2018 International Conference on Computing, Power and Communication Technologies (GUCON)

Galgotias University, Greater Noida, UP, India. Sep 28-29, 2018

Fuzzy Logic based Improved P&O MPPT

Technique for Partial Shading Conditions
Ratnesh Kumar Bhavnesh Kumar Swaroop D
Division of ICE Division of ICE Division of ICE
Netaji Subhas Institute of Technology, Netaji Subhas Institute of Technology, Netaji Subhas Institute of Technology,
Delhi, INDIA Delhi, INDIA Delhi, INDIA
ratnesh0725@gmail.com kumar_bhavnesh@yahoo.co.in swaroopdevaraju@gmail.com

Abstract—This paper presents a hybrid method of method, all peaks can be detected successfully but it takes a
maximum power point tracking (MPPT) under partial shading large time to perform it.
conditions for photovoltaic (PV) power generators. The
perturb and observe (P&O) algorithm is one of extensively
used scheme for MPPT, but in case of partial shading
conditions, the P-V curve have multiple peaks and this
conventional method may not converge to the global maximum
power point (GMPP). Here, whole search space of MPP is
divided into three regions and a fuzzy logic controller has been
proposed to identify the region of GMPP. After identification
P&O algorithm is used to track the GMPP. Detailed analysis of
the proposed scheme is done using MATLAB/Simulink
software.

Keywords- Fuzzy Logic Controller (FLC), Maximum power

point tracking (MPPT), Photovoltaic (PV), Partial shading
conditions (PSC’s), Perturb & Observe (P&O).

I. INTRODUCTION Fig. 1. Block diagram of a MPPT System

Nowadays, demand of photovoltaic (PV) systems is In order to have better performance and better overall
increasing rapidly due to environment friendly energy efficiency during these abnormal conditions, some intelligent
conversion process and the features of solar energy such as techniques such as Fuzzy Logic (FL), neural network has
plentiful amount, free of cost, sustainable and global also been applied [8]. In [9], a FL based algorithm is
availability. The contribution of PV cell energy is expected presented where slope of PV curve is taken as input while
to reach 800GW by the year of 2030 [1]. variations in duty ratio of a DC-DC converter is taken as the
Solar panel formed by interconnecting multiple solar output. Both input and output have 7 membership functions
cells is used to convert the solar energy into electrical and the system has a set of 7 rules. In this method, the entire
energy. Despite of various advantages, these solar cells have search space has been scanned and stored global MPP on PV
low energy conversion efficiency and high initial cost. curve through corresponding duty ratio. Another FL for MPP
Furthermore, output power varies nonlinearly with the output tracking has been proposed in [10] considering error in
voltage. Therefore, to fetch maximum power from solar power and change in error having 7 membership functions
panel, maximum power point tracking (MPPT) techniques and 49 rules. A FL assisted P&O method is also proposed in
are applied in these systems. Block diagram of a PV system which slope of PV curve and old step size (perturbation
using a MPPT is shown in Fig1. voltage) are inputs and change in step size is output. Each
input has 3 membership functions, output has 5 membership
In past, many MPPT techniques have been proposed for functions and a set of 9 rules have been defined. These FL
PV systems. These techniques include Perturb and Observe based MPP tracker provide superior performance than
(P&O) [2], Incremental conductance (INC) [3], Hill climbing classical approaches but consumes more time in scanning
[4], FOCV (Fractional Open Circuit Voltage) [5], FSCC due to large search space.
(Fractional Short-Circuit Current) [6]. Although, these
techniques track the maximum power point (MPP) during In this paper, a new method has been introduced to track
uniform and slow changing weather conditions, but may not the global MPP. Here, FL is used to identify the region of
converge to global MPP during partial shading and fast global MPP and a modified P&O algorithm is used to track
changing weather conditions. Under uniform irradiance the MPP in the identified search region.
conditions, only one MPP exists in P-V curve, but in case of This paper is organized as follows: section II describes
non-uniform irradiance (partial shading), there are. multiple the modelling of P-V cell and different partial shading
peaks that occur in P-V curve due to the limitation of string conditions, it also contains the characteristics of PV system
current. In [7], authors proposed a modified P&O algorithm, under different irradiance patterns. Section III has the
in which duty cycle of a DC-DC converter is varied between proposed MPPT technique, section IV has simulation results
the boundaries (highest and lowest) to detect all possible and section V has the conclusion of the research paper.
maximum power points in the P-V curve. Through this

