NOVATEUR PUBLICATIONS

International Journal of Research Publications in Engineering and Technology [IJRPET]

ISSN: 2454-7875
VOLUME 3, ISSUE 6, Jun.-2017
MAXIMUM POWER POINT TRACKING CONTROL FOR PHOTOVOLTAIC
SYSTEMS: NEURO-FUZZY APPROACH
MR. SWAPNIL L. DHAGE
M.E. Student, G. H. Raisoni Institute of Engineering and Technology, Pune Email ID: [email protected]

MRS. SANGITA B. PATIL

Assistant Professor, G. H. Raisoni Institute of Engineering and Technology, Pune Email ID: [email protected]

ABSTRACT: algorithms methods that may not be appropriate for

The proper function of photovoltaic (PV) systems realistic implementations. It creates use of a thinking that is
needs the design of an maximum power point tracking neither actual nor definitely inexact, but only to a certain
system (MPPT) to draw out the maximum possible level actual or inexact. The controller design is simple,
power from the photovoltaic array. This paper mainly consisting of the conversion of a linguistic rules set
investigates the efficiency of a neuro-fuzzy logic into an automatic control algorithm. It does not require
controller over conventional ones intended to track precise data, its reasoning schemes being based on
the maximum power point (MPP). The proposed model uncertain or imprecise information. The output control is a
used for simulation studies is implemented using smooth control function despite a variety of input
Matlab/Simulink. A comparison between the classical variations. The fuzzy logic controller processes user defined
Perturb and Observe (P&O) algorithm and fuzzy rules that can be easily modified, to improve the system
algorithm in terms of MPPT accuracy is provided. performance [2-5].
Results prove that the proposed model is simple, Moreover, both Fuzzy Logic (FL) and Neural Network
reliable and allows simulation under different (NN) controls have been preferred for the MPPT of the PV
operating conditions. system over the last several years. The FL MPPT controller
KEYWORDS: photovoltaic; modeling; simulation; is one of the most promising control schemes for the
neuro-fuzzy logic; MPPT. unpredictable PV system, but it requires a priori knowledge
of the system input/output relationship [28]. Similarly, the
I. INTRODUCTION: NN enhances the efficiency of the system by adopting a
Solar output power of PV systems is generally multilayer structure; though, each kind of PV array has to
influenced by atmospheric factors like variable be periodically trained to formulate the control rules;
temperature and solar irradiation. Different control therefore, its limitation is versatility [29]. The shortcomings
techniques could be used on get rid of the miss functioning: of NN and FL are overwhelmed by hybridizing NN and FL in
controlling the input to the PV array and controlling the the Neuro-Fuzzy Controller (NFC) [24,30–32]. NFC
power output from the PV array. The combinations of these combines the explicit knowledge of FC, which is understood
two may also be considered for maintaining constant with the implicit knowledge of NN, which is acquired by
output power at load [1]. The irradiance dependent input to learning. The hybrid neuro-fuzzy is becoming the most
the PV systems is controlled in order to be kept as high as preferred choice for tracking the MPP of PV over the last
possible, no matter the changes in weather conditions. Due few decades. Nevertheless, the neuro-fuzzy system have
to nonlinear I-V and P-V characteristics of the PV array, the problems of getting trapped in local minima of the search
MPPT becomes more challenging. Such non-linear and non- space and low convergence speed [33]. In recent years,
minimum phase characteristics further confuse the MPPT several researchers have developed different techniques for
of the boost converter [5]. To overcome these problems, solving the local minima problem in the neuro-fuzzy system
different conventional and intelligent MPPT algorithms In this work, a high performance neuro-fuzzy MPPT control
have been proposed such as Incremental Conductance (IC) method is proposed. A five-layer based neuro fuzzy control
[6–8], Open Circuit Voltage (OCV) [9], Short Circuit Current is made to monitor the MPP of the PV system. The
(SCC) [10], Perturb and Observe (P&O) [11], fuzzy logic information extracted from the fuzzy control is used to
[12–15], feedback linearization [16], neural network [17– initialize the factors of the suggested framework or
22], neuro-fuzzy [23–25] and sliding mode [26,27]. proposed structure. An on-line learning algorithm based on
Nevertheless, there still remains the concern of fast and the embedded Neural Fuzzy (NF) gradient-decent-based
accurately determining the locus of the MPP during high back-propagation algorithm is derived to update the
weather variations and external load changes occurring. parameters of the proposed structure adaptively.
The fuzzy logic control is selected as it is effective and Compared with the conventional and intelligent MPPT
versatile, robust and flexible, without including control algorithms, such as PID-based P&O, the proposed MPPT
76 | P a g e
 NOVATEUR PUBLICATIONS
International Journal of Research Publications in Engineering and Technology [IJRPET]
ISSN: 2454-7875
VOLUME 3, ISSUE 6, Jun.-2017
controller over performs in terms of efficiency, power respond differently to the changes in solar insolation and
quality and MPP error. The paper is structured as follows: cell temperature and this can cause the MPPT system to
Section 2 focuses on the PV system. Section 3 presents the become inefficient. The basic layout of the photovoltaic
proposed controller design. Section 4 describes the system proposed includes a PV panel, a boost converter, a
performance of the proposed solution using detailed MPPT controller and a storage device (Fig. 1). The MPPT
simulations, followed by the conclusion in Section 5. control is performed using two different methods.

