high voltage gain interleaved boost converter with anfis ...joics.org/gallery/ics-1981.pdf · anfis...
TRANSCRIPT
High Voltage Gain Interleaved Boost Converter with ANFIS Based
MPPT Controller for Fuel Cell Based Electric Vehicle Applications
Manasamudram Kiran1, *Penagaluru Suresh2, Dr. Sudheer Kasa3 1PG Student, 2,3Associate Professor
Department Of EEE
S.V. College of Engineering, Tirupati-517507, India [email protected], *[email protected], [email protected]
Abstract
Because of the more energetic guidelines on Carbon gasoline emanations and efficiency,
Fuel Cell Electric Vehicles (FCEV) are ending up an increasing number of main stream in the
automobile enterprise. This paper suggests a neural framework based totally Maximum Power
Point Tracking (MPPT) controller for 1.26 kW Proton Exchange Membrane Fuel Cell (PEMFC),
giving electric vehicle powertrain via a high voltage-gain DC-DC bolster converter. The proposed
ANFIS based MPPT controller uses the Maximum Power Point (MPP) of the PEMFC. High
buying and selling repeat and high voltage advantage DC-DC converters are essential for the
catalyst of FCEV. In order to perform excessive voltage increment, a 3-arrange high voltage
expansion Interleaved Boost Converter (IBC) is in like manner predicted for FCEV device. The
interleaving framework diminishes the information cutting-edge swell and voltage weight at the
strength semiconductor devices. The introduction examination of the FCEV structure with RBFN
based MPPT controller is differentiated and the Fuzzy Logic Controller (FLC) and ANFIS in
MATLAB/Simulink arrange.
Key words_ Fuel Cell Electric Vehicle, High Voltage Gain IBC, PEMFC, MPPT, RBFN
I. INTRODUCTION
Because of the ecological contamination and limited stores of non-renewable energy
sources, car businesses are demonstrating more enthusiasm for Fuel Cell Electric Vehicles
(FCEV). The fast progressions in power gadgets and energy component advancements have
engaged the critical improvement in FCEVs [1-2]. Energy components have the benefits of clean
control age, high dependability, high productivity and low commotion [3]. Contingent upon the
sort of electrolyte substance energy units are classified into various kinds, for example, Proton
Exchange Membrane Fuel Cell (PEMFC), Alkaline Fuel Cell (AFC), Phosphoric Acid Fuel Cell
(PAFC), Solid Oxide Fuel Cell (SOFC) and Molten Carbonate Fuel Cell (MCFC). Among these,
PEMFCs are commanding the car business because of their low working temperature and the
speedy startup [4]. The yield voltage of energy component relies upon film water substance and
cell temperature. Prominently, power modules have non-straight voltage-current qualities. Thus,
Journal of Information and Computational Science
Volume 9 Issue 12 - 2019
ISSN: 1548-7741
www.joics.org691
there is just a solitary one of a kind working point accessible for energy units with the most extreme
yield voltage and power. The most extreme power point following (MPPT) procedure is important
to separate the greatest power from the energy component at various working conditions. In the
writing, different MPPT strategies are accessible like bother and watch (P&O), practical swarm
enhancement (PSO), gradual conductance (INC), fluffy rationale control (FLC), sliding mode
control, neural system (NN) to follow greatest power point (MPP) [5]. Among these accessible
MPPT calculations, P&O is straightforward, mainstream and simple to execute. P&O [6] and
gradual conductance techniques produce motions at consistent state which will decrease the
proficiency of the energy component framework. To beat this issue, fluffy rationale controller and
neural system [6] calculations are acquainted with track the MPP with expanded proficiency and
precision. In this paper, spiral premise capacity arranges (RBFN) base MPPT controller is
proposed to follow the MPP of the PEMFC. The power train engineering of FCEV is appeared in
Fig. 1. A pile of PEMFC produces an unregulated low DC yield voltage. So, a lift or venture up
DC-DC converter is required to support and control the PEMFC yield voltage. Lift converter is
broadly utilized as a front-end control conditioner for the energy unit. For low control applications,
the ordinary lift converter Is applied as a energy electronic interface while for excessive manipulate
packages assist converter likely may not be best resulting from its low cutting-edge handling
potential and warm administration problems [7]. To beat these problems numerous high voltage
advantage DC-DC converters are deliberate in the writing. In [8], a quadratic lift converter made
out of two lift converters is proposed to accomplish high voltage gain. Be that as it may, utilizing
of two lift converters may lessen the general effectiveness of the framework. A fell 2-stage
interleaved DC-DC support converter is proposed in [9]. In any case, this topology experiences
poor dependability and less productivity. In [10], a lift converter with voltage multiplier cell is
proposed to accomplish high voltage gain, yet the voltage increase of single multiplier cell isn't
much enough to drive the powertrain of FCEV. Detached converters with coupled inductors or
high recurrence transformers are proposed to accomplish high voltage gain in [11].