978-1-5386-4491-1/18/$31.00 ©2018 IEEE 775

 II. SYSTEM DESCRIPTION
A. Basic characteristics of a PV cell
A PV cell can be represented by an equivalent single
diode model [11] shown in Fig2. The symbols used in the
model are as follows:
G : Irradiance
I ph : Photovoltaic Current Source;
Id : Diode Current;
I sh : Shunt Current;
: Shunt Resistance; Fig. 3. P-V characteristic of solar cell
R sh
Rs : Series Resistance;
I : Cell Current;
V : Cell Voltage;

Fig. 4. I-V characteristic of solar cell

Fig. 2. Equivalent circuit of solar cell
B. Partial shading condition
Using KCL (Kirchhoff’s Current Law), we can write the To generate a higher voltage, several PV modules can be
equation of photocurrent which is given in equation (1), connected in series and in parallel to satisfy the voltage or
I I ph  Id  I sh (1) current ratings of the load. In case of PSC’s, if there is at
least one PV module which is less illuminated, it will
The diode current is given by equation (2) produce lesser current. In this situation, the shaded module
will work as a load and consume some power generated by
ª § V  IR s · º
(2) other PV modules. Here the string current is limited by the
Id I s «exp ¨ ¸  1»
¬« © VT ¹ »¼ current of the less illuminated modules. This problem can be
avoided by connecting a parallel diode with the PV module
And the shunt current is given by equation (3), which is also known as bypass diode. This bypass diode
V  IR s permits the string current to circulate in the right path even
I sh (3) there is one or more than one modules are partially or
R sh
completely shaded. Because of bypass diodes, there are more
From equations (1), (2) and (3), the characteristic can be than one peaks occurs in P-V curve, known as local and
written as, global maxima points occur in the P–V characteristics curve.
ª § V  IR s · º § V  IR · Four PV arrays configurationhaving 3 modules in series
s (3S configuration) with different shading patterns are shown
I I ph  I s «exp ¨ ¸  1»  ¨ ¸ (4)
¬« © V T ¹ »¼ © R sh ¹ in Fig 5(a, b, c, d). P-V curves of these patterns are shown in
Fig 5. Here IR1,IR2 and IR3 are in w / m 2 .
A Simulink / MATLAB model is designed on the basis
of characteristic equation (4). Parameters of the designed
Simulink model are given in table I. At the standard test
conditions, it has the following P-V and I-V characteristics
shown in the Fig (3), and Fig (4).

TABLE I: PARAMETERS OF DESIGNED SIMULINK MODEL

Parameters Name Value
NS No of cell in a module 60
VOC Open circuit voltage 38.1 V
ISC Short circuit current 8.58 A
a Ideality factor of Diode 1.3
KI Temperature coefficient 0.0638 V/ Ԩ
RS Series resistance 0.20 Ω Fig. 5. PV arrays configuration with different PSC’s
RSH Shunt resistance 4000 Ω
Tref Standard test temperature 25 Ԩ
GS Standard test irradiance 1000 W/m2

776
 In PV array configuration ‘a’, there is no shaded module In case of pattern 1, the GMPP occurs in search region 1
and it has only one peak, while in configuration ‘b’, there is (SR1), in pattern 2, it occurs in SR2 and in case of pattern 3,
one shaded module i.e. two different irradiance levels (600 it occurs in SR3.
and 1000 w / m 2 )exist in the array and it is having two
peaks. In case of configuration ‘c’ and ‘d’, each module has
different shading intensity i.e. three different irradiance level
and these configurations have three peaks. So the number
peaks in P-V curve is same as the number of different
irradiance level in PV string. Fig6 shows PV curve pattern in
different shading conditions.

