Comparative Analysis of Fuzzy and Neural Controller For A Boost Converter

You might also like

Download as pptx, pdf, or txt
Download as pptx, pdf, or txt
You are on page 1of 24

Comparative Analysis of

Fuzzy and Neural


Controller for a Boost
Converter
By
Muhammad Waqas Qaisar, Hassan Mujtaba, Muhammad Tanveer Riaz, Muhammad
Shahid, Ahmad Abdul Ghani. Muhammad Abbas Khan, Kashif Hussain

Comparative Analysis of Fuzzy and Neural Controller for a


1
Boost Converter
Outline
Introduction & Background
Boost Converter
Neural Network Controller
Fuzzy Logic Controller
Proposed System
Results & Discussion
Conclusion
Future Work
Comparative Analysis of Fuzzy and Neural Controller for a Boost Converter 2
This research aims to advance the understanding and
application of fuzzy control, adaptive fuzzy logic, and
NNCs in the field of DC-DC converters, with broader
implications in electrical engineering and related
domains.
Introduction DC-DC converters maintain a constant output voltage
& despite load changes and unregulated input voltage.
Buck converters regulate output as DC with unregulated
Background DC input, while boost converters increase input voltage
for higher output.
Fuzzy control has advantages over linear P.I.D. and P.I.
controllers, offering superior performance across diverse
input voltage and load conditions.

Comparative Analysis of Fuzzy and Neural Controller for a Boost


3
Converter
Research focuses on designing and implementing fuzzy
logic controllers for DC-DC converters. Fuzzy control
enables compatibility with Simulink parameter settings.
Adaptive fuzzy logic controllers are being developed for
DC-DC converters, with studies exploring hybrid neuro-
Introduction fuzzy controls.
Neural network controls (NNC) are intelligent controls
& capable of modifying controller parameters, suitable for
Background nonlinear systems.
Non-negativity constraint (NNC) has been introduced
and successfully tested in simulated environments for
DC-DC converters.
NNCs find applications in drives, power electronics, and
academic performance improvement.

Comparative Analysis of Fuzzy and Neural Controller for a


4
Boost Converter
 A boost converter, also called a step-up converter,
increases the DC voltage input.
 The main components of a boost converter include a
transistor, a diode acting as a switch, and a battery.
 Capacitors with appropriate voltage and capacitance
Boost values are used to minimize output voltage ripple.
 When the switch is turned off, more current flows into
Converts the inductor.
 Current can only leave the inductor through the free-
wheeling diode connected in parallel with the output
when the switch is not activated.
 The capacitor can act as an energy transmitter when
triggered.

Comparative Analysis of Fuzzy and Neural Controller for a


5
Boost Converter
Boost Converts

Comparative Analysis of Fuzzy and Neural Controller for a


6
Boost Converter
 Artificial Neural Networks (ANNs) simulate brain
functions and can solve complex problems and
generating accurate models from complex natural
systems.
 ANNs consist of multiple artificial neurons, also known
Neural as processing elements (PE), with individual transfer
functions, input weights, and single outputs.
Network  ANNs are trained by identifying patterns and
Controller relationships in data rather than being explicitly
programmed.
 Feedforward and feedback propagation ANN models
are commonly used as they offer advantages such as
precision, processing speed, fault tolerance, validation,
and convergence.

Comparative Analysis of Fuzzy and Neural Controller for a


7
Boost Converter
Neural Network
Controller

Comparative Analysis of Fuzzy and Neural Controller for a


8
Boost Converter
 Fuzzy logic control utilizes a set of rules based on
human intuition for system behavior.
 MATLAB and Simulink are used to investigate the
dynamic behavior of a DC-DC converter.
 Fuzzy logic controllers can handle both insignificant
Fuzzy Logic and significant signals simultaneously, improving the
Controller converter's reliability.
 The buck converter's fuzzy controller accepts input
from two independent inputs: voltage difference and
sum of multiple errors.
 The fuzzy controller adjusts the duty cycle to alter the
voltage output.

Comparative Analysis of Fuzzy and Neural Controller for a


9
Boost Converter
 The fuzzification block includes a fuzzy knowledge-
based block, fuzzy inference engine, and defuzzification
block.
 Fuzzy logic components include fuzzy sets, membership
functions, linguistic variables, and fuzzy rules.
 A rule base contains if/then constructs to determine
Fuzzy Logic actions based on specific conditions.
 Seven variables, represented by linguistic values (N.B.,
Controller N.M., N.S., ZE, P.B., PM, P.S.), are used to determine
error and change in error.
 Fuzzification converts input variables into membership
values between 0 and 1 based on plant output and
reference comparison. The inference mechanism makes
decisions on potential outcomes, and a control signal is
sent to the plant

Comparative Analysis of Fuzzy and Neural Controller for a


10
Boost Converter
Fuzzy Logic The basic components of a fuzzy controller in block

Controller diagram form is presented here.

