of the output voltage is a Positive Big value will damage the loads.
and Change in Error 1e(t) is a Positive Big value, then
the changes in the Duty cycle must be a Negative Big 2) OPEN-LOOP CONTROL OF SITO CONVERTER UNDER
value. VARIABLE SOURCE AND VARIABLE LOAD 1 CONDITION
Correspondingly, rules are framed for negative small (NS), When both source and load are varied for open loop operation
zero error (ZE), positive small (PS), and positive big (PB) of SITO converter by varying input voltage ranges 40 V, 48 V,
output linguistic variables respectively. Here, we formulate and 60 V for the time intervals t = 0.3 s, t = 0.6 s, and t =
25 fuzzy set rules using the input membership functions. All 0.9 s. Simultaneously, 30% decrement of load 1 at time t =
the rules are framed and listed in the Table 3. from the basis 0.5 s is also carried out, as shown in Figure 7(c). For this case,
of paper [29], [30] for easy implementation. output voltages and currents obtained are varying in nature.
A significant perturbation may result in undesirable Hence, these variable voltages and currents are not suitable
overshoot. Similarly, in comparison to a more substantial for auxiliary power loads. The open loop study of the SITO
perturbation, adding a perturbation of small magnitude to the converter infers that the obtained voltages and currents do
positive big output value of the FLC will take a longer time not match the required voltages and currents. Next, closed-
to track the reference value. loop research is carried out for the same converter under
The gain values in the existing PI controller are obtained by constant source, constant loads, variable source, and variable
applying Ziegler-Nichols thumb rule. whereas the tuning of loads conditions. An effective controller must perform the
gain values of FLC is done by the interaction of membership operation and bring better results to extract the required
functions based on the user. output voltage from the converter.

107592 VOLUME 12, 2024

Abishek M et al.: Voltage Regulation of APUs for Electric Vehicle Applications Using FLC

FIGURE 6. The membership functions of the FLC, inputs: (a). Error, (b). Change in Error, output: (c). Duty cycle, and rule base: (d). Fuzzy
rules.

B. CLOSED-LOOP CONTROL OF SITO CONVERTER t = 0.75 s, respectively. The output voltages vary concerning
Under closed-loop conditions, the required output voltages variation in source voltage, as shown in Figure 7(e). This
are obtained by using FLC. The output voltages obtained are variation is significant and exceeds the output voltage ripple
24.05 V, 14.41 V, and 5 V, and their corresponding current value of 1.2 V for load 1. Similarly, for load 2, the permissible
values are 4.176 A, 3.002 A, and 2.405 A, respectively, output ripple voltage is 0.72 V, and for load 3, it is 0.25 V.
as shown in Figure 7(d). The gain values of FLC are decided The gain values for the PI controller are determined by
by the membership functions assigned to it. While comparing Ziegler-Nicholas method, and the corresponding gain values
the output voltages of open-loop and closed-loop condition, are Kp1 = 0.002, Ki1 = 0.21, Kp2 = 0.083, Ki2 = 1.614, Kp3 =
the obtained voltage for closed-loop are better than under 0.21, and Ki3 = 1 respectively.
open-loop condition. To further verify the effectiveness of
the controllers, the SITO converter is subjected to source 2) CLOSED-LOOP CONTROL OF SITO CONVERTER BY FLC
and load variation. During this condition, transient analysis FOR VARIABLE SOURCE AND CONSTANT LOAD CONDITION
is done for both FLC and PI controllers under small For the given condition of closed-loop control of the SITO
perturbation for source and loads. The controller’s objective converter controlled by FLC for variable source and constant
is to supply constant output voltages to the loads irrespective load application, The FLC provides the required output
of disturbances created by external parameters. voltages and currents, as shown in Figure 7(f). The supply
voltage is intermittent. The value ranges from 40 V, 48 V,
1) CLOSED LOOP CONTROL OF SITO CONVERTER BY PI and 60 V at intervals of 0.25 s, 0.5 s and 0.75 s, respectively.
CONTROLLER FOR VARIABLE SOURCE AND CONSTANT The transient response for varying the source is negligible
LOAD CONDITION for load 1, load 2, and load 3 when compared with the PI
The variable source voltages are 40 V, 48 V, and 60 V, and controller discussed in the previous section. FLC maintains
voltage fluctuation occurs at t = 0.25 s, t = 0.5 s, and the output voltage constant, ensuring that the obtained output

VOLUME 12, 2024 107593

 Abishek M et al.: Voltage Regulation of APUs for Electric Vehicle Applications Using FLC

FIGURE 7. (a). Open-loop waveform of SITO converter, (b). Open-loop source variation for constant loads, (c). Open-loop source and load variation,
(d). Closed-loop waveform of SITO converter, (e). Closed-loop source variation for constant loads by PI controller, and (f). Closed-loop source
variation for constant loads by FLC.

