2020 2nd International Conference on Sustainable Technologies for

Industry 4.0 (STI), 19-20 December, Dhaka

Comprehensive Analysis of MPPT Techniques

using Boost Converter for Solar PV System
Redwan Ahmed, Shadhon Chandra Mohonta
Department of Electrical and Electronic Engineering, Rajshahi University of Engineering & Technology, Rajshahi, Bangladesh
2020 2nd International Conference on Sustainable Technologies for Industry 4.0 (STI) | 978-1-6654-0489-1/20/$31.00 ©2020 IEEE | DOI: 10.1109/STI50764.2020.9350398

Email: [email protected], [email protected]

Abstract—An efficient maximum power point tracking By comparing instantaneous and incremental conductance of
(MPPT) controller is a crucial part of solar photovoltaic (PV) a PV array, a new MPPT algorithm is developed and
system, which can handle the non-linear characteristics of a discussed [6]. The software flowcharts of two methods
solar PV array. In this study, a proposed solar PV system with (IncCond and P&O) are displayed and their advantages and
boost converter and dc load is modeled and compared disadvantages are discussed in detail [7]. An MPPT
according to the operating characteristics (i.e. voltage, current, technique based on open circuit voltage is proposed for solar
power and efficiency under varying solar irradiance and cell systems to reduce sampling period and sampling interval of
temperature) using five conventional and modern MPPT the PV array voltage [8]. A PV system adopting FOCV
techniques. The proposed solar PV system and MPPT
MPPT technique is proposed which can be used in IOT
controllers are modeled and simulated in MATLAB Simulink
platform. Among these five MPPT controllers, artificial neural
based applications [9]. The performances of three types of
network (ANN) MPPT controller provides the highest efficiency MPPT control algorithms (P&O, IncCond and FSCC) are
of 97.55% and produces less voltage and power fluctuations. numerically discussed under steady-state and dynamic
The novelty of this paper is that it focuses on the key atmospheric conditions [10]. A model of ANN MPPT
characteristics and simulated results of the five MPPT method is proposed based on the architecture multi-layer
techniques to make a comparison between them. perception (MLP) whose training is done by practical data
[11]. A residential solar PV system is designed using ANN
Keywords— solar photovoltaic (PV) system, maximum power MPPT controller for both MPP tracking and inverter control
point tracking (MPPT), boost converter, MATLAB hence performance of the system is evaluated through digital
simulation [12].
I. INTRODUCTION
Energy is a vital part of our daily life. Nowadays, In this paper, different components of a solar PV system
renewable energy is a great concern to the energy specialists and operating characteristics of five MPPT controllers are
because the sources of this energy are inexhaustible. As the discussed. A proposed solar PV system with five MPPT
existed energy sources (i.e. coal, natural gas, oil) are techniques at varying irradiance and temperature conditions
depleting day by day, it is important to find suitable is also modeled and simulated. Moreover, the performances
renewable energy sources. In the year 1990, only 1% of of these models are analyzed and compared according to the
electrical energy was produced from renewable energy key parameters. This paper offers a design of Artificial
sources and became 5% in the year 2012. Only 8 years later Intelligence (AI) based MPPT controller which will aid
it has become 10%. It is estimated that it will reach at 17% in Industry 4.0.
2040 [1]. Solar energy is a clean, easily available and II. SOLAR PHOTOVOLTAIC SYSTEM
imperishable renewable energy source. If solar power can be
converted to electrical power efficiently, then the A. PV Modeling
requirement of energy can be fulfilled where grid power is A semiconductor device which converts light energy into
less than the demand or unavailable. In recent years, electrical energy is referred to as the photovoltaic cell or
necessity of solar power is increasing more rapidly for photoelectric cell. An assembly of photovoltaic cells are
remote power applications. Power generation based on solar mounted in a framework in a PV module. A PV panel is
PV system plays an important role in Industry 4.0 revolution. constructed with a number of PV modules. An array is a
Digitalized, efficient and cost effective PV system will set up collection of panels and can be modeled with parallel and
smart systems and accelerate Industry 4.0 revolution. series modules. A solar cell can be represented by a parallel
Maximum power point tracking (MPPT) controllers have connection of a current source and a p-n diode,
been integrated with solar photovoltaic (PV) array to get Rs Ipv
maximum output power with optimum efficiency. 19
methods have been discussed to track maximum power point
(MPP) for PV arrays and a comparison is made between ID Ish
them [2]. A survey of 30 different MPPT techniques are
performed based on their operating characteristics,
Iph Rsh Vpv
advantages and disadvantages for PV power systems [3]. A
brief overview of P&O and Hill Climbing MPPT techniques
is demonstrated hence simulated outputs are used to make a
comparison between them [4]. A PV system with P&O
algorithm is proposed which feeds a dc series motor then
transient and steady-state analysis of the motor is done [5]. Fig. 1. Equivalent model of solar cell [13]