Fig. 8. P-V curve for different patterns

III. MODIFIED PERTURBE AND OBSERVE (P & O)

TECHNIQUE WITH FUZZY CONTROLLER
A. Proposed Technique:
Fig. 6. P-V Curves under different partial condition’s After analysing the PV curve pattern, it has been noticed
that the GMPP occurs in the specific search region based on
C. Study of Characteristic Curves Under PSC’s irradiance levels on modules. So, in our proposed technique
a fuzzy logic controller is used to divide the entire search
A PV array having 3S configuration shown in Fig.7, has
space into three different search regions and it is also used to
been considered for the analysis. Here IR1, IR2 and IR3 are
select the particular search region where GMPP has been
the irradiance values on respective modules. There are three
occurred. Once the search region is selected, a modified P &
patterns of irradiance level shown in the table II. The P-V
O algorithm is used to detect the GMPP.
curves of these configurations have been shown in the Fig 8.
Here the entire search space has been divided into three B. Fuzzy Logic Controller:
search regions (SR1, SR2 and SR3) on the basis of open
FLC became very popular in recent years, because it
circuit voltage (Voc) of PV module. So, the search regions
works with imprecise inputs, it does not need accurate
have been defined as SR1 (0V to 38.1V), SR2 (38.2V to
mathematical model, simple in designing and it can also
76.2V) and SR3 (76.3V to 114.3V).
dealwith non-linearity of the system. It includes three steps
that is fuzzification, inference and defuzzification.
In fuzzification step, mathematical inputs are converted
into linguistic variables. Here, triangular membership
function has been used. Three inputs of the fuzzy IR1, IR2
and IR3 (ranging from 0 to 1000 w/m2) have been converted
in linguistic variable such LL (low level), ML (medium
level) and HL (high level).Further the output V r e f is
converted into low(L), medium(M) and high(H) according to
our desired response. Fig.9 and Fig.10 show the membership
functions of input and output of fuzzy system.

Fig. 7. Irradiance pattern on modules

TABLE II: PATTERN CONSIDERED FOR ANALYSIS

Pattern/Irradiance IR1 IR2 IR3

Pattern 1 200 300 900
Pattern 2 100 600 800 Fig. 9. Membership function of IR1, IR2 and IR3
Pattern 3 500 700 950

777
 C. P & O Technique:
Perturb and observe algorithm is the most suitable
algorithm to track the MPPT in the non-linear PV curve.
Once the search region is selected, the P & O algorithm is
called to track the GMPP. In this method, perturbation is
done in PV array voltage and array power is observed. If the
power is increased by perturbation, then the perturbation is
performed in the same direction [5]. Once the MPP is
reached, the power at the next instant is decreased and then
the direction of perturbation is reverse. At the steady state,
this method oscillates around the MPP. To overcome this
Fig. 10. Membership function of
problem, perturbation size is kept very small.
V ref

D. Flowchart of the proposed algorithm:

In inference step, a set of fuzzy rules are defined
according to our prior knowledge of output pattern. In this
step, Mamdani fuzzy inference method has been used. Table-
III shows the 27 rules defined for the proposed fuzzy logic
controller. In the last step, MOM (mean of maxima) method
has been used for the defuzzification. Table-IV shows the
defuzzification outline for the proposed algorithm.
TABLE III: RULE BASE FOR FUZZY CONTROLLER

Rule Linguistic Linguistic

Number Inputs Output

IR1 IR2 IR3 V ref

1 HL HL HL H
2 HL HL ML H
3 HL HL LL M
4 HL ML HL H
5 HL ML ML M
6 HL ML LL M
7 HL LL HL M
8 HL LL ML M
9 HL LL LL L
10 ML HL HL H
11 ML HL ML M
12 ML HL LL M
13 ML ML HL M
14 ML ML ML H
15 ML ML LL M
16 ML LL HL M
17 ML LL ML M
18 ML LL LL L
19 LL HL HL M
20 LL HL ML M
21 LL HL LL L
22 LL ML HL M
23 LL ML ML M Fig. 11. Flowchart of proposed technique
24 LL ML LL H
25 LL LL HL L IV. SIMULATION STUDY AND RESULT
26 LL LL ML L In this work, MATLAB Simulink tool is used for the
27 LL LL LL H
study of the model. In case of pattern 1, the valve of V r e f is
0V i.e. the search region 1 will be selected and after that
TABLE IV: DEFUZZFICATION OF V ref
modified P & O is called to track the GMPP. The tracking
Linguistic Variable CrispValue SearchRegion result of pattern 1 is shown in the Fig12. In case of pattern 2
(V ref ) (V ref ) (SR) and pattern 3, the values of V r e f are 38.1V and 76.2V
L 0 SR1 respectively. Fig13, 14 are the tracking results of pattern 2
M 38.2 SR2 and pattern 3 respectively. Complete analysis result table of
H 76.3 SR3 the considered patterns is shown in the table V.