107594 VOLUME 12, 2024

Abishek M et al.: Voltage Regulation of APUs for Electric Vehicle Applications Using FLC

FIGURE 8. (a). 30% Decrement of load 1, (b). 30% Decrement of load 2, (c). 30% Decrement of Load 3, (d). 30% Increment of load 1, (e). 30%
Increment of load 2, and (f). 30% Increment of load 3.

VOLUME 12, 2024 107595

 Abishek M et al.: Voltage Regulation of APUs for Electric Vehicle Applications Using FLC

voltages will not affect the auxiliary power units in an electric

vehicle.

3) PERFORMANCE COMPARISON OF PI AND FLC UNDER

CLOSED-LOOP OPERATION OF SITO CONVERTER FOR
CONSTANT SOURCE AND VARIABLE LOADS CONDITION
This section includes the performance comparison of PI
and FLC of the SITO converter operating under closed-loop
control for constant source and variable load conditions, and
the corresponding waveforms are portrayed in Figure 8. The
load variation of the SITO converter comprises the following
conditions: 30% decrement of load 1, which involves the
decremental of resistive load 1 to 30% from its original value
of R1 = 5.76  to R1 = 4.032  at period t = 0.5 s. The
current value suddenly increases at that instant to compensate
for the power. The output voltage remains constant, as shown
in Figure 8(a). A 30% decrement of load 2 includes the
resistive load 2 that is decreased to 30% from its original
value of R2 = 4.8  to R2 = 3.36  at period t = 0.5 s at
that duration, current increases and output voltage remains
constant, as shown in Figure 8(b). Similar to load one and
load 2 decrements, a 30% decrement of resistive load 3 is
also executed from its original value of R3 = 2.083  to
R3 = 1.458  at t = 0.5 s, the corresponding current value
gets increased, as shown in Figure 8(c). 30% increment of
load 1 comprises the resistive load 1, which is increased to
30% from its original value of R1 = 5.76  to R1 = 7.488 
at t = 0.5 s, as shown in Figure 8(d). During this condition, the
current value will decrease. In Figure 8(e), a 30% increment
FIGURE 9. (a). Transient response of PI and FLC during variation of load
of resistive load 2 is carried out from its original value of R2 = 1 at t = 0.5 s, (b). Rise time of PI and FLC during the variation of load 1.
4.8  to R2 = 6.24  at t = 0.5 s; it gives rise to a decrement
in the current value while load variation occurs. A 30%
increment of load 3 represents a small perturbation in resistive TABLE 4. Transient performance of controllers for load 1 variation of
load 3 from R3 = 2.083  to R3 = 2.7 , and current changes SITO converter.

are portrayed in Figure 8(f). The comparison of the output

voltages tracking their reference values is plotted for both PI
and FLC for better understanding. From this, it is inferred
that FLC performs better than PI controller concerning load
variation. Also, it can be noticed that variation in either one
load does not affect the other loads. This ensures that the issue
of cross-regulation is also solved in this approach. The sudden
increment and decrement of current values at t = 0.5 s were
obtained due to power imbalance while varying the loads.
time are similar for both PI and FLC controllers. Thereby, the
C. TRANSIENT ANALYSIS OF PI AND FLC comparison of all the transient parameters along with output
When the SITO converter is subjected to disturbance by voltage and current for load 1 are entered in Table 4.
decreasing the load 1 at a time interval (t = 0.5 s), both PI and Figure 9(b) shows the sudden response of the voltage and
FLC come into effect simultaneously. Figure 9(a) illustrates current ensuing when the switches are turned ON imme-
the transient response of PI and FLC for variation of load 1. diately. During this condition, the PI controller produces a
PI controller oscillates during load variation, requiring more voltage and inrush current surge that will damage the loads.
time to reach the steady state. Meanwhile, FLC reaches a FLC gives a smooth response and reaches a steady state;
steady state without disturbing the load. In addition, FLC therefore, it will not affect the loads. In addition, the variation
produces a quick response, due to which its rise time is in either of the loads will not affect the other loads. Hence,
minimal. Meanwhile, in the PI controller, the rise time is the issue of cross-regulation is solved for the SITO converter.
longer than FLC, and there is a steady state error, which Similarly, the variation of loads is done for other converters.
results in a drop in the output voltage. The overshoot and peak The simulation result shows that FLC performs better than