II. MPPT CONTROL: A. P&O ALGORITHM:

A successful MPPT-PV system design must take into First the P&O algorithm is used. P&O is the most
consideration a few requirements. Stability is the most common algorithm because of its ease of implementation,
fundamental design requirement of a dynamic control despite its drawbacks such as slow response speed,
system. In PV power systems, the switching mode oscillation around the MPP in steady state, and even
converters are nonlinear systems and the output tracking in wrong way under rapidly changing atmospheric
characteristics of solar array are also nonlinear. Therefore, conditions [6], [7]. Although it is known it does not provide
stability is a critical factor to evaluate a PV MPPT control the best results in all situations, it is mainly used to test the
systems dealing with non-linearity. Besides this, in MPPT MPPT controller functionality in the PV system. The
control systems, a good dynamic response is desirable for algorithm operates by periodically perturbing the control
the fast tracking requirement. A good MPPT control variable and comparing the instantaneous PV output power
algorithm needs to respond quickly to rapidly changing after perturbation with that before. Thus, the direction of
atmospheric conditions like temperature and illumination the next perturbation that should be used is determined [6-
and track the maximum power points quickly. It is also 9]. If the change in power ΔP > 0, the direction of the next
important to design a MPPT control system robust to any perturbation keeps the same algebraic sign. That should
kind of disturbances. The disturbances can arise from place the operation point closer to MPP. If ΔP < 0, the
various sources, one of the most common problems is that algebraic sign of the perturbation should be reversed.
the PV modules manufactured by different technologies

Figure 1. Photovoltaic system with fuzzy logic/ P&O control for MPPT.

B. FUZZY LOGIC BASED ALGORITHM: FUZZIFICATION:

The second method of tracking the MPP is based on the The fuzzification involves conversion of digital data in
fuzzy logic. The MPPT controller is generally composed of linguistic data. There are two inputs of the fuzzy logic
three main units: the fuzzification, the rule base and the controller - P and E. P is the power generated by the solar
defuzzification (Fig. 2) [10], [11]. panel and E denotes the error at k-th time sample, defined
as E(k)=P(k)-P(k-1). The inputs P and E are converted to
fuzzy membership values on the fuzzy subsets (Figure 3 (a)
Rule base and (b)). Seven membership functions are used for the
input Negative Big (NB), Negative Medium (NM), Negative
Small (NS), Zero (Z), Positive Small (PS), Positive Medium
Inputs Output
(PM) and Positive Big (PB). The E input is coded using three
Fuzzyfication Inference Defuzzyfication
membership functions expressed as linguistic variables
denoted Negative - NEG, Zero - Z, Positive – POZ. The input
Figure 2. Block scheme of a fuzzy controller.

