Electrical Power and Energy Systems 39 (2012) 21–28

Contents lists available at SciVerse ScienceDirect

Electrical Power and Energy Systems

journal homepage: www.elsevier .com/locate / i jepes

Maximum power point tracking using fuzzy logic control

Mohamed M. Algazar a, Hamdy AL-monier b, Hamdy Abd EL-halim a, Mohamed Ezzat El Kotb Salem c,⇑a Electrical Power & Machines Dept., Faculty of Engineering, Al-Azhar University, Egyptb Electronics & Communication Dept., Naser High Institute for Engineering and Technology, Egyptc B. Sc.Electrical Power & Machines Engineering, Al Azhar University, Cairo 2004, Egypt

a r t i c l e i n f o a b s t r a c t

Article history:Received 12 July 2011Received in revised form 10 November 2011Accepted 5 December 2011Available online 29 February 2012

Keywords:Photovoltaic (PV)Maximum power point tracking (MPPT)Fuzzy logic control (FLC)DC–DC Cúk converterDC water pumpMATLAB/SIMULINK

0142-0615/$ - see front matter � 2011 Elsevier Ltd. Adoi:10.1016/j.ijepes.2011.12.006

⇑ Corresponding author. Tel.: +20 01001832433.E-mail address: moezzat3@yahoo.com (M.E.E.K. Sa

This paper proposes an intelligent control method for the maximum power point tracking (MPPT) of aphotovoltaic system under variable temperature and insolation conditions. This method uses a fuzzylogic controller applied to a DC–DC converter device. The different steps of the design of this controllerare presented together with its simulation. The PV system that I chose to simulate to apply my techniqueson it is stand-alone PV water pumping system. Results of this simulation are compared to those obtainedby the system without MPPT. They show that the system with MPPT using fuzzy logic controller increasethe efficiency of energy production from PV.

� 2011 Elsevier Ltd. All rights reserved.

1. Introduction

In our Arabian nation peopled areas is very small comparing tototal area because fresh water resources concentrate in theseareas. Now, the existing fresh water almost enough for our needs,but in the near future with increasing in people numbers it willbe huge problem. On the other hand we have shining sun allthe year, so we can use stand alone PV-powered water pumpingsystem to get water in non peopled areas. Unfortunately theactual energy conversion efficiency of PV module is rather low.So to overcome this problem and to get the maximum possibleefficiency, the design of all the elements of the PV system hasto be optimized.

In order to increase this efficiency, MPPT controllers are used.Such controllers are becoming an essential element in PV systems.A significant number of MPPT control have been elaborated sincethe seventies, starting with simple techniques such as voltageand current feedback based MPPT to more improved power feed-back based MPPT such as the perturbation and observation (P&O)technique or the incremental conductance technique [1,2]. Re-cently intelligent based controls MPPT have been introduced.

In this paper, an intelligent control technique using fuzzy logiccontrol is associated to an MPPT controller in order to improve en-ergy conversion efficiency.

ll rights reserved.

lem).

2. The proposed system

The proposed system in this paper is stand-alone DC waterpumping without backup batteries. As shown in Fig. 1, the systemis very simple and consists of subsystems:

2.1. PV module

BP Solar BP SX 150S PV module is chosen for a MATLAB simula-tion model. The module is made of 72 multi-crystalline silicon so-lar cells in series and provides 150 W of nominal maximum power[14]. Table 1 shows its electrical specification.

2.2. DC–DC Cúk converter

The basic operation of Cúk converter and derivation of the volt-age transfer function is explained in Fig. 2:

The voltage transfer function of Cúk converter is written as [15]:

Vo

Vs¼ D

1� Dð1Þ

Its relationship to the duty cycle (D) is:

� If 0 < D < 0.5 the output is smaller than the input.� If D = 0.5 the output is the same as the input.� If 0.5 < D < 1 the output is larger than the input.

Fig. 1. Block diagram of the proposed PV water pumping system.

Fig. 2. Cúk converter waveforms: (a

Table 1Module electrical characteristics data of PV module taken from the datasheet [14].

Electrical characteristics

Maximum Power (Pmax) 150 WVoltage at Pmax (Vmp) 34.5 VCurrent at Pmax 4.35 AWarranted minimum Pmax 140 WShort-circuit current (Isc) 4.75 AOpen-circuit voltage (Voc) 43.5 VMaximum system voltage 600 VTemperature coefficient of Isc (0.065 ± 0.015)%/�CTemperature coefficient of Voc �(160 ± 20)mV/�CTemperature coefficient of power �(0.5 ± 0.05)%/�CNOCT5 47 ± 2 �C

22 M.M. Algazar et al. / Electrical Power and Energy Systems 39 (2012) 21–28

) switch off; (b) switch on [15].

Table 2Design specification of the Cúk converter [5].

Specification

Input voltage (Vs) 20–48 VInput current (Is) 0–5 A (<5% Ripple)Output voltage (Vo) 12–30 V(<5% Ripple)Output current (Io) 0–5 A (<5% Ripple)Maximum output power (Pmax) 150 WSwitching frequency (f) 50 KHzDuty cycle (D) 0.1 6 D 6 0.6

Fig. 3. Kyocera SD 12–30 water pump performance chart [16].