778
 TABLE V: RESULT TABLE OF DIFFERENT PATTERNS V. CONCLUSION
Pattern/ IR IR IR SR PM P P VM PP I M PP In this paper, the effect of the partial shading is analysed
Irradiance 1 2 3
(W a tt ) (V o lt ) ( A m p .) and their simulation responses are discussed. Partial shading
makes detection of global MPP complex for MPPT
Pattern 200 30 90 SR1 211.48 29.26 7.28 controllers and so reduces the system efficiency. Fuzzy logic
1 0 0 based MPP tracker is proposed to assist classical perturb &
Pattern 100 60 80 SR2 310.27 62.59 4.96 observe algorithm to search global MPP. The proposed
2 0 0 hybrid MPPT technique improves the efficiency of
Pattern 500 70 95 SR3 409.86 97.24 4.17 photovoltaic system under partial shading condition and
3 0 0 takes lesser time. Simulated results verify the effectiveness
of the proposed algorithm.
REFERENCES
[1] B. Subudhi and R. Pradhan, "A Comparative Study on Maximum
Power Point Tracking Techniques for Photovoltaic Power Systems,"
in IEEE Transactions on Sustainable Energy, vol. 4, no. 1, pp. 89-98,
Jan. 2013.
[2] M. I. Bahari, P. Tarassodi, Y. M. Naeini, A. K. Khalilabad and P.
Shirazi, "Modeling and simulation of hill climbing MPPT algorithm
for photovoltaic application," 2016 International Symposium on
Power Electronics, Electrical Drives, Automation and Motion
(SPEEDAM), Anacapri, 2016, pp. 1041-1044.
[3] M. A. Elgendy, B. Zahawi and D. J. Atkinson, "Assessment of
Perturb and Observe MPPT Algorithm Implementation Techniques
for PV Pumping Applications," in IEEE Transactions on Sustainable
Fig. 12. Pattern 1 tracking result. Energy, vol. 3, no. 1, pp. 21-33, Jan. 2012.
[4] F. Liu, S. Duan, F. Liu, B. Liu and Y. Kang, "A Variable Step Size
INC MPPT Method for PV Systems," in IEEE Transactions on
Industrial Electronics, vol. 55, no. 7, pp. 2622-2628, July 2008.
doi: 10.1109/TIE.2008.920550
[5] A. Thangavelu, S. Vairakannu and D. Parvathyshankar, "Linear open
circuit voltage-variable step-size-incremental conductance strategy-
based hybrid MPPT controller for remote power applications," in IET
Power Electronics, vol. 10, no. 11, pp. 1363-1376, 9 9 2017.
[6] A. Sandali, T. Oukhoya and A. Cheriti, "Modeling and design of PV
grid connected system using a modified fractional short-circuit
current MPPT," 2014 International Renewable and Sustainable
Energy Conference (IRSEC), Ouarzazate, 2014, pp. 224-229.
[7] E. Koutroulis and F. Blaabjerg, "A New Technique for Tracking the
Global Maximum Power Point of PV Arrays Operating Under Partial-
Shading Conditions," in IEEE Journal of Photovoltaics, vol. 2, no. 2,
Fig 13. Pattern 2 tracking result. pp. 184-190, April 2012.
[8] A. Sayal, "MPPT techniques for photovoltaic system under uniform
insolation and partial shading conditions," 2012 Students Conference
on Engineering and Systems, Allahabad, Uttar Pradesh, 2012, pp. 1-6.
[9] N. Shah and R. Chudamani, "A novel algorithm for global peak
power point tracking in partially shaded grid-connected PV
system," 2012 IEEE International Conference on Power and Energy
(PECon), Kota Kinabalu, 2012, pp. 558-563.
[10] Hasan Mahamudul, Mekhilef Saad, and Metselaar Ibrahim Henk,
“Photovoltaic System Modeling with Fuzzy Logic Based Maximum
Power Point Tracking Algorithm,” International Journal of
Photoenergy, vol. 2013, Article ID 762946, 10 pages, 2013.
[11] M. F. Nayan and S. M. S. Ullah, "Modelling of solar cell
characteristics considering the effect of electrical and environmental
parameters," 2015 3rd International Conference on Green Energy
and Technology (ICGET), Dhaka, 2015, pp. 1-6.

Fig. 14. Pattern 3 tracking result.

779