77 | P a g e
 NOVATEUR PUBLICATIONS
International Journal of Research Publications in Engineering and Technology [IJRPET]
ISSN: 2454-7875
VOLUME 3, ISSUE 6, Jun.-2017
variables are fuzzified by using trapezoidal MFs for P and R1: If P is NB and E is NEG then Iref is NB;
triangular for E. R2: If P is NB and E is Z then Iref is NB;
INFERENCE: R3: If P is NB and E is POZ then Iref is NM; etc.
The second step is the inference, where the fuzzified Figure 4 shows the membership functions of input and
variables are compared with predefined sets in order to get output variables. On the ox axis the universe of discourse is
the appropriate response. It is responsible of the represented, while on oy axis there is the membership
interpretation of the rules using information collected in grade taking values between 0 and 1.
the knowledge base to compute the fuzzy set output.
The fuzzy rules set are a collection of expert control
knowledge allowing the fuzzy control objectives
achievements. The control rules base is set up using IF-
THEN rules, based on expert experience and engineering Figure 3. Graphical construction of the control
knowledge. Inference fuzzy rules for the PV system include signal in the MPPT controller (generated in the Matlab
13 fuzzy control rules. Fuzzy Logic Toolbox).
Mamdani fuzzy inference method is used with Max-Min
operation fuzzy combination. This method implies the
output membership function to be fuzzy sets. After the
aggregation process, there is a fuzzy set for every output
variable, leading to the necessity of a defuzzification. The
operations used in the inference process are: And method is
min; Or method is max; Implication is min; Aggregation is
max.

DEFUZZIFICATION:
(a)
Defuzzification of the inference engine evaluates the
rules based on a set of control actions for a given fuzzy
inputs set. This operation converts the inferred fuzzy
control action into a numerical value at the output by
forming the “union” of the outputs resulting from each rule.
In other words, the deffuzification plays the role of a
linguistic-to-numerical data converter. The center of area
(COA) algorithm is used for defuzzification of output
control parameter Iref. The duty cycle of the boost (b)
converter is adjusted thought Iref such that the system
operates at the maximum power point. The coding of the
membership functions for the otput Iref is identical to that
of the input P: Negative Big (NB), Negative Medium (NM),
Negative Small (NS), Zero (Z), Positive Small (PS), Positive
Medium (PM) and Positive Big (PB). In Table 1 it is
summarized the different fuzzy rules used in the fuzzy
controller to track the maximum power point.
Tabel 1 Fuzzy rules (c)
P NB NM NS Z PS PM PB Figure 4. Membership functions for: (a) input P, (b)
E
input , c) output Iref.
NEG NB NB NM Z PS PM PB
Figure 3 represents the graphical construction of the
Z NB NM NS PS PS PM PB
algorithm in the core of the controller. Each of the thirteen
POZ NM NM Z PS PS PB PB
rows refers to one rule. With two inputs and one output
Every rule of the rule base of the fuzzy logic system the input-output mapping is a surface. Figure 5 is a mesh
establishes a fuzzy relation between the input fuzzy sets plot of the relationship between P and E on the input side,
and the output fuzzy set Iref. There are 13 rules in the and controller output Iref on the output side. The plot
system rule base that make up the control strategy. The 13 results from the rule base with thirteen rules previously
rules are presented to the end-user in if-then format like presented. The surface is more or less regular. The
the one below: horizontal plateaus are due to flat peaks on the input sets.

