TITLE
DESIGN AND IMPLEMENTATION OF
(MPPT) PV SYSTEM USING ARTIFICIAL
NEURAL NETWORK TECHNIQUES
DATE:30/7/2017
ARBAMINCH,ETHIOPIA
OUTLINE
INTRODUCTION
OBJECTIVE OF PROJECT
STATEMENT OF THE PROBLEM
LITERATURE REVIEW
SCOPE
METHODOLOGY
INTRODUCTION
Solar energy is a vital component of the global shift towards renewable resources,
offering a clean and sustainable solution to our energy needs.
In Africa, particularly in Ethiopia, the potential for solar power is immense, with annual
average solar radiation of approximately 5.5 kWh/m²/day in Arba Minch.
A significant portion of the rural population in Ethiopia still lacks access to modern
energy sources.
The existing photovoltaic (PV) systems in these areas often suffer from low efficiency
due to the absence of advanced maximum power point tracking (MPPT) controllers.
MPP(Maximum power point):is a point at which the product of voltage and
current gets a maximum value or a operating point at which maximum power
can be extracted from from the system.
MPPT(Maximum Power Point Tracking) is a technology used in solar
photovoltaic (PV) systems to optimize the power output from solar panels by
constantly adjusting the operating point to ensure they are always producing
the maximum possible power under varying environmental conditions like
sunlight intensity and temperature.
Our research addresses this challenge by proposing an Artificial Neural
Network (ANN)-based MPPT controller combined with a boost converter to
optimize power extraction from PV modules under varying solar conditions.
OBJECTIVE OF THE PROJECT
main objective
The primary objective of this project is to design and implementation of ANN
based MPPT to maximize power output from solar PV systems
Specific objective:
To create an ANN-based MPPT system with data collection, model training,
and integration with a DC-DC converter.
To evaluate the system's performance and compare it with conventional MPPT
methods.
To simulate the MPPT PV system using MATLAB Simulink for performance
evaluation.
STATEMENT OF THE PROBLEM
Traditional MPPT algorithms often suffer from slow tracking, oscillations
around the MPP, and inefficiency under rapid irradiance variations.
The conventional MPPT techniques struggle to maintain optimal power output
from Photovoltaic (PV) panels
This project aims to develop an ANN-based MPPT system to overcome these
limitations and improve energy conversion efficiency.
LITERATURE REVIEW
we have looked at many books so far out of them some are listed below;
[1] Design and Implementation of Solar based Maximum Power Point Tracking
using Machine Learning.(published 2023)
[2] Design and implementation of an ANN-based MPPT controller for solar-
powered BLDC motor systems.(published 4 April 2025)
[3] MPPT Algorithms based on Artificial Neural Networks for PV system under
partial shading condition.(published 2020)
SCOPE
This project aims to boost solar PV efficiency in Arba Minch, Ethiopia, using an ANN-
based MPPT controller to aid rural electrification.
METHODOLOGY
Block diagram of the work
G
Boost
Solar panel
converter Load
T
Neural Network
Function Of The Elements
1. Solar Panel (PV Module)
A solar panel is a device that captures sunlight and turns it into electricity.
2. Boost Converter
A boost converter is a tool that takes the electricity from the solar panel and
makes it stronger by increasing its voltage.
3. Neural Network (ANN Controller)
A neural network is a smart controller—like a brain for the system. It uses
artificial intelligence (AI) to make decisions and improve how the system
works.