The high voltage increase is accomplished by modifying the transformer turns proportion
[12]. Be that as it may, these separated converters are progressively costly contrasted with non-
disconnected DC-DC converters. Thus, this paper proposes a high voltage increase three-stage
non-disengaged interleaved support converter (IBC) for energy unit applications to accomplish
low exchanging pressure and high voltage gain. Interleaving system expands the unwavering
quality of the energy unit and gives high control capacity.
Journal of Information and Computational Science
Volume 9 Issue 12 - 2019
ISSN: 1548-7741
www.joics.org692
Fig 1. Conventional configuration of fuel cell fed BLDC motor driven electric vehicle.
The yield voltage of the proposed converter is given to the electric engine through an inverter for
impetus of the vehicle. The electric engine assumes a significant job in FCEVs. A sufficient engine
extensively diminishes the expense and size of the fuel cell. In past, most of automakers are utilized
DC engines for electric vehicle applications. Antagonistically, DC engines have high upkeep cost
and low proficiency because of the brushes and turning gadgets.
At present, perpetual magnet BLDC engine is for the most part utilizing in FCEV
applications because of straightforward control, high unwavering quality and high roughness.
Fig.2 demonstrates the proposed BLDC engine driven FCEV framework with three-stage high
voltage gain IBC. It comprises of a 1.26 kW PEMFC, three-stage high voltage gain IBC, voltage
source inverter (VSI) and a BLDC engine. The three-stage IBC works as an interface among
PEMFC and VSI. RBFN based MPPT calculation is intended to remove the greatest power from
the energy component. Three-stage IBC supplies capacity to the BLDC engine through VSI. The
switches of the VSI are constrained by utilizing electronic recompense of BLDC engine. The
engine shaft is associated with vehicle wheels for the drive. The remainder of the paper is
composed as pursues.
PEMFC demonstrating is talked about in Section II; displaying of the proposed converter
is canvassed in Section III; MPPT and BLDC engine control methods are depicted in Section IV;
reproductions and results are examined in Section V and the ends condensed in Section VI.
Journal of Information and Computational Science
Volume 9 Issue 12 - 2019
ISSN: 1548-7741
www.joics.org693
Fig 2. Proposed configuration of fuel cell fed BLDC motor driven electric vehicle.
II. FUEL Cell MODELING
A power gadget is an electrochemical contraption that changes over hydrogen fuel into
power. The commitments to the vitality part are air and fuel and these are changed over into water
and power through an engineered reaction. A lone vitality segment includes two terminals (anode
and cathode) and an electrolyte. The electrolyte separates the positive and negative charged
particles of the hydrogen fuel. Exactly when the hydrogen and oxygen are supported into the cell,
control is created at the yield of the cell inside seeing an electrolyte. Power gadget makes just
warmth and water as the wastage of the creation response. The cell voltage of PEMFC is given as
VFC = ENernst − Vact − Vohm − Vcon
(1)
Where ENernst is the open-circuit (or reversible) thermodynamic voltage and is given as
ENernst=1.229-8.5*10−4(T-298.15) +4.308*10−5T (ln)+0.5ln (P02)
(2)
Where T is absolute temperature (K), PO2 and PH2 are oxygen and hydrogen partial pressures
(atm) respectively. Activation voltage Vact is the combination of both anode and cathode
activation overvoltage and is expressed as
Vact = −[δ1 + δ2Tln (C02) + δ4Tln (IFC)]
(3)
Where i d (i= 1,2,3,4) is empirical coefficient for each cell and CO2 is the dissolved oxygen
concentration at the liquid/gas interface and is calculated by using the following expression
Journal of Information and Computational Science
Volume 9 Issue 12 - 2019
ISSN: 1548-7741
www.joics.org694
C02 =P02
(5.08∗106)∗exp (−498
T)
(4)
Ohmic overvoltage Vohm is expressed as
Vohm = IFC(Rc + RM)
(5)
Where RM is the electron flow equivalent resistance and RC is the proton resistance. RC is
considered as constant.