78 | P a g e
 NOVATEUR PUBLICATIONS
International Journal of Research Publications in Engineering and Technology [IJRPET]
ISSN: 2454-7875
VOLUME 3, ISSUE 6, Jun.-2017
E
 jK (t  1)   jK (t )   * (3)
 jK
Where, η and η* = learning rate.
Learning rate parameter is selected by the user and, as it
can be deduced from equation (2), it plays an important
role in the convergence of the network in terms of success
and speed. For our experiments the most commonly used
parameters are selected. The inspection of advanced
possibilities related to neural network learning procedures
confirms a broad field of investigation and could be,
therefore, a point of further experimentation. In the back
Figure 5. Control surface
reproduction learning algorithm online training is usually
considerably quicker than batch training, especially in the
C. NEURAL NETWORK BASED CONTROL:
case of large training sets with many similar training
illustrations. On the other hand, results of the training with
back propagation and update after every pattern
presentation, heavily depend on a proper choice of the
parameter η. The back propagation weight update rule, also
called generalized delta-rule, for the NN software reads as
follows:
wij   j oi
(4)
vij   *  k yi
Figure 6. Fully connected neural network structure
 f   net j  t j  o j 
The computation of neural control is based on 
j   (5)
 f   net j   o j wij
fully connected neural network structure, which is
consists of an input layer with two neurons (n), one  k
hidden layer with four neurons (h) and a single neuron in  f   netk  tk  yk 
output layer (m). The structure of NN presented in 
k   
control configuration as depicted in Fig. 5 with x is the n × f  netk   yk v jk

1 input vector and y is a m × 1 diagonal vector. Here, ω k

and ϑ denotes the input-to-hidden layer and hidden -to- Where

output layer weights respectively in feed forward NN η = learning factor (a constant).
structure. The essence of Back propagation learning δj = Error
algorithm is the recurring application of the sequence oi = output of the preceding unit i.
concept to estimate the impact of each weight in the tj = teaching input of unit j.
network with respect to an arbitrary error function E: i = index of a predecessor to the current unit j with link
wij
E E ai neti j = index of the current unit.
 (1) The output signal from neural network structure is
wij ai neti wij employed for improving tracking response of MPPT.
Where
wij = weight from neuron j to neuron i. III. SIMULATION RESULTS:
ai = activation value. In this section, the solar system implemented in
Neti = weighted sum of the inputs of neuron i. Matlab/Simulink is evaluated. Two MPPT methods are
Once the partial derivative of each weight is known, the aim studied by simulation: the Neuro-Fuzzy and the P&O. Two
of reducing the error function is achieved by performing a situations are simulated. The first test is when the solar
simple gradient descent: irradiation is kept constant. The standard conditions are
E considered: the temperature = 25°C and the solar
wij (t  1)  wij (t )   (2) irradiation level = 1000W/m2. In the second case the
wij
system behavior is evaluated when solar illumination
change takes place and the temperature is kept constant.
79 | P a g e
 NOVATEUR PUBLICATIONS
International Journal of Research Publications in Engineering and Technology [IJRPET]
ISSN: 2454-7875
VOLUME 3, ISSUE 6, Jun.-2017
The purpose of these simulations is to observe the time
response and accuracy in terms of tracking the MPPT. For
building the prototype 10W PV module is used considering
the hardware development cost. The PV module has
specification: solar panel power 10 W, open circuit voltage
(Voc) 21.65 V, short circuit current (Isc) 0.69A, rated
voltage 17.85 V, rated current 0.65 A. Figure 7 and 8 shows
I-V Characteristics and P-V characteristics of 10W PV
Module with constant radiation and varying solar radiation
respectively. Figure 9 and 10 shows tracking of MPP at
solar radiation 400 W/m2 and 700 W/m2 with change in
external temperature. The tracking response is enhanced
due to neural network structure with back propagation Figure 9. Tracking of MPP at 400 W/m2 &
algorithm. changing temperature conditions

Figure 10. Tracking of MPP at 700 W/m2 &

Figure 7. I-V Characteristics of 10W PV Module
changing temperature condition

Figure 8. P-V Characteristics of 10W PV Figure12. Tracking of MPP at 700 W/m2 & changing
Module temperature condition