Fig. 4. PV module is directly connected to a (variable) resistive load [5].

Fig. 5. I–V curves of BP SX 150S PV module [14] and various resistive loadssimulated with the MATLAB model (1 KW/m2, 25 �C) [5].

M.M. Algazar et al. / Electrical Power and Energy Systems 39 (2012) 21–28 23

Here, a Cúk converter is designed based on the specificationshown in Table 2.

2.3. FL controller

This is the main subject of this paper; I will discuss it in the nextsections.

2.4. DC water pump

Fig. 3 shows the relationship between flow rate of water and to-tal dynamic head for the Kyocera SD 12–30 solar pump to be mod-eled. It has the normal operating voltage of 12–30 V and themaximum power of 150 W.

3. Maximum power point tracking algorithms

When a PV module is directly coupled to a load, the PV module’soperating point will be at the intersection of its I–V curve and theload line which is the I–V relationship of load. For example in Fig. 4,a resistive load has a straight line with a slope of 1/Rload as shownin Fig. 5. In other words, the impedance of load dictates the oper-ating condition of the PV module. In general, this operating point isseldom at the PV module’s MPP, the optimal adaptation occursonly at one particular operating point, called Maximum PowerPoint (MPP) and noted in our case Pmax.

To overcome this problem, it is necessary to add an adaptationdevice, MPPT controller with a DC–DC Cúk converter, between thesource and the load [3].

Furthermore the location of the MPP in the I–V plane is notknown beforehand and always changes dynamically dependingon irradiance and temperature [6,7]. For example, Fig. 6 shows aset of PV I–V curves under increasing irradiance at the constanttemperature (25 �C), and Fig. 7 shows the I–V curves at the sameirradiance values but with a higher temperature (50 �C). Thereare observable voltage shifts where the MPP occurs. So, the MPPTcontroller is also required to track the new modified maximumpower point in its corresponding curve whenever temperatureand/or insolation variation occurs.

Many MPPT control techniques have been conceived for thispurpose these last decades [1,4].

Maximum power point tracking algorithms:

� Perturb and observe algorithm.� Incremental conductance algorithm.� Parasitic capacitances.� Constant voltage control.� Constant current control.� Pilot cell.� Artificial intelligent method. (Fuzzy logic control, neural

network. . .)� They can be classified as:� Voltage feedback based methods which compare the PV operat-

ing voltage with a reference voltage in order to generate thePWM control signal of the DC–DC converter [8].� Current feedback based methods which use the PV module

short circuit current as a feedback in order to estimate the opti-mal current corresponding to the maximum power.� Power based methods which are based on iterative algorithms

to track continuously the MPP through the current and voltagemeasurement of the PV module [5].

4. Fuzzy logic MPPT controller

Recently fuzzy logic controllers have been introduced in thetracking of the MPP in PV systems [9,10]. They have the advantageto be robust and relatively simple to design as they do not requirethe knowledge of the exact model. They do require in the otherhand the complete knowledge of the operation of the PV systemby the designer.

The proposed system in this thesis consist of two input vari-ables: error (E) and change of error (CE), and one out variable: dutyratio or duty cycle (D), as shown in Fig. 8.

Fig. 6. I–V curves for varying irradiance and a trace of MPPs (25 �C) [5].

Fig. 7. I–V curves for varying irradiance and a trace of MPPs (50 �C) [5].

Fig. 8. General diagram of a fuzzy controller [3].

24 M.M. Algazar et al. / Electrical Power and Energy Systems 39 (2012) 21–28

4.1. Fuzzification

Membership function values are assigned to the linguistic vari-ables, using five fuzzy subsets: NB (negative big), NS (negativesmall), ZE (zero), PS (positive small), and PB (positive big). The par-tition of fuzzy subsets and the shape of membership function,which can adapt shape up to appropriate system, are shown inFig. 9. The value of input error (E) and change of error (CE) are nor-malized by an input scaling factor. In this system the input scalingfactor has been designed such that input values are between �1and 1.

Fig. 9. Fuzzy logic membership functions for inputs and output variables.

Table 3Fuzzy rule table.

E CE

NB NS ZE PS PB

NB ZE ZE NB NB NBNS ZE ZE NS NS NSZE NS ZE ZE ZE PSPS PS PS PS ZE ZEPB PB PB PB ZE ZE

M.M. Algazar et al. / Electrical Power and Energy Systems 39 (2012) 21–28 25

The triangular shape of the membership function of this arrange-ment presumes that for any particular input there is only one dom-inant fuzzy subset. I chose the number of fuzzy subset depending onthe required accuracy, some paper [11] using seven subset for thesame problem but the increasing in accuracy is very small.