978-1-6654-0489-1/20/$31.00©20XX IEEE

Authorized licensed use limited to: Univ of Calif Santa Barbara. Downloaded on May 19,2021 at 00:49:46 UTC from IEEE Xplore. Restrictions apply.
 III. MAXIMUM POWER POINT TRACKING TECHNIQUES
A. MPPT Technique
A less amount of solar irradiation can be converted into
electrical energy by a typical solar panel. MPPT controller
can improve the efficiency of solar panel using maximum
power transfer technique (i.e. Thevenin’s theorem).
It states that when the Thevenin impedance of the
converter circuit equals the load impedance, output power of
the array will be maximum. Some popular and effective
MPPT techniques are:
Fig. 2. I-V and P-V characteristics of solar PV cell [3]
1. Perturbation and Observation (P&O)/ Hill Climbing
as shown in Fig. 1. Current through the diode is given by (1) 2. Incremental Conductance (IncCond)
[13]. 3. Fractional Open Circuit Voltage (FOCV)
4. Fractional Short Circuit Current (FSCC)
 q (Vpv + Ipv Rs )
ID = I0  e  


k B T  — 1 (1) 5. Artificial Neural Network (ANN)
 
B. P&O/ Hill Climbing MPPT Technique
While, the output current of the cell,
P&O and Hill Climbing are one of the earliest & simplest
Ipv = Iph — ID — Ish (2) MPPT techniques. These are easily implemented and cost
effective. Still this techniques are used by the most
So, commercial systems.
 q (Vpv + Ipv Rs )
Ipv = Iph — I0  e   Vpv + Ipv Rs
 There is a slight difference between these two techniques.

k B T  — 1 — (3)
  Rsh The operating voltage of the PV array is perturbed in P&O
technique, while the duty ratio of the power converter is
Where, Iph is the maximum short circuit current of the solar perturbed in Hill Climbing technique [4]. In P&O algorithm,
cell, q is the charge of electron, I0 is the reverse saturation the control parameter is periodically perturbed (i.e.
current of a p-n junction, T is the temperature in Kelvin scale decrementing or incrementing) according to the difference of
and kB is the Boltzman constant. PV output power of that cycle and previous perturbation
Fig. 2 shows non-linear characteristics (i.e. I-V and P-V) cycle as in Fig. 5(a). The subsequent perturbation for the
of a solar PV cell. A certain point on the P-V curve shows control parameter will be retained in the same direction if
maximum PV output power. This is the maximum power there is an increase in the PV output power while reaching
point (MPP). Solar PV array should be driven at MPP the MPP. If there is a decrease in the PV output power the
voltage (VMPP) to produce maximum PV output power. perturbation will be reversed. System waveforms oscillate
around the MPP due to the continuous perturbations of the
B. Boost Converter control parameter at every cycle of MPPT though the solar
Boost converter is a switching mode regulator that steps irradiance and cell temperature are constant or slowly
up voltage from its input to its output [14]. Dc-dc boost varying. This oscillation causes PV power loss. By
converter is used to adjust the supplied voltage to load. It decreasing the step size of the control parameter this problem
also allows the array to perform at its MPP by determining can be solved but this process will increase the tracking time.
the duty cycle [13]. Fig. 3 shows a boost converter is a When the atmospheric conditions are changed rapidly, P&O
combination of capacitor, inductor, diode and switch. algorithm deviates from MPP.
Characteristics equation of a boost converter is shown by the C. Incremental Conductance MPPT Technique
following (4) [14].
IncCond is one of the most reliable and convenient
1 MPPT techniques which is used enormously in commercial
Vo = Vi (4)
1— D applications. This method can avoid the P&O algorithm
drawbacks. IncCond algorithm is based on the fact that
Where, D is the duty cycle, Vi is the input voltage and Vo is according to the difference of array terminal voltage and
the output voltage of the boost converter. MPP voltage, the terminal voltage will be adjusted at every
cycle of MPPT [6].
Inductor Diode