RM =ρm L
A
(6)
Where L is membrane thickness (cm), A denotes active area of membrane (cm2) and m r is the
membrane specific resistivity (Ω-cm) and is given as
ρm =181.6[1+0.03J+0.062(
T
303)2(J)2.5]
[G−0.634−3J]exp [4.18(1−303
T)]
(7)
Where G is water content of the membrane and J is current density and is expressed as
J =IFC
A
(8)
Finally, the concentration overvoltage Vcon can be calculated from the following expression
Vcon = −RT
nF ln [1-
J
Jmax]
(9)
Where F is Faraday’s constant, R is universal gas constant and Jmax is maximum greatest current
thickness. A DC-DC converter is associated with the yield of the power device to keep up a
consistent voltage over the DC connect.
III. THREE-PHASE HIGH VOLTAGE GAIN IBC
The proposed converter includes three switches (S1, S2 and S3) and three diodes (D1, D2
and D3). L1, L2 and L3 are the filtering inductors of stage 1, arrange 2 and stage 3 independently.
VFC is the information voltage, VO is yield voltage and R is the stack resistor. The going with
suppositions are considered for the examination of proposed high voltage gain IBC: I. Inductors
of all the three phases are believed to be immaculate (L1= L2 =L3=L). ii. Isolating capacitors C1
and C2 are considered as same (C1 = C2=C). iii. The proposed converter reliably works in
Continuous Conduction Mode (CCM). iv. The voltage and current swells over the capacitor and
inductor are believed to be close to nothing. The switches S1, S2 and S3 are turned ON by utilizing
two entry beats which are 180˚ stage moved. One portal heartbeat is given to the switch S2 and
another gateway beat with 180˚ stage move is given to both the switches S1 and S3.
Journal of Information and Computational Science
Volume 9 Issue 12 - 2019
ISSN: 1548-7741
www.joics.org695
(a)
(b)
(c)
Fig.3. Modes of operation of 3-phase high voltage gain IBC.
Mode-1 (to ≤ t ≤ t1):
During this mode, all the three switches S1, S2 and S3 are switched ON and all the three
diodes D1, D2, and D3 are reverse biased as shown in Fig. 3(a). The input voltage source
VFC charges the inductors L1, L2 and L3. The current through these inductors I1, I2 and
I3 increased linearly with a slope of (VFC/L). The input capacitor Cin is disconnected from
the load as well as from the supply. The output capacitors C1 and C2 supplies energy to
the load resistor and the voltage of output capacitors VC1 and VC2 decreases with a slope
of (VO/RC).
Mode-2 (t1 ≤ t ≤ t2):
In this mode, the switch S2 is switched ON and the switches S1 and S3 are switched OFF.
The diodes
D1 and D3 are forward biased and the diode D2 is reverse biased as shown in Fig. 3(b).
The current through the inductors L1 and L3 decreased with a slope of (VFC –VCin)/L and
(VFC – VC2)/L respectively. The current through the inductor L2 increases with a slope
of (VFC/L). The capacitor C1 supplies the energy to the load and the capacitors C2 and
Cin are charged by the input voltage VFC.
Mode-3 (t3 ≤ t ≤ t4):
This mode is similar to mode-1. All the three switches S1, S2 and S3 are switched ON and
all the three diodes D1, D2 and D3 are switched OFF.
Journal of Information and Computational Science
Volume 9 Issue 12 - 2019
ISSN: 1548-7741
www.joics.org696
Mode-4 (t4 ≤ t ≤ t5):
In this mode, the switch S2 is switched OFF and the switches S1 and S3 are switched ON.
The diodes D1 and D3 are reverse biased and the diode D2 is conducting as shown in Fig.
3(c). The input voltage source VFC charges the inductors L1 and L3 and the current
through these inductors increases with a slope of (VFC/L). The current through the inductor
L2 decreases with a slope of (VFC+VCin-VC1)/L. The capacitors C2 and Cin supplies
energy to the load. Capacitor C1 gets charged by the input voltage VFC.
A. ANALYSIS OF THE CONVERTER
To enhance the examination of the converter, inductors, capacitors and strength
semiconductor gadgets are believed to be flawless and the converter working in CCM. The static
voltage benefit (M) of the DC-DC converter is obtained with the aid of applying volt-2nd stability
on inductors L1, L2 and L3. By making use of volt second parity to the inductor L1, we get.