107596 VOLUME 12, 2024

Abishek M et al.: Voltage Regulation of APUs for Electric Vehicle Applications Using FLC

PI controller, which is best for voltage regulation of auxiliary [3] H. Zhang, D. Dong, M. Jing, W. Liu, and F. Zheng, ‘‘Topology derivation
power units in an electric vehicle. of multiple-port DC–DC converters based on voltage-type ports,’’ IEEE
Trans. Ind. Electron., vol. 69, no. 5, pp. 4742–4753, May 2022.
[4] C. Ji, M. Smith, K. M. Smedley, and K. King, ‘‘Cross regulation in flyback
VIII. CONCLUSION converters: Analytic model and solution,’’ IEEE Trans. Power Electron.,
In this paper, robust FLC has been designed to ensure smooth vol. 16, no. 2, pp. 231–239, Mar. 2001.
[5] P. Patra, A. Patra, and N. Misra, ‘‘A single-inductor multiple-output
transient performance concerning input voltage fluctuation switcher with simultaneous buck, boost, and inverted outputs,’’ IEEE
and load variation. Also, FLC is implemented on the SITO Trans. Power Electron., vol. 27, no. 4, pp. 1936–1951, Apr. 2012.
converter to achieve the desired output voltages for the [6] Y. Zheng, J. Guo, and K. N. Leung, ‘‘A single-inductor multiple-
application of auxiliary power units in an electric vehicle. output buck/boost DC–DC converter with duty-cycle and control-current
predictor,’’ IEEE Trans. Power Electron., vol. 35, no. 11, pp. 12022–12039,
The complete analysis of the SITO converter includes circuit Nov. 2020.
topology, modes of operation, and small signal modelling. [7] G. Chen, Y. Deng, J. Dong, Y. Hu, L. Jiang, and X. He, ‘‘Integrated
The parameters are designed considering the auxiliary power multiple-output synchronous buck converter for electric vehicle power
supply,’’ IEEE Trans. Veh. Technol., vol. 66, no. 7, pp. 5752–5761,
loads of an electric vehicle, such as the windscreen wiper Jul. 2017.
motor, air conditioner, and USB. To verify the effectiveness [8] Z. Saadatizadeh, P. C. Heris, E. Babaei, and M. Sabahi, ‘‘A new nonisolated
of the proposed approach, simulation is carried out on single-input three-output high voltage gain converter with low voltage
stresses on switches and diodes,’’ IEEE Trans. Ind. Electron., vol. 66, no. 6,
the MATLAB/Simulink R2018a platform under various pp. 4308–4318, Jun. 2019.
conditions, such as open-loop control, closed-loop control [9] X. Zhang, B. Wang, X. Tan, H. B. Gooi, H. H. Iu, and T. Fernando,
that includes constant sources and loads, and variable ‘‘Deadbeat control for single-inductor multiple-output DC–DC converter
sources and loads. The simulation result demonstrates that with effectively reduced cross regulation,’’ IEEE J. Emerg. Sel. Topics
Power Electron., vol. 8, no. 4, pp. 3372–3381, Dec. 2020.
the proposed design yields superior quantitative outcomes [10] O. Ray, A. P. Josyula, S. Mishra, and A. Joshi, ‘‘Integrated dual-output
compared to the existing values. Specifically, the proposed converter,’’ IEEE Trans. Ind. Electron., vol. 62, no. 1, pp. 371–382,
design achieves an output voltage (Vo1 ) of 24 V, a rise time Jan. 2015.
[11] G. Chen, Z. Jin, Y. Deng, X. He, and X. Qing, ‘‘Principle and topology
(Tr ) of 0.02 sec, and a steady-state error (ess ) of 0, whereas synthesis of integrated single-input dual-output and dual-input single-
the existing values are an output voltage (Vo1 ) of 23.4 V, output DC–DC converters,’’ IEEE Trans. Ind. Electron., vol. 65, no. 5,
a rise time (Tr ) of 0.345 sec, and a steady-state error (ess ) pp. 3815–3825, May 2018.
of 0.6 V. The simulation results show that FLC is more [12] H. Zhang, D. Dong, W. Liu, H. Ren, and F. Zheng, ‘‘Systematic synthesis
of multiple-input and multiple-output DC–DC converters for nonisolated
effective than the PI controller in tracking the reference applications,’’ IEEE J. Emerg. Sel. Topics Power Electron., vol. 10, no. 6,