Resistor
Dc Voltage
Switch
Source
Capacitor

Fig. 3. Boost converter circuit [14] Fig. 4. Characteristics of PV array power curve [4]

Authorized licensed use limited to: Univ of Calif Santa Barbara. Downloaded on May 19,2021 at 00:49:46 UTC from IEEE Xplore. Restrictions apply.
 Start
Start Hidden
Sample V(n), I(n) layer
P(n) = V(n) * I(n) Sample V(n), I(n)
dP = P(n) - P(n-1) ∆V = V(n) - V(n-1)
dV = V(n) - V(n-1) ∆I = I(n) - I(n-1) Input 1
layer W11
Yes W21 Output
dP = 0 No Yes Irradiance
∆V = 0 (G)
W12 b21
layer
No 1
Yes Yes W13
∆I/∆V = - I/V ∆I = 0 VMPP
Yes No W22
dP > 0 2 1
No No W14
No dV > 0 No Temperature
dV > 0 Yes
∆I > 0 (T) b22
W2n b31
Yes Yes No 2 W15
∆I/∆V > - I/V
W1(2n)
For Hill Climbing For Hill Climbing No
D(n) = D(n-1) - D(step) D(n) = D(n-1) + D(step) Yes
For P&O For P&O n
V(n) = V(n-1) + V(step) V(n) = V(n-1) - V(step)
V(n) = V(n-1) + V(step) V(n) = V(n-1) - V(step)
b2n

P(n-1) = P(n) V(n-1) = V(n)