VL1 = VFC(t1 − t0) + (VFC − VCin)(t2 − t1) + VFC(t3 − t2) + VFC = 0
(10)
From Eq. (10), input capacitor voltage VCin is acquired as
VCin =VFC
(1−D) (11)
By applying volt-second balance to the inductor L2, we get
VL2 = VFC(t1 − t0) + (VFC)(t2 − t1) + VFC(t3 − t2) + (VFC + VCin − VC1) = 0
(12)
By applying volt-2d balance to the inductor L2, we get
VC1 =VFC
(1−D)+VCin (13)
Journal of Information and Computational Science
Volume 9 Issue 12 - 2019
ISSN: 1548-7741
www.joics.org697
Fig.4. Steady-state waveforms of 3-Phase high voltage gain
From Eq. (11) and Eq. (13) we get
VC1 =2VFC
(1−D) (14)
By making use of volt-2nd balance to the inductor L3, we get
VL3 = VFC(t1 − t0)+(VFC−VC2)(t2−t1)+VFC(t3−t2)+VFC(t4−t3)=0 (15)
From Eq. (15), capacitor C2 voltage is acquired as
VC2=
VFC(1−D)
(16)
The output voltage of the converter is acquired via using the Eq. (17)
vo = VC1 + VC2 − VFC
(17)
From Eqs. (14) (16) and (17), the converter static voltage benefit M is acquired as
M=V0
VFC=
(2+D)
(1−D) (18)
.
The switches S1 and S3 are switched OFF in mode-2 and stays switched ON in all of the
different modes. From Fig. 3(b), the voltage stress of switches S1 and S2 may be expressed as
VS1 = VCin =VFC
(1−D) (19)
VS2 = VC1 − VCin =VFC
(1−D) (20)
The transfer S2 is switched OFF in simplest mode-4. From Fig. Three(c), the voltage
pressure of transfer S3 is expressed as
Journal of Information and Computational Science
Volume 9 Issue 12 - 2019
ISSN: 1548-7741
www.joics.org698
VS3 = VC2 =VFC
(1−D) (21)
In the same way the voltage stress of the diodes D1, D2 and D3 can be derived and they expressed
as
VD1 = −VC1=−2VFC
(1−D) (22)
VD2 = VCin − VC2=−VFC
(1−D) (23)
VD3 = VC2= VFC
(1−D) (24)
Inductance L is designed by using the usage of the input current ripple (ΔI). The maximum
enter present day ripple is believed as 20% of the enter contemporary. The value of the input
inductor is calculated by the use of the Eq. (25).
L=L1=L2=L3=DVFC
∆Ifs (25)
Similarly, the input and output capacitors are designed by means of the use of the voltage
ripples throughout the enter and output capacitors. The voltage ripple (ΔV) is taken into
consideration as 10% of the enter voltage
Cin =Vo
R∆Cinfs (26)
C=C1 = C2 =DVo
R∆Vfs (27)
IV. CONTROLLER DESIGN
Two control procedures are utilized for the proposed arrangement. One is for to follow
the most extreme intensity of the energy component and another is for BLDC engine activity.
A. RBFN BASED MPPT CONTROLLER
MPPT is needful for energy component framework to remove the most extreme power
from it at various temperature conditions. For the proposed setup RBFN based MPPT controller is
created and the outcomes are contrasted and FLC. RBFN is a sort of feedforward neural system
model and has both directed and solo learning stages. RBFN commonly involves three layers: a
data layer, a covered layer and a yield layer as appeared in Fig.6. The shrouded layer comprises of
non-direct spiral premise actuation work though the yield layer is straight one . The hubs in the
info layer are utilized to transmit the contributions to the concealed layer. The net info and yield
of the information neuron are spoken to as
xi(1)
(n) = neti(1) (28)
yi(1)
(n) =fi(1)
[neti(1)
(n)] = neti(1)
(n) i=1,2 (29)
Journal of Information and Computational Science
Volume 9 Issue 12 - 2019
ISSN: 1548-7741
www.joics.org699
Where xi (1) is info layer, yi (1) is shrouded layer and neti (1) is entirety of the information layer.
Each hub in the shrouded layer executes as Gaussian capacity. The Gaussian capacity is utilized
as a participation work in the RBFN.
Fig.6. RBFN structure.
netj(2)= -(X-Mj)
TEj(X-Mj) (30)
yj(2)(n) = fj
(2) [netj(2)(n)]xp[netj
(2)(n)] ,j=1,2,......