Figure11. Tracking of MPP at 700 W/m2 & changing Figure13. Tracking of MPP at 700 W/m2 & changing
temperature condition temperature condition
80 | P a g e
 NOVATEUR PUBLICATIONS
International Journal of Research Publications in Engineering and Technology [IJRPET]
ISSN: 2454-7875
VOLUME 3, ISSUE 6, Jun.-2017
Figure 11, 12, 13 shows tracking responce of power Tracking Algorithm for PV Arrays”, 35th Annul IEEE
based on neuro fuzzy, fuzzy logic & P and O methode Power Electronics Specialists Conference, 2004, Vol. 3,
respectively. The comparative analysis illustrate that, P and pp. 2005-2010.
O methode gives Oscillatory response, which can be smooth 8) K. H. Hussein, I. Muta, T. Hoshino, and M. Osakada
with the help of fuzzy logic concept. But in case of fuzzy, “Maximum photovoltaic power tacking: An algorithm
desired power point tracking is not possible so that, Neural for rapidly changing atmospheric conditions”, Proc.
network combined in parallel with fuzzy logic controller. IEEE - Generation. Transmission. Distribution, Vol.
142, 1\10. 1, Jan 1995, pp. 59-61.
IV. CONCLUSION: 9) W. Xiao, W.G. Dunford, “A Modified Adaptive Hill
In this work, an intelligent neuro-fuzzy direct method Climbing MPPT Method for Photovoltaic Power
with high adaptive capability is designed for the MPPT of a Systems,” 35th Annual IEEE Power Electronics
PV system. A five-layer NFC is adopted as the process Specialists Conference, 2004, pp. 1957-1963.
feedback controller. The proposed control is initialized 10) “Analysis of Perturb and Observe Maximum Power Point
from the traditional fuzzy control by means of expert Tracking Algorithm for Photovoltaic Applications,” 2nd
knowledge, which decreases the weight of the lengthy pre- IEEE International Conference on Power and Energy
learning. With a produced learning plan, the factors are (PECon 08), December 1-3, 2008, pp. 237-242.
modified in the proposed structure adaptively by 11) A. Chouder, F. Guijoan and S. Silvestre, “Simulation of
monitoring and modifying the tracking error. The simulator fuzzy-based MPP tracker and performance comparison
results show that the Neuro-fuzzy control algorithm with perturb & observe method,” Revue des Energies
considerably enhances the performance during the tracking Renouvelables, Vol. 11, N°4, 2008, pp. 577 – 586.
phase as compared to a conventional algorithm of the 12) W. C. So, C. K. Tse and Y. S. Lee, "Development of a
maximum power point tracking (MPPT) in photovoltaic Fuzzy Logic Controller for DC-DC converters: Design,
power systems. It provides fast response times and stability Computer Simulation, and Experimental Evaluation,”
for changing environmental conditions. Stability and IEEE Transactions on Power Electronics, Vol.11, No.1,
robustness is proven even in the case of a luminosity January 1996, pp. 24-32.
variation. 13) Alajmi, B.N.; Ahmed, K.H.; Finney, S.J.; Williams, B.W.
Fuzzy-Logic-Control Approach of a Modified Hill-
REFERENCES: Climbing Method for Maximum Power Point in
1) U. R. Yaragatti, A. N. Rajkiran, B. C.Shreesha, “A novel Microgrid Standalone Photovoltaic System. IEEE Trans.
method of fuzzy controlled maximum power point Power Electron. 2011, 26, 1022–1030.
tracking in photovoltaic system,” IEEE International 14) Shiau, J.K.; Wei, Y.C.; Chen, B.C. A study on the fuzzy-
Conference on Industrial Technology, 2005, pp 1421 – logic-based solar power MPPT algorithms using
1426. different fuzzy input variables. Algorithms 2015, 8,
2) M.S. Aït Cheikh, C. Larbes, G.F. Tchoketch Kebir and A. 100–127.
Zerguerras, “Maximum power point tracking using a 15) Ramalu, T.; Mohd Radzi, M.; AtiqiMohd Zainuri, M.;
fuzzy logic control scheme,” Revue des Energies Abdul Wahab, N.; Abdul Rahman, R. A photovoltaic-
Renouvelables, 2007, Vol. 10, N°3, pp 387 – 395. based SEPIC converter with dual-fuzzy maximum power
3) G.Y. Ayvazyan1, G.H. Kirakosyan2, and A.H. point tracking for optimal buck and boost operations.
Vardanyan1, “Maximum Power Operation of PV System Energies 2016, 9, 604.
Using Fuzzy Logic Control,” Armenian Journal of 16) Hadjaissa, A.; Ait cheikh, S.M.; Ameur, K.; Essounbouli,
Physics, vol. 1, 2008, pp. 155-159. N. A GA-based optimization of a fuzzy-based MPPT
4) A. El Hajjaji, M. BenAmm controller for a photovoltaic pumping system, Case
5) ar, J. Bosche, M. Chaabene, and A. Rabhi, “Integral study for Laghouat, Algeria. In Proceedings of the 8th
Fuzzy Control for Photovoltaic Power Systems,” IFAC Conference on Manufacturing Modelling,
Sustainability in Energy and Buildings, Springer Berlin Management and Control MIM 2016, Troyes, France,
Heidelberg, 2009, pp. 219-228. 28–30 June 2016; pp. 692–697.
6) N. Patcharaprakiti, S. Premrudeepreechacharn, Y. 17) Salimi, M.; Siami, S. Cascade nonlinear control of DC-DC
Sriuthaisiriwong, “Maximum power point tracking buck/boost converter using exact feedback
using adaptive fuzzy logic control for grid-connected linearization. In Proceedings of the 2015 4th
photovoltaic system,” Renewable Energy, 2005, Vol 3, International Conference on Electric Power and
No 11, pp. 1771-1788. Energy Conversion Systems (EPECS), Sharjah, UAE,
7) Xuejun Liu and Luiz A. C. Lopes, “An Improved 24–26 November 2015; pp. 1–5.
Perturbation and Observation Maximum Power Point