V(n-1) = V(n) I(n-1) = I(n) Fig. 7. Structure of a neural network [11]
Return Return
It is a voltage based MPPT technique. The basic idea of
(a) Algorithm for P&O technique (b) Algorithm for IncCond technique this technique is that the PV array voltage shows a linearity
with the array open circuit voltage at maximum power point
Fig. 5. Algorithm for P&O and IncCond MPPT technique [5], [6] for different irradiation and temperature levels [8]. This
The slope of the PV array power curve at the MPP is relationship can be described by the following (7) [9].
zero, positive on the left, and negative on the right of the VMPP = k 1 × VOC (7)
MPP [3], as shown in Fig. 4. Since slope of P-V curve,
Where, VMPP is the maximum power point voltage, VOC is the
dP d ( IV ) dI ∆I open circuit voltage of the PV array, k1 is the voltage factor.
= = I +V ≈ I +V (5)
dV dV dV ∆V This equation (7) looks very simple but it is difficult to
dP determine the accurate value of k1. Since, k1 relies on the PV
At MPP, = 0 . So, (5) can be written as, array characteristics, it is calculated previously by
dV
empirically determining VMPP and VOC at different irradiance
∆I I and temperature levels for a specific PV array [2]. Generally
=— , at MPP the value of k1 varies between 0.7 and 0.8 [9]. By shutting
∆V V
down the boost converter instantly, VOC is measured. This
∆I I process is performed periodically and as a result some
>— , left of MPP (6)
∆V V problems are occurred, including temporary loss of power.
Pilot cells can be a solution of this problem. Photovoltaic
∆I I cells having equivalent characteristics as those in the array,
<— , right of MPP
∆V V but having electrical separation from the array is referred to
as pilot cells. These cells should be chosen carefully so that
So, the array will be driven at the maximum power point
these cells can represent same characteristics as in main PV
when to the negative of the output conductance is equal the
array. The PV array actually never operates at the MPP,
variance of the output conductance [7]. According to
since (7) is only an approximation.
IncCond algorithm as in Fig. 5(b), PV array voltage (VPV)
equals to MPP voltage (VMPP) at MPP. So, this algorithm E. Fractional Short Circuit Current MPPT Technique
continuously decrements or increments VPV to reach VMPP. FSCC is a swift but an unsophisticated technique of
Step size determines the speed of IncCond MPPT controller. tracking the maximum power point and the algorithm is
Larger step size enables fast tracking but the array may not shown in Fig. 6(b). It is a current based MPPT technique.
be driven at the MPP and there will be an oscillation around The basic idea of this technique is that the PV array current
MPP. So, a trade-off is required between speed and accuracy. shows a linearity with the array short circuit current at
D. Fractional Open Circuit Voltage MPPT Technique maximum power point for different irradiation and
temperature levels [10]. This relationship can be described
FOCV technique is the simplest MPPT technique and the
by the (8) [10].
algorithm is in Fig. 6(a).
Start Start
IMPP = k 2 × ISC (8)
Measure Array Voltage, Vpv Measure Array Current, Ipv Where, IMPP is the maximum power point current, ISC is the
Isolate Array from MPPT Isolate Array from MPPT

Record Open Circuit Voltage, Voc Record Short Circuit Current, Isc

Vref = k1 * Voc Iref = k2 * Isc

Wait x Sec Wait x Sec

Vpv = Vref Ipv = Iref
Yes Yes

No No
E

Vpv > Vref Ipv > Iref

Yes No Yes No

Decrease Vpv Increase Vpv Decrease Ipv Increase Ipv

(a) Algorithm for FOCV technique (b) Algorithm for FSCC technique

Fig. 6. Algorithm for FOCV & FSCC MPPT techniques [8], [10] Fig. 8. MATLAB Simulink model for the proposed PV system

Authorized licensed use limited to: Univ of Calif Santa Barbara. Downloaded on May 19,2021 at 00:49:46 UTC from IEEE Xplore. Restrictions apply.
Fig. 9. Irradiance and temperature signal

short circuit current of the PV array, k2 is the current factor. (a) Structure of neural network (b) Training system