(31)
Where Mj and Σj are mean and standard deviation of the Gaussian capacity separately. The yield
layer has
single hub k, produces the straight control signal
netk(3) =Ejwjyj(2) (32)
yk(3) =f(3)k [net k(3)(n)] =net(3)k(n) (33)
Where wj is the connective weight lattice among yield and concealed layer. In this paper current
and voltage of energy unit are taken as contributions to the RBFN controller and it produces
obligation cycle (D) as the yield as appeared in Fig. 7.
Fig.7. RBFN based MPPT architecture for the fuel cell.
Journal of Information and Computational Science
Volume 9 Issue 12 - 2019
ISSN: 1548-7741
www.joics.org700
B. ELECTRONIC COMMUTATION
The control sign to the switches of the VSI are gotten from the BLDC engine electronic
substitution. Three lobby sensors are utilized to produce three corridor sign contingents upon the
engine rotor position for every interim of 60˚. These created lobby signs are changed over into
changing heartbeats to the VSI by utilizing a decoder circuit. The exchanging conditions of VSI
are recorded in Table 1.
TABLE 1. Switching states for electronic commutation of BLDC motor
V. Adaptive Neuro-Fuzzy Inference System
An adaptable neuro-cushioned assurance arranges or adaptable structure based fleecy
pondering system (ANFIS) is a kind of false neural shape that depends on Takagi–Sugeno quiet
enrollment coordinate. The strategy was made in the mid-Nineteen Nineties. Since it makes each
neural frameworks and padded clarification measures, it can get the upsides of both in a lone
structure. Its enlistment structure thinks about to a machine of padded IF–THEN determinations
which have learning ability to off base nonlinear cutoff focuses. Along these lines, ANFIS is
analyzed to be a completed estimator. For utilizing the ANFIS as to a couple of recognition an
evidently decent and most ideal way, you can utilize the top-notch parameters got by methods for
characteristic calculation. ANFIS: Artificial Neuro-Fuzzy Inference Systems
1. ANFIS are a class of adaptable structures which can be all round that truly matters
foggy to delicate fame frameworks.
2. ANFIS adapt to Sugeno e Tsukamoto padded models.
3. ANFIS utilizations a mutt getting the hold of figuring.
In the circle of electronic thinking neuro-padded translates blends of trickery neural
structures and woolen assistance. Neuro-pleasant hybridization gets a handle on a cream sharp
Journal of Information and Computational Science
Volume 9 Issue 12 - 2019
ISSN: 1548-7741
www.joics.org701
shape that synergizes these two methods by method for joining the human-like contemplating style
cushioned systems with the getting to know and connectionist shape of neural structures. Neuro-
cushioned hybridization is all around named as Fuzzy Neural Network (FNN) or Neuro-Fuzzy
System (NFS) in the made paintings. Neuro-cushioned structure (the greater well-known term is
utilized from this time ahead) wires the human-like considering style warm frameworks the usage
of agreeable sets and a semantic version which includes a approach of IF-THEN cushioned
benchmarks. The crucial idea of neuro-soft systems is that they're paying little thoughts to what
you resemble at it approximates with the capability to request interpretable IF-THEN runs the
display.
The possibility of neuro-sensitive structures wires two conflicting requirements in open to
acting: interpretability versus exactness. All round that in reality topics, one of the properties wins.
The neuro-woolen in cushioned indicating research area is restrained into two zones: semantic
touchy showing this is revolved round interpretability, by using and massive the Mamdani display;
and right feathery demonstrating that relies upon upon exactness, in a general sense the Takagi-
Sugeno-Kang (TSK) layout.
Watching out for fuzzification, cushioned inducing and defuzzification via multi-layers
feed-forward connectionist frameworks. It must be raised that interpretability of the Mamdani-
type neuro-agreeable structures may be misplaced. To refresh the interpretability of neuro-
cushioned systems, sure value determinations should be taken, wherein important bits of
interpretability of neuro-agreeable systems are in like manner separated. A propelling appraisal
line watches out for the fact’s movement mining case, wherein neuro-cushioned structures are
successively resuscitated with new pushing towards fashions on asking for and on-the-fly. Thusly,
shape restores don't just merge a recursive distinction in model parameters, but moreover a unique
development and pruning of version with a selected intense goal to oversee thought flow and
continuously changing framework direct sufficiently and to maintain the frameworks/fashions
"within the gift style" at something factor.