81 | P a g e
 NOVATEUR PUBLICATIONS
International Journal of Research Publications in Engineering and Technology [IJRPET]
ISSN: 2454-7875
VOLUME 3, ISSUE 6, Jun.-2017
18) Khaldi, N.; Mahmoudi, H.; Zazi, M.; Barradi, Y. The 30) Liu, Y.H.; Liu, C.L.; Huang, J.W.; Chen, J.H. Neural-
MPPT control of PV system by using neural networks network-based maximum power point tracking methods
based on Newton Raphson method. In Proceedings of for photovoltaic systems operating under fast changing
the 2014 International Renewable and Sustainable environments. Sol. Energy 2013, 89, 42–53.
Energy Conference (IRSEC), Ouarzazate, Morocco, 17– 31) Vazquez, J.R.; Martin, A.D.; Herrera, R.S. Neuro-Fuzzy
19 October 2014; pp. 19–24. control of a grid-connected photovoltaic system with
19) Mohapatra, A.; Nayak, B.; Mohanty, K.B. Performance power quality improvement. In Proceedings of the 2013
improvement in MPPT of SPV system using NN IEEE EUROCON, Zagreb, Croatia, 1–4 July 2013; pp.
controller under fast changing environmental condition. 850–857.
In Proceedings of the 2016 IEEE 6th International 32) Shanthi, T.; Vanmukhil, A.S. ANFIS Controller based
Conference on Power Systems (ICPS), New Delhi, MPPT Control of Photovoltaic Generation System. Res. J.
India, 4–6 March 2016; pp. 1–5. Appl. Sci. 2013, 8, 375–382.
20) Ou, T.C.; Hong, C.M. Dynamic operation and control of 33) Lin, F.J.; Shen, P.H. Adaptive fuzzy-neural-network
microgrid hybrid power systems. Energy 2014, 66, 314– control for a DSP-based permanent magnet linear
323. synchronous motor servo drive. IEEE Trans. Fuzzy Syst.
21) Hong, C.M.; Ou, T.C.; Lu, K.H. Development of intelligent 2006, 14, 481–495.
MPPT (maximum power point tracking) control for a 34) Atakulreka, A.; Sutivong, D. Avoiding local minima in
grid-connected hybrid power generation system. Energy feedforward neural networks by simultaneous learning.
2013, 50, 270–279. In Australasian Joint Conference on Artificial
22) Ma, S.; Chen, M.;Wu, J.; Huo,W.; Huang, L. Augmented Intelligence; Springer: Berlin, Germany, 2007; pp. 100–
Nonlinear Controller for Maximum Power-Point 109.
Tracking with Artificial Neural Network in Grid- 35) Choi, B.; Lee, J.H.; Kim, D.H. Solving local minima
Connected Photovoltaic Systems. Energies 2016, 9, problem with large number of hidden nodes on two-
1005. layered feed-forward artificial neural networks.
23) Messalti, S.; Harrag, A.; Loukriz, A. A new variable step Neurocomputing 2008, 71, 3640–3643.