This equation (8) looks very simple but it is difficult to Fig. 10. Structure and training system of the proposed neural network
determine the accurate value of k2. Since, k2 relies on the PV
array characteristics, it is calculated previously by process. This is called feed-forwarding. Training procedure
empirically determining IMPP and ISC at different irradiance of a neural network is given below:
and temperature levels for a specific PV array [2]. Generally
the value of k2 varies between 0.78 and 0.92 [10]. By Step 1: Initialize the weights.
shutting down the boost converter instantly and shorting the Step 2: Calculate the error.
array, ISC is measured which causes some temporary loss of
power. Pilot cells can be used to reduce the power loss. If j and n are the node number and total number of nodes
There are difficulties in measuring ISC at the time of in the previous layer, then output of i-th node of p-th layer
operation of the system. It requires an additional switch. The can be calculated using (9) [11].
PV array actually never operates at the MPP, since (8) is
 n
only an approximation. y ip = (
g   x j W ijp + bip ) (9)
F. Artificial Neural Network MPPT Technique  j =1 
Fig. 7 shows the structure of Artificial Neural Network
(ANN), which is a modern MPPT technique with high Where, y ip is the output variable, x j is the input variable,
convergence speed. This method is gaining popularity day by W ijp is the synaptic weight, bip is the bias and g is the
day because the maximum power point is accurately tracked
activation function. Levenberg-Marquardt equation for error
by it. A neural network imitates biological neural network
calculation is given in (10) [11].
(i.e. human brain). Human brain observes the surrounding
events and learns from experiences. A neural network stores
information in patterns. The processing unit is called neuron e=
1 n p
(
 y k — g (x k ,W )
N k =1
2
) (10)
in an ANN. Adjustable weights are taken to connect neurons.
Biases are also added to the neurons. Signals are passed Where, e is the mean square error (MSE), y kp is the
neuron to neuron. Activation function for a neuron
determines a threshold value for the neuron to activate. There measured output, g(xk,W) is the model estimated output or
are three layers in a layered neural network which are input, target and N is the number of training data.
hidden, and output layers. Trial and error method can be used Step 3: Calculate the weight updates.
to determine the number of hidden layers [12]. In each layer,
the number of nodes are user dependent. Input of the ANN The back-propagation algorithm distributes the error(e)
MPPT controller can be atmospheric data like temperature towards the input hidden neurons to calculate the weight
(T) and irradiance (G), or array parameters like short circuit updates , using the following (11) [11].
current (ISC) and open circuit voltage (VOC), or any
∂e
combination of these. Several reference signals like duty ∆ W n = δ ∆ W n −1 — η (11)
cycle or reference voltage are the output of the controller. ∂W
Neural networks work in two steps. At first, the network Where, ∆Wn is the change of weight for iteration (n), ∆Wn-1 is
has to be trained using some data sets. Backpropagation is an the change of weight for iteration (n-1), δ is a term of speed
algorithm to train (adjust weights) a neural network. When and η is the the training rate.
the network is trained, it is ready to perform for similar types Step 4: Adjust the weight updates.
of data inputs that have been used in training
Step 5: Repeat step 2 to 4 until the error reaches an
TABLE I. COMPONENTS OF THE BOOST CONVERTER acceptable level.
Parameters Value When the training process is completed, the network can be
Load resistance, RLoad 20 Ω used for tracking the maximum power point.
Inductance, L 2 mH
Input capacitance, C1 100 μF IV. MODELING OF PV SYSTEM WITH MPPT CONTROLLER
Output capacitance, C2 100 μF
Resistance in series with inductance, R2 0.1 Ω A proposed PV system is modeled in MATLAB Simulink
Resistance in series with input capacitance, R1 0.1 mΩ platform as in Fig. 8. The proposed PV system is formed
Resistance in series with output capacitance, R3 0.1 mΩ with PV array, boost converter, MPPT controller and dc
Switching frequency, fs 200 kHz load.

Authorized licensed use limited to: Univ of Calif Santa Barbara. Downloaded on May 19,2021 at 00:49:46 UTC from IEEE Xplore. Restrictions apply.
Fig. 11. Output voltage with various MPPT controller
Fig. 13. PV power with various MPPT controller