ANFIS MPPT controller
ANFIS is the combination of both neural networks and fuzzy inference sytem. Thus, it offers the
benefits of inference mechanism of fuzzy system and learning ability of neural network. Fig.8
shows the typical ANFIS structure of the Sugeno fuzzy system. It has two inputs (X and Y) and
single output (F) with 5 layers. The fuzzy if-then rules of the Sugeno fuzzy system are given as
follows:
Rule 1: If X is A1 and Y is B1, then F1 ¼ p1X þ q1Y þ r1
Rule 2: If X is A2 and Y is B2, then F1 ¼ p2X þ q2Y þ r2
In Fig.8, square specifies an adaptive node, whereas a circle specifies a fixed node. The purpose
of each layer in the ANFIS
Journal of Information and Computational Science
Volume 9 Issue 12 - 2019
ISSN: 1548-7741
www.joics.org702
Layer 1: In this layer the input values are converted to fuzzy values. Each node in fuzzification
layer is an adaptive node with a node function given as
O1,i =µAi(X) ,for i= 1,2
(34)
O1,i = µBi-2 (Y) ,for i = 3, 4 (35)
Fig.8 ANFIS architecture.
where O1;i is membership value for the inputs X and Y, m is membership function. The subscripted
i and 1 indicates the node and layer numbers respectively.
Layer 2: This layer is called as product layer. Each node in this layer is fixed node. The output
of this layer is product of every incoming signals. The output of each node indicates the firing
strength of each rule.
O2,i = wi = µAi(X)µBi(Y) i = 1,2
(36)
Layer 3: The third layer in Sugeno fuzzy system is normalization layer. Each node in this layer
is a fixed node. This layer normalizes the firing strength by dividing rule's firing strength with
sum of firing strengths of all the rules.
𝑂3𝑖 = 𝑤𝑖 =𝑤1
𝑤1+𝑤2 (37)
Layer 4: It is defuzzification layer. Each node in this layer is an adaptive node with node
function given as:
O4,i = wifi =wi ( piX + qiY + ri ) i =1, 2
(38)
Layer 5: It is output layer and has only a fixed node. It calculates the overall output value by
summing all the incoming signals.
Journal of Information and Computational Science
Volume 9 Issue 12 - 2019
ISSN: 1548-7741
www.joics.org703
O𝑜5 = ∑ 𝑊𝑖𝐹𝑖𝑖 =∑ 𝑊𝑖𝐹𝑖𝑖
∑ 𝑊𝑖𝑖 (39)
Input variables to the ANFIS controller are PEMFC voltage (VFC) and current (IFC) and the
output is duty cycle (K).
Input_1
Input_2
Output
VI.SIMULATION RESULTS
The performance of the proposed BLDC motor driven FCEV system is analyzed by using
the MATLAB/Simulink platform. To analyze the dynamic response of the FCEV system, sudden
changes in the temperature of the fuel cell is considered as follows: T= 300˚K for a period of 0 to
0.3sec, T= 340˚K for a period of 0.3 sec to 0.6 sec and T= 320˚K for a period of 0.6sec to 0.9 sec
as shown in Fig. 9. For the different temperatures the output current, voltage and power waveforms
of the fuel cell are as shown in Fig. 10. Fuel cell generates a power of 1080W for 0 to 0.3 sec,
970W for 0.3sec to 0.6 sec and 1220W for 0.6sec to 0.9sec.Fig. 11 shows The DC link current,
voltage and power by using the FLC base MPPT technique. It generates a power of 775W, 1140W
and 900W for the temperatures of 300˚K, 340˚K and 320˚K respectively.
The DC link output current, voltage and power using proposed RBFN based MPPT
controller are shown in the Fig. 12. The proposed controller gives 780W for the temperature of
300˚K, 1150W for 340˚K and 910W for the temperature of 320˚K. In Fig. 13, the performance of
the RBFN based MPPT controller for fuel cell is compared with fuzzy logic based MPPT
controller. From Fig. 13, it is observed that proposed controller generates the high DC link power
than the FLC. The comparative analysis of FLC and RBFN controllers are listed in Table 3. The
starting and steady-state characteristics of the BLDC motor at different temperatures of the fuel
Journal of Information and Computational Science
Volume 9 Issue 12 - 2019
ISSN: 1548-7741
www.joics.org704
cell are as shown in the Fig. 14. The motor parameters such as stator current (Isa), back EMF (E),
electromagnetic torque (Te) and load torque (TL) are presented at dynamic temperature conditions
of the fuel cell. The BLDC motor has a speed of 3300 rpm for 0 to 0.3sec, 2400 rpm for 0.3sec to
0.6sec and 3700 rpm for 0.6sec to o.9sec. The torque of the BLDC motor remains constant for
varying speed conditions.