voltage. As a result, auxiliary power units in electric vehicles pp. 6470–6481, Dec. 2022.
can utilise it. In the future, we will implement the proposed [13] M. Premkumar, U. Subramaniam, H. H. Alhelou, and P. Siano, ‘‘Design
and development of non-isolated modified SEPIC DC–DC converter
methodology in a real-time hardware setup. Also, the design topology for high-step-up applications: Investigation and hardware imple-
methodology will be extended to include multiple outputs mentation,’’ Energies, vol. 13, no. 15, p. 3960, Aug. 2020.
for the voltage regulation of auxiliary power units in electric [14] M. Dhananjaya and S. Pattnaik, ‘‘Design and implementation of a SIMO
vehicle applications. DC–DC converter,’’ IET Power Electron., vol. 12, no. 8, pp. 1868–1879,
Jul. 2019.
[15] C. Y. Angundjaja, Y. Wang, and W. Jiang, ‘‘Power management for
IX. LIMITATIONS AND FUTURE SCOPE connected EVs using a fuzzy logic controller and artificial neural
In the proposed methodology, we have considered the network,’’ Appl. Sci., vol. 12, no. 1, p. 52, Dec. 2021.
[16] R. Soni and B. Mehta, ‘‘Diagnosis and prognosis of incipient faults and
idealized converter for designing the controller to regulate insulation status for asset management of power transformer using fuzzy
the output voltages of the auxiliary power units. In practical logic controller & fuzzy clustering means,’’ Electr. Power Syst. Res.,
scenarios, non-ideal conditions will arise, and determining vol. 220, Jul. 2023, Art. no. 109256.
[17] M. A. Hannan, Z. A. Ghani, M. M. Hoque, P. J. Ker, A. Hussain,
the right gain values for the controller to regulate the output and A. Mohamed, ‘‘Fuzzy logic inverter controller in photovoltaic
voltages presents significant challenges. Despite the fact applications: Issues and recommendations,’’ IEEE Access, vol. 7,
that FLC outperforms the existing PI controller, it relies pp. 24934–24955, 2019.
on membership functions, created solely through human [18] R. A. Plazas-Rosas, É. Franco-Mejía, and M. L. Orozco-Gutierrez,
‘‘Methods and challenges for the fault detection and diagnosis in
interaction, to determine its gain values. The future scope power electronic converters: An overview,’’ in Proc. IEEE ANDESCON,
will concentrate on designing an effective, robust controller Nov. 2022, pp. 1–6.
that takes into account model uncertainties and external [19] I. A. Ayad, E. Elwarraki, S. U. Ali, S. M. Qaisar, A. Waqar, M. Baghdadi,
and A. Alzahrani, ‘‘Optimized nonlinear integral backstepping con-
disturbances, with the aim of measuring the performance troller for DC–DC three-level boost converters,’’ IEEE Access, vol. 11,
index accurately. Finally, the real-time hardware setup will pp. 49794–49805, 2023.
be implemented. [20] U. Hussan, M. A. Majeed, F. Asghar, A. Waleed, A. Khan, and M. R. Javed,
‘‘Fuzzy logic-based voltage regulation of hybrid energy storage system
in hybrid electric vehicles,’’ Electr. Eng., vol. 104, no. 2, pp. 485–495,
REFERENCES Apr. 2022.
[1] C. Wang, P. Zheng, and J. Bauman, ‘‘A review of electric vehicle auxiliary [21] A. Korompili and A. Monti, ‘‘Review of modern control technologies
power modules: Challenges, topologies, and future trends,’’ IEEE Trans. for voltage regulation in DC/DC converters of DC microgrids,’’ Energies,
Power Electron., vol. 38, no. 9, pp. 11233–11244, Jan. 2023. vol. 16, no. 12, p. 4563, Jun. 2023.
[2] A. S. Samosir and A. H. M. Yatim, ‘‘Dynamic evolution control for [22] M. B. Nejad, S. M. Ghamari, and H. Mollaee, ‘‘Adaptive neuro-fuzzy
synchronous buck DC–DC converter: Theory, model and simulation,’’ inference systems controller design on buck converter,’’ J. Eng., vol. 2023,
Simul. Model. Pract. Theory, vol. 18, no. 5, pp. 663–676, May 2010. no. 10, Oct. 2023, Art. no. e12316.