The solar PV module used in this experiment is 1Soltech

1STH-215-P. The PV array is comprised of two parallel
strings and two series connected modules per string. The
irradiance and temperature signals are generated with signal
builder block as in Fig. 9. Table I shows the value of
different components of the boost converter. Without MPPT
controller, initial duty cycle for the boost converter is
assumed 55% which remains constant at any atmospheric
conditions. For P&O and IncCond algorithm, initial duty
cycle is assumed 55%. Step value for P&O is 0.005. Step and
threshold value for IncCond are 0.006 and 0.001
respectively. For FOCV algorithm, the voltage factor is
assumed 0.8. Current factor is assumed 0.85 for FSCC Fig. 14. Output power with various MPPT controller
algorithm. Pilot cells are used for measuring VOC and ISC.
ANN algorithm is implemented by MATLAB nftool toolbox Mean value of PV power and output power are 548.9W
as in Fig. 10(a). The proposed neural network has 3 layers. and 540.2W which are very good. In the PV system with
Input, hidden and output layers contain 2, 10 and 1 nodes FOCV MPPT controller shows that there are large number of
respectively. 295 input and output data points are taken for ripples in voltage, current and power curves due to false
the training process at different irradiation and temperature tracking of the MPP. Mean value of PV power and output
as in Fig. 10(b). The training is done by Levenberg- power are 522.5W and 513.7W which are not satisfactory at
Marquardt training algorithm. The training process takes all. Also, there are large number of ripples in voltage, current
1000 epochs to train the neural network and gives the best and power curves due to false tracking of the MPP in PV
performance. Mean square error for training, validation and system with FSCC MPPT controller. Mean value of PV
testing are 0.0000184, 0.000057954 and 0.0000732 power and output power are 514.6W and 507.4W which are
respectively. Since MSE is so small, the network is well not good. PV system with ANN MPPT controller produces
trained. very small amount of ripples in voltage, current and power
curves. Mean value of PV power and output power are
V. SIMULATION & ANALYSIS 552.1W and 543.1W which are greater than any other MPPT
Fig. 11-14 shows comparison between MPPT controllers controller discussed in this paper. The overall simulated
with respect to output voltage, output current, PV power and performances of all MPPT techniques are enlisted in Table
output power respectively. PV system without MPPT II.
controller possesses serious ripples in voltage, current and
TABLE II. PERFORMANCE COMPARISON OF MPPT TECHNIQUES
power curves. Mean value of PV power and output power
are 442W and 432.8W which are very poor. PV system with Paramet No
P&O
Inc
FOCV FSCC ANN
P&O MPPT controller produces considerable ripples in er MPPT Cond
voltage, current and power curves. Mean value of PV power Mean PV
voltage 40.31 57.4 57.59 55 59.59 57.61
and output power are 545W and 536.3W which are good. PV (V)
system with IncCond MPPT controller shows that there are Mean
less number of ripples in voltage, current and power curves. output
93.82 101.6 102 99.75 99.11 102.4
voltage
(V)
Mean PV
power 442 545 548.9 522.5 514.6 552.1
(W)
Mean
output
432.8 536.3 540.2 513.7 507.4 543.1
power
(W)
Efficieny
73.2 96.82 97.21 94.14 93.07 97.55
(%)
Efficieny
improved 32.26 32.80 28.61 27.14 33.27
Fig. 12. Output current with various MPPT controller (%)