size neural networks MPPT controller: Review, 36) Lo, J.T.H.; Gui, Y.; Peng, Y. Overcoming the Local-
simulation and hardware implementation. Renew. Minimum Problem in Training Multilayer Perceptrons
Sustain. Energy Rev. 2017, 68, 221–233. with the NRAE Training Method; International
24) Subiyanto, S.; Mohamed, A.; Hannan, M.A. Intelligent Symposium on Neural Networks; Springer: Berlin,
maximum power point tracking for PV system using Germany, 2012; pp. 440–447.
Hopfield neural network optimized fuzzy logic 37) Lee, C.Y.; Lin, C.J. A wavelet-based neuro-fuzzy system
controller. Energy Build. 2012, 51, 29–38. and its applications. Intell. Autom. Soft Comput. 2007,
25) Khosrojerdi, F.; Taheri, S.; Cretu, A.M. An adaptive 13, 385–403.
neuro-fuzzy inference system-based MPPT controller for 38) Abiyev, R.H.; Kaynak, O. Identification and Control of
photovoltaic arrays. In Proceedings of the 2016 IEEE Dynamic Plants Using Fuzzy Wavelet Neural Networks.
Electrical Power and Energy Conference (EPEC), In Proceedings of the 2008 IEEE International
Ottawa, ON, Canada, 12–14 October 2016; pp. 1–6. Symposium on Intelligent Control, San Antonio, TX,
26) Abu-Rub, H.; Iqbal, A.; Ahmed, S.M. Adaptive neuro- USA, 3–5 September 2008; pp. 1295–1301.
fuzzy inference system-based maximum power point 39) Cao, C.; Ma, L.; Xu, Y. Adaptive control theory and
tracking of solar PV modules for fast varying solar applications. J. Control Sci. Eng. 2012, 2012, 827353.
radiations. Int. J. Sustain. Energy 2012, 31, 383–398. 40) Rouzbehi, K.; Miranian, A.; Luna, A.; Rodriguez, P.
27) Rekioua, D.; Achour, A.Y.; Rekioua, T. Tracking power Identification and maximum power point tracking of
photovoltaic system with sliding mode control strategy. photovoltaic generation by a local neuro-fuzzy model. In
Energy Procedia 2013, 36, 219–230. Proceedings of the IECON 2012—38th Annual
28) Valencia, P.A.O.; Ramos-Paja, C.A. Sliding-mode Conference on IEEE Industrial Electronics Society,
controller for maximum power point tracking in grid- Montreal, QC, Canada, 25–28 October 2012; pp. 1019–
connected photovoltaic systems. Energies 2015, 8, 1024.
12363–12387. 41) Pradhan, R.; Subudhi, B. Design and real-time
29) Cheng, P.C.; Peng, B.R.; Liu, Y.H.; Cheng, Y.S.; Huang, implementation of a new auto-tuned adaptive MPPT
J.W. Optimization of a fuzzy-logic-control-based MPPT control for a photovoltaic system. Int. J. Electr. Power
algorithm using the particle swarm optimization Energy Syst. 2015, 64, 792–803.
technique. Energies 2015, 8, 5338–5360.

82 | P a g e