FLOW CHART
START
Irradiance(G)
and Train ANN using G,T
Temperature( and Vmp
T)
if N<M Testing ANN process
NO
yes
Calculate Vmp Apply MPPT
N=N+1 End
Types of pv module
S.No parameters name values
1 module 1soltech 1STH-250WH
2 Pmax 250
3 Cell per module 36
4 Voc 37.3
5 Vmax 30.7
6 Vref 30
7 Isc 8.66
8 Imax 8.15
9 Irradiance(W/m2 1000
Mathematical calculation
𝑁𝑜𝑚𝑖𝑛𝑎𝑙𝑝𝑜𝑤𝑒𝑟 𝑓𝑟𝑜𝑚𝑝𝑣 𝑠𝑦𝑠𝑡𝑒𝑚
𝑁𝑡𝑜𝑡𝑎𝑙=
𝑁𝑜𝑚𝑖𝑛𝑎𝑙 𝑝𝑜𝑤𝑒𝑟 𝑜𝑓 𝑠𝑒𝑙𝑒𝑐𝑡𝑒𝑑𝑚𝑜𝑑𝑢𝑙𝑒
780 𝑊
=3 .12 ≅ 3
250 𝑊
𝑉𝑜𝑙𝑡𝑎𝑔𝑒 𝑜𝑓 𝑠𝑦𝑠𝑡𝑒𝑚
𝑁𝑠𝑒𝑟𝑖𝑒𝑠=
𝑂𝑝𝑒𝑛𝑐𝑖𝑟𝑐𝑢𝑖𝑡 𝑉𝑜𝑙𝑡𝑎𝑔𝑒 𝑜𝑓 𝑠𝑒𝑙𝑒𝑐𝑡𝑒𝑑𝑚𝑜𝑑𝑢𝑙𝑒
=
𝑉𝑠𝑡𝑟𝑖𝑛𝑔=𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑠𝑒𝑟𝑖𝑒𝑠𝑚𝑜𝑑𝑢𝑙𝑒 ∗𝑉𝑎𝑚𝑝
𝑉𝑠𝑡𝑟𝑖𝑛𝑔=3∗30.7
𝑉𝑠𝑡𝑟𝑖𝑛𝑔=98 .77 𝑣
𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑡𝑜𝑡𝑎𝑙𝑚𝑜𝑑𝑢𝑙𝑒
𝑁𝑠𝑡𝑟𝑖𝑛𝑔=
𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑠𝑒𝑟𝑖𝑒𝑠𝑚𝑜𝑑𝑢𝑙𝑒
3
𝑁𝑠𝑡𝑟𝑖𝑛𝑔 =
3
𝑁𝑠𝑡𝑟𝑖𝑛𝑔=1
module
𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑡𝑜𝑡𝑎𝑙𝑚𝑜𝑑𝑢𝑙𝑒=1∗3
𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑡𝑜𝑡𝑎𝑙𝑚𝑜𝑑𝑢𝑙𝑒=𝟑𝒎𝒐𝒅𝒖𝒍𝒆𝒔
Design of boost converter
Duty cycle (D)
𝑉𝑖𝑛
𝐷=100 % ∗( 1− )
𝑉𝑜𝑢𝑡
Input current(Iin) 𝑃𝑜𝑢𝑡
𝐼𝑖𝑛=
𝑉𝑝𝑣
𝐼𝑖𝑛= ❑
❑
Then,the current ripple ,which is 10% of input current,is calculated as follow;
10
∆ 𝐼= 𝐼𝑖𝑛
100
The other is ripple voltage,which is 1% of out put voltage is calculated as
follow;
1
∆𝑉 = 𝑉𝑜𝑢𝑡
100
Out put current(Iout)
𝑃𝑟𝑎𝑡𝑒𝑑
𝐼𝑜𝑢𝑡=
𝑉𝑜𝑢𝑡
The most important parameters we are finding for is the following two
parameters.
Inductor and
Capacitor
Inductor is used to store magnetic energy.
Capacitor is used to store electrical energy.
In order to design the DC-DC boost converter knowing the value of those
two parameters is necessary things
The value of those two parameters is calculated as follow;
𝑉𝑖𝑛∗(𝑉𝑜𝑢𝑡 −𝑉𝑖𝑛)
𝑳=
𝐹𝑠𝑤 ∗ ∆ 𝐼 ∗𝑉𝑜𝑢𝑡
¿
𝐼𝑜𝑢𝑡 ∗(𝑉𝑜𝑢𝑡 −𝑉𝑖𝑛)
𝑪=
𝐹𝑠𝑤 ∗ ∆𝑉 ∗𝑉𝑜𝑢𝑡
=
The Schematic Diagram of an ANN
Input Layers Hidden Layers Output Layer
Irradiance(G)
Duty cycle(D)
Temperature(T)
Figure 2: schematic diagram of ANN
Input Layer: Think of this as the "ears" of the system. It listens to the information
coming in like sensor data or user commands and passes it along.
Hidden Layers: These are the "brain" of the ANN. They process the input by working
together like a team of problem-solvers, figuring out patterns or solutions behind the
scenes.
Output Layer: This is the "mouth" of the system. It takes what the hidden layers
figured out and delivers the final result like a command to move a robot or adjust a
setting.
!!!
O U
Y
N K
H A
T