Authorized licensed use limited to: Univ of Calif Santa Barbara. Downloaded on May 19,2021 at 00:49:46 UTC from IEEE Xplore. Restrictions apply.
 Though the mean PV voltage of ANN is less than Power Point Tracking Circuit," IEEE Transactions on Circuits and
FSCC, the mean output voltage, PV power and output Systems II: Express Briefs, vol. 66, no. 2, pp. 257-261, Feb. 2019.
power are higher than other techniques. The efficiencies [10] M. B. Danoune, A. Djafour, A. Gougui, N. Khelfaoui and H.
Boutelli, "Study and Performance Analysis of Three Conventional
without MPPT, with FSCC, with FOCV, with IncCond, MPPT Algorithms Used in Photovoltaic Applications," in 2018
P&O and with ANN MPPT controller are 73.2, 96.82, International Conference on Communications and Electrical
97.21, 94.14, 93.07, 97.55% respectively. It is evident that Engineering (ICCEE), El Oued, Algeria, December 17-18, 2018,
P&O, IncCond and ANN MPPT controllers increase the pp. 1-7.
performance of the solar systam than the rest of MPPT [11] S. Farhat, R. Alaoui, A. Kahaji and L. Bouhouch, "Estimating the
controllers. Moreover, IncCond MPPT controller performs photovoltaic MPPT by artificial neural network," in 2013
International Renewable and Sustainable Energy Conference
better than P&O MPPT controller. But, the ANN MPPT (IRSEC), Ouarzazate, March 7-9, 2013, pp. 49-53.
controller provides the highest efficiency of 97.55% and [12] J. Lakshmi P.N. and M. R. Sindhu, “Residential Solar Photovoltaic
performs better than other MPPT controllers. System With Artificial Neural Network Based Controller,” in 2018
International Conference on Control, Power, Communication and
VI. CONCLUSION Computing Technologies (ICCPCCT), Kannur, March 23-24, 2018,
In this study, a solar PV system without and with pp. 416-420.
MPPT controllers is simulated and investigated. A [13] Y. Chaibi, A. Allouhi, M. Salhi and A. El-jouni, “Annual
performance analysis of different maximum power point tracking
comparative analysis shows that MPPT techniques have techniques used in photovoltaic systems,” Springer Transactions
increased the performance of the proposed PV system. on Prot Control Mod Power Syst, vol. 4, no. 15, Aug. 2019.
Conventional MPPT controllers (i.e. P&O, IncCond, [14] Muhammad H. Rashid, Power Electronics Circuits, Devices, And
FOCV and FSCC) are simple and easy to be implemented Applications , 3rd ed. Pearson Education Inc., 2004.
but cannot perform well under rapid changing
atmospheric conditions. On the contrary, ANN MPPT
controller can track MPP accurately at any atmospheric
conditions. It is obvious that efficiency, PV power and
output power for the solar PV system with ANN MPPT
controller is larger than any other controllers. Moreover,
with the ANN MPPT controller there are less fluctuations
in output characteristics, which makes this technique more
suitable for the proposed PV system. This paper can
function as a suitable reference for future work in the field
of Artificial Intelligence based PV power generation.
REFERENCES
[1] (accessed April 9, 2020) World Electricity Generation. [Online].
Available:https://www.researchgate.net/figure/WORLD-
ELECTRICITY-GENERATION-1990-2040-SOURCE-
IEA_fig1_297684038.
[2] T. Esram and P. L. Chapman, "Comparison of Photovoltaic Array
Maximum Power Point Tracking Techniques," IEEE Transactions
on Energy Conversion, vol. 22, no. 2, pp. 439-449, Jun. 2007.
[3] A. N. A. Ali, M. H. Saied, M. Z. Mostafa, and T. M. Abdel-
Moneim, "A survey of maximum PPT techniques of PV
systems," in 2012 IEEE Energytech, Cleveland, OH, May 29-31,
2012, pp. 1-17.
[4] Fangrui Liu, Yong Kang, Yu Zhang, and Shanxu Duan,
"Comparison of P&O and hill climbing MPPT methods for grid-
connected PV converter," in 2008 3rd IEEE Conference on
Industrial Electronics and Applications, Singapore, June 3-5, 2008,
pp. 804-807.
[5] T. O. Sweidan and M. S. Widyan, "Perturbation and observation as
MPPT algorithm applied on the transient analysis of PV-powered
DC series motor," in 2017 8th International Renewable Energy
Congress (IREC), Amman, March 21-23, 2017, pp. 1-6.
[6] K. H. Hussein, I. Muta, T. Hoshino, and M. Osakada, "Maximum
photovoltaic power tracking: an algorithm for rapidly changing
atmospheric conditions," IEE Proceedings - Generation,
Transmission and Distribution, vol. 142, no. 1, pp. 59-64, Jan.
1995.
[7] S. Qin, M. Wang, T. Chen, and X. Yao, "Comparative analysis of
incremental conductance and perturb-and-observation methods to
implement MPPT in photovoltaic system," 2011 International
Conference on Electrical and Control Engineering, Yichang,
September 16-18, 2011, pp. 5792-5795.
[8] J. Ahmad, "A fractional open circuit voltage based maximum
power point tracker for photovoltaic arrays," in 2010 2nd
International Conference on Software Technology and
Engineering, San Juan, PR, October 3-5, 2010, pp. V1-247-V1-
250.
[9] T. Hsu, H. Wu, D. Tsai and C. Wei, "Photovoltaic Energy
Harvester With Fractional Open-Circuit Voltage Based Maximum

Authorized licensed use limited to: Univ of Calif Santa Barbara. Downloaded on May 19,2021 at 00:49:46 UTC from IEEE Xplore. Restrictions apply.