The DC link output current, voltage and power using proposed ANFIS based MPPT
controller are shown in the Fig. 18. The proposed controller gives 790W for the temperature of
300˚K, 1160W for 340˚K and 920W for the temperature of 320˚K. In Fig. 17, the performance of
the ANFIS based MPPT controller for fuel cell is compared with fuzzy logic and RBFN based
MPPT controller. From Fig. 20, it is observed that proposed controller generates the high DC link
power than the RBFN and FLC. The comparative analysis of FLC, RBFN and ANFIS controllers
are listed in Table 2.
Fuzzy logic Simulation results:
Fig.9.Temperature changes in PEMFC
system.
Fig.10.Fuel cell output current, voltage and
power at different temperatures
Fig.11.DC link output current, voltage and
power at different temperatures using FLC.
Fig.12.BLDC parameters using FLC.
Journal of Information and Computational Science
Volume 9 Issue 12 - 2019
ISSN: 1548-7741
www.joics.org705
RBFN simulation results:
Fig.13.Fuel cell output current, voltage and
power at different temperatures using RBFN
Fig.14. DC link output current, voltage and
power at different temperatures using
RBFN.
Fig.15.BLDC parameters using RBFN Extension
Simulation results by using ANFIS Controller
Fig.16.Temperature changes in PEMFC
system.
Fig.17. Fuel cell output current, voltage and
power at different temperatures by using
ANFIS controller
Journal of Information and Computational Science
Volume 9 Issue 12 - 2019
ISSN: 1548-7741
www.joics.org706
Fig.18.DC link output current, voltage and
power at different temperatures using
ANFIS.
Fig.19.BLDC parameters using ANFIS
Fig.20.Comparison of power by using fuzzy controller, RBFN Controller, ANFIS controller
Table.2. Comparison of DC link power with both RBFN and Fuzzy based MPPT controllers
Parameter
Period (sec) 0 to 0.3 0.3 to0.6 0.6 to 0.9 0 to 0.3 0.3 to0.6 0.6 to 0.9 0 to 0.3 0.3 to0.6 0.6 to 0.9
Fuel cell temperature (˚K) 300 340 320 300 340 320 300 340 320
DC link power (W) 775 1140 900 780 1150 910 790 1160 920
Time taken to reach MPP (sec) 0.12 0.42 0.7 0.09 0.38 0.68 0.04 0.33 0.65
1.26 kW PEMFC with fuzzy 1.26 kW PEMFC with RBFN 1.26 kW PEMFC with ANFIS
VII. CONCLUSION
In this paper, a three-arrange excessive voltage gain DC-DC converter is proposed for
FCEV packages. The proposed converter has dwindled the energy section enter current swells and
the voltage weight on the power semiconductor switches. The ANFIS based MPPT has maximum
brilliant strength from the energy phase at extraordinary temperatures. The proposed MPPT
framework is differentiated and the FLC MPPT controller. The reenactment effects monitor that
the ANFIS based MPPT controller has pursued the maximum high-quality electricity point
snappier whilst diverged from the smooth technique of reasoning controller. Also, fantastic
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.90
500
1000
1500
2000
2500
3000
time(s)
pow
er(
w)
with fuzzy controller
with RBFN controller
with ANFIS controller
Journal of Information and Computational Science
Volume 9 Issue 12 - 2019
ISSN: 1548-7741
www.joics.org707
execution traits of the BLDC motor, as an instance, electromagnetic torque, speed and returned
EMF are broke down at exceptional temperatures of the energy unit system.
REFERENCES
[1] Penagaluru Suresh, and T. Gowri Manohar. "Review on Renewable Source Integrated
Topologies with Power Quality Enhancing Strategies." International Journal of Renewable
Energy Research, Vol. 8, No.4, 2018, Pp: 2350-2366.
[2] Sudheer, Kasa. "Control strategies for Power Quality Upgradation in Renewable Source
Connected Grid Systems." (2018).
[3] Ramasamy, Sudha, and Kasa Sudheer. "Enhancement of Power Quality in Renewable Source
Integrated Three Phase System with ICC Controller." Journal of Green Engineering 8, no. 1
(2018): 71-88.
[4] Suresh, P., & Manohar, T. G, “Effective Renewable Source Integration using Unified Power
Quality Conditioner with Power Quality Enhancement in Three Phase System” In MATEC Web
of Conferences EDP Sciences. Vol. 225, p. 05014 (2018).