The input variables error (E) and change of error (CE) for the fuz-zy logic controller can be calculated as follows:

EðKÞ ¼ DIDVþ I

V¼ DP

DV¼ DP

DIð2Þ

CEðKÞ ¼ EðKÞ � EðK � 1Þ ð3Þ

where I is the output current from PV array; DI is I(k) � I(k–1); V isoutput voltage from PV array; DV is the V(k) � V(k–1).

4.2. Inference method

The composition operation by which a control output can begenerated. Several composition methods such as MAX–MIN andMAX–DOT have been proposed in fuzzy tool box in MATLAB. Thecommonly used method is MAX–MIN is used in this thesis. Theoutput membership function of each rule is given by the MIN (min-imum) operator and MAX (maximum) operator. Table 3 shows therule table for fuzzy logic controller.

4.3. Defuzzification

Defuzzificaion for this system is the centre of gravity to com-pute the output of this FLC which is the duty ratio (cycle). The

centre of gravity method is both very simple and very fast method.The centre of gravity defuzzification method in a system of rules byformally given by:

D ¼Pn

j¼1lðDjÞ � DjPn

j¼1lðDjÞð4Þ

Duty ratio, the output of fuzzy logic control uses to controlthrough PWM which generated pulse to control MOSFET switchin DC–DC converter.

4.4. Fuzzy logic control simulation in MATLAB/SIMULINK

Fig. 10 shows details of FLC block in MATLAB/SIMULINK. As Imentioned in beginning of this section, FLC for MPPT have two in-put variables and one out variables. The first step in shown block isto calculate the input variables using Eqs. (1) and (2), then usingFLC block which programmed as mentioned in previous sectionsto calculate duty ration which is the Pulse Width Modulator input.

Fig. 12. Result of system with irradiance 1000 W/m2 and temperature 25 �C.

26 M.M. Algazar et al. / Electrical Power and Energy Systems 39 (2012) 21–28

5. Simulation of PV water pump system with MPPT

In this Section 1 will simulate the system in MATLAB/SIMULINK.Fig. 11 shows block diagram stand-alone PV water pumping sys-tem with MPPT using FLC.

I will first test the whole system with two certain atmosphericconditions and comparing the result with theoretical data and di-rect-coupled system.

Fig. 12 shows the test result for the first test scenario with irra-diance 1000 W/m2 and temperature 25 �C. The PV system withoutMPPT has poor efficiency because of mismatching between the PVmodule and the DC pump motor load. The power generated by PVsystem with MPPT (assume loss-less converter) is approximatelyequal the theoretical maximum generated power at this atmo-spheric condition. Assuming a DC–DC converter has efficiencymore than 95%, the system can increase the overall efficiency com-pared to the system without MPPT.

Fig. 13 shows the test result for the second test scenario withirradiance 500 W/m2 and temperature 15 �C. The result is showthat, the PV system with MPPT can increase the overall efficiencycompared to the system without MPPT, like that shown in the firsttest scenario Fig. 12.

Now for final test scenario, I will simulate the system usingatmospheric data (irradiation and temperature) from NationalRenewable Energy Laboratory [12] for one day as shown in Figs.14 and 15 to provide a comparison with the system without MPPTin terms of energy produce and flow rates of water pumped.

The day that I chose to using it’s atmospheric data for test thesystem is a very sunny day as shown in Fig. 14 (the irradiance

Fig. 10. Fuzzy logic c

Fig. 11. PV water pump system with MP

increase gradually from sun rise 5:00 AM to noon then decreasegradually to sun set 6:42 PM). The reason for this choice is to viewclearly the effect of using MPPT in the system.

The effect of changing temperature over day time for choosingday is small because the temperature value varies in the mostday hour around four degrees (18–21 �C), as show in Fig. 15.

ontroller block.

PT simulated in MATLAB/SIMULINK.

Fig. 14. Irradiance vs. time data from 5:00 AM to 6:42 PM for sunny day simulationpurpose.

Fig. 15. Temperature vs. time data from 5:00 AM to 6:42 PM for sunny daysimulation purpose.

Fig. 16. Generating power of PV system for day time period (from 5:00 AM to6:42 PM).

Fig. 17. Flow rates of PV system for day time period (from 5:00 AM to 6:42 PM).

Fig. 13. Result of system with irradiance 500 W/m2 and temperature 15 �C.

M.M. Algazar et al. / Electrical Power and Energy Systems 39 (2012) 21–28 27

Fig. 16 shows that, the generated power of PV system withMPPT is greater than the direct-coupled system. The difference willbe maximum at noon time. It shows also that the system withMPPT has smooth curve over the day, but direct-coupled systemhas oscillation over the day time. This oscillation will affect onthe performance of DC pump the long life run and its life time.

As shown in Fig. 17, the flow rate of Kyocera SD 12–30 waterpump is proportional to the power delivered. When the total dy-namic head is 30 m, the flow rate per watt is approximately86.7 cm3/W min. The minimum power requirement of pump mo-tor is 35 W [13]; therefore as long as the output power is higherthan 35 W, it pumps water with the flow rate above.

Fig. 17 shows also that the direct-coupled PV water pumpingsystem has a severe disadvantage because in the morning thepump will delay starting while the same