VOLUME 12, 2024 107597

 Abishek M et al.: Voltage Regulation of APUs for Electric Vehicle Applications Using FLC

[23] M. Dhananjaya, D. Potnuru, P. Manoharan, and H. H. Alhelou, ‘‘Design AMUTHA PRABHA N received the B.E. degree
and implementation of single-input-multi-output DC–DC converter topol- in ECE from Bharathiar University, and the M.E.
ogy for auxiliary power modules of electric vehicle,’’ IEEE Access, vol. 10, degree in applied electronics and the Ph.D. degree
pp. 76975–76989, 2022. from Anna University. She was the Head of the
[24] W.-M. Zheng, W.-L. Zeng, C.-W. U, C.-S. Lam, Y. Lu, S.-W. Sin, Department. She is currently the Associate Dean
M.-C. Wong, and R. P. Martins, ‘‘Analysis, design and control of an of the School of Electrical Engineering. Well
integrated three-level buck converter under DCM operation,’’ J. Circuits, versed and up-to-date in her field. In the teaching
Syst. Comput., vol. 29, no. 1, Jan. 2020, Art. no. 2050011.
profession, she has a vast experience of more than
[25] S. Maniktala, Switching Power Supplies A-Z. Amsterdam,
25 years. She has handled a wide range of subjects
The Netherlands: Elsevier, 2012.
for undergraduate students from the ECE, EEE,
[26] M. Dhananjaya, D. Ponuru, T. S. Babu, B. Aljafari, and H. H. Alhelou,
‘‘A new multi-output DC–DC converter for electric vehicle application,’’ and EIE streams, and postgraduate students from communication systems.
IEEE Access, vol. 10, pp. 19072–19082, 2022. She laurels the best lecture award for continuous two years. An ardent
[27] A. S. Samosir, T. Sutikno, and L. Mardiyah, ‘‘Simple formula for designing research scholar, she is guiding research in wireless communication and
the PID controller of a DC–DC buck converter,’’ Int. J. Power Electron. wireless sensor networks, image and signal processing, and electric vehicles.
Drive Syst. (IJPEDS), vol. 14, no. 1, p. 327, Mar. 2023. She has also won awards, such as the best lecturer and best paper at
[28] H. Sucu, T. Göktaş, and M. Arkan, ‘‘Design, simulation and application of various levels. She has published about 50 papers in national/ international
buck converter with digital PI controller,’’ Balkan J. Electr. Comput. Eng., journals and conferences. She has organized and conducted various national
vol. 9, no. 2, pp. 106–113, Apr. 2021. and international conferences and workshops in the fields of VLSI design,
[29] P. Mattavelli, L. Rossetto, G. Spiazzi, and P. Tenti, ‘‘General-purpose fuzzy wireless networks, and advanced communication systems. She has also
controller for DC–DC converters,’’ IEEE Trans. Power Electron., vol. 12, played a vital role in conducting and coordinating various other events.
no. 1, pp. 79–86, Jan. 1997. Her research interests include among other things, wireless LAN networks,
[30] S. K. Gadari, P. Kumar, K. Mishra, A. R. Bhowmik, and A. K. Chakraborty, and discrete time linear systems. She is a fellow of IETE. She is also a
‘‘Detailed analysis of fuzzy logic controller for second order DC–DC prominent member of MISTE. She also serves as a Ph.D. committee member,
converters,’’ in Proc. 8th Int. Conf. Power Syst. (ICPS), Dec. 2019, pp. 1–6. and member of various panels and committees of educational institutions in
various capacities.

JITENDRA KUMAR GOYAL received the bach-

elor’s degree in electronics and communication
engineering from Lingaya’s University, Faridabad,
Haryana, India, in 2013, the master’s degree in
electronics and instrumentation from the YMCA
University of Science and Technology, Faridabad,
ABISHEK M was born in Karnataka, Tamil Nadu, in 2015, and the Ph.D. degree in system engineer-
India. He received the B.E. degree in electrical and ing from Indian Institute of Technology (BHU),
electronics engineering from the SNS College of Varanasi, India, in 2021. He was an Assistant
Engineering, Coimbatore, Tamil Nadu, in 2017, Professor with the Department of Instrumentation
and the M.E. degree in power electronics and and Control, NIT, Jalandhar, India, from 2015 to 2016. He was a Postdoctoral
drives from the College of Engineering Guindy, Fellow with LS2N Ecole Centrale De Nantes, France. He is currently
Anna University, Chennai, Tamil Nadu, in 2020. an Assistant Professor with the School of Electrical Engineering, VIT,
He is currently pursuing the Ph.D. degree with the Vellore, Tamil Nadu, India. He has published many research papers in
School of Electrical Engineering, Vellore Institute reputed journals and international conferences. His research interests include
of Technology, Vellore, Tamil Nadu. His research designing robust controllers, sliding mode control, linear matrix inequalities,
interests include electric vehicles, power converters, auxiliary power units, control applications to electric vehicle, and power system applications.
multi-port dc-dc converters, and renewable energy sources and controllers.

107598 VOLUME 12, 2024