[5] Kasa, Sudheer, and Sudha Ramasamy. "Mitigation of current harmonics in renewable
source integrated grid topology with fuzzy based dynamic shunt active filter." In 2017
Innovations in Power and Advanced Computing Technologies (i-PACT), pp. 1-6. IEEE, 2017.
[6] Kasa, Sudheer, Prabhu Ramanathan, Sudha Ramasamy, and D. P. Kothari. "Effective grid
interfaced renewable sources with power quality improvement using dynamic active power
filter." International Journal of Electrical Power & Energy Systems 82 (2016): 150-160.
[7] Kasa, Sudheer, and Sudha Ramasamy. "Mitigation of current harmonics in renewable source
integrated grid topology with fuzzy based dynamic shunt active filter." In 2017 Innovations in
Power and Advanced Computing Technologies (i-PACT), pp. 1-6. IEEE, 2017.
[8] Kasa, Sudheer, Prabhu Ramanathan, Sudha Ramasamy, and D. P. Kothari. "Effective grid
interfaced renewable sources with power quality improvement using dynamic active power
filter." International Journal of Electrical Power & Energy Systems 82 (2016): 150-160.
[9] Sudheer, Kasa, Duvvuru Madhurita, Amudala Chandana, Marella Thanesh, and M.
Karunakar Babu. "Intelligent Street Light System For Smart Cities.", International Journal of
Applied Science and Computations, Vol. 6, Issue 5,2019, Pp.684-690.
[10] Devi, S. Lakshmi, and K. Sudheer. "Neuro Fuzzy Control Single–Stage Single Phase AC-DC
Converter for High Power factor." , International Journal Of Engineering Research & Technology
Volume 02, Issue 09 (September 2013):3183-3194.
Journal of Information and Computational Science
Volume 9 Issue 12 - 2019
ISSN: 1548-7741
www.joics.org708
[11] RAMASAMY, SUDHA, and Kasa Sudheer. "Mitigating voltage imperfections with
photovoltaic fed ANFIS based ZSI-DVR in three phase system." International Journal of
Renewable Energy Research (IJRER) 7, no. 4 (2017): 2103-2110.
[12] Sudheer, Kasa, and Ramasamy Sudha. "Enhancement of Power Quality in Multi Feeder
Three Phase System with Photovoltaic fed ANFIS-Unified Multi Converter Controller."
In MATEC Web of Conferences, vol. 225, p. 03015. EDP Sciences, 2018.
[13] Sudheer, Kasa, Sarabu Kavitha, Panchavati Yashaswini, Reddycherla Aswini, and Ponnuru
Karthik. "POSLLC Based PV Fed D-STATCOM for PQ Improvement in Distribution System using
Neuro-Fuzzy Controller." International Journal of Applied Science and Computations, Vol. 6,
Issue 5,2019, Pp.691-698.
[14] Garca, P., Torreglosa, J.P., Fernndez, L.M., Jurado, F., 2013. Control strategies for high-
power electric vehicles powered by hydrogen FC, batteryand SC. Expert Syst. Appl. 40 (September
(12)), 4791–4804.
[15] Glavin, M.E., Chan, P.K.W., Armstrong, S., Hurley, W.G., 2008. A standalone photovoltaic
supercapacitor battery hybrid energy storage system.In: Proc. 13th Power Electron. Motion
Control Conf., September 1–3, pp. 1688–1695.
[16] Guiting, X., Yan, Z., Dakang, Z., 2013. Synthetically control of a hybrid PV/FC/SC power
system for standalone applications. J. Appl. Sci. 5 (5),1796–1803.
[17] Suresh, P., and T. Gowri Manohar. "A novel fuzzy based transformer less PV inverter for
reactive power control in grid connected PV system." In 2017 IEEE International Conference on
Power, Control, Signals and Instrumentation Engineering (ICPCSI), pp. 2463-2472. IEEE, 2017.
[18] Sudheer, Kasa, Sarabu Kavitha, Panchavati Yashaswini, Reddycherla Aswini, and Ponnuru
Karthik. "POSLLC Based PV Fed D-STATCOM for PQ Improvement in Distribution System
using Neuro-Fuzzy Controller." International Journal of Applied Science and Computations,
Vol. 6, Issue 5,2019, Pp.691-698.
Journal of Information and Computational Science
Volume 9 Issue 12 - 2019
ISSN: 1548-7741
www.joics.org709