Comparative Analysis of Fuzzy and Neural Controller for a


11
Boost Converter
 The study aims to determine which nonlinear
controller, either fuzzy or neural network, works better
when the output load or boost converter characteristics
Proposed change.
 In this paper, a boost converter controller using both a
System fuzzy logic controller and a neural network controller.
 The fuzzy controller uses a standard set of rules, while
the neural network controller employs two hidden
layer networks.

Comparative Analysis of Fuzzy and Neural Controller for a Boost Converter 12


Proposed System

Comparative Analysis of Fuzzy and Neural Controller for a


13
Boost Converter
Simulation Results &
Discussion
 We used MATLAB R2021b for the
simulation of the boost converter
based on fuzzy logic and a neural
network.
 Figure 1, Figure 2, and Figure 3, the
examples of the membership
functions Error, Change in Error, and
Output using trapezoidal and
triangular functions with a gap of 0.5
and a range of [-1, 1] are present.

Comparative Analysis of Fuzzy and Neural Controller for a


14
Boost Converter
The fuzzy controller created in the boost converter’s
Simulation Results & Simulink model regulates the duty cycle of the MOSFET,
Discussion which is designed according to the converter’s parameters
listed in Table.

Comparative Analysis of Fuzzy and Neural Controller for a


15
Boost Converter
Simulation Results & To simulate the boost converter Simulink model,
we entered the desired table parameters into the
Discussion fuzzy controller.

Comparative Analysis of Fuzzy and Neural Controller for a


16
Boost Converter
This Figure depicts, via the output of the boost
Simulation Results & converter, how the fuzzy logic controller maintains
Discussion a constant output level for the system. At 0.12
seconds into the measurement, this occurs.

Comparative Analysis of Fuzzy and Neural Controller for a


17
Boost Converter
The figure demonstrates the use of the Neural Network
Simulation Results & Controller in Simulink to simulate the boost converter
model.
Discussion This figure depicts how to change the classifier’s settings,
such as the number of hidden units in a neural network.

Comparative Analysis of Fuzzy and Neural Controller for a


18
Boost Converter
The ANN’s linear regression model, which determines the
Simulation Results & relationship between the response and predictors are shown in
Figure, as opposed to previous slide Figure, which depicts how
Discussion to change the classifier’s settings, such as the number of hidden
units in a neural network.

Comparative Analysis of Fuzzy and Neural Controller for a


19
Boost Converter
This figure depicts the output of the boost converter after the
Simulation Results & 0.023-second settling time required by the neural network
controller to maintain the desired output level. The fuzzy
Discussion controller converges on the target value more slowly than the
neural network controller, which has a steady-state error.

Comparative Analysis of Fuzzy and Neural Controller for a


20
Boost Converter
Simulation Results & A neural network controller performs less effectively than a
fuzzy controller. The output is shown in pervious slide and the
Discussion results are available in in table which shows us the comparative
analysis of both the controllers.

Comparative Analysis of Fuzzy and Neural Controller for a


21
Boost Converter
Conclusion

 The suggested system utilizes a Fuzzy Logic and Neural Network controller for
managing boost converters.
 Fuzzy Logic and Neural Network controllers are easier to understand and install
compared to traditional P, PI, and PID Controllers as they do not require a
mathematical model of the converter.
 A shared set of rules can assist the Fuzzy Logic controller in delivering the desired
output value accurately.
 The Fuzzy Logic controller effectively controls the converter's input, despite having
a long settling time and no steady-state error.
 On the other hand, the Neural Network Controller exhibits a steady-state inaccuracy
when compared to the Fuzzy Logic controller.
Comparative Analysis of Fuzzy and Neural Controller for a
22
Boost Converter
 Our main goals for the fuzzy controller for the future work are:
 To increase rule base
 To Reduce settling time
 To Eliminate the steady-state error.
 Our main goals for the neural network controller for the future work
Future Work are:
 To increase accuracy through additional layers and different e
 To refine training data set
 To reduce steady-state error.
 Future we will do the Hardware-level implementation for both
controllers to achieve stability and faster performance.

Comparative Analysis of Fuzzy and Neural Controller for a Boost Converter 23


Thank You

Comparative Analysis of Fuzzy and Neural Controller for a Boost Converter 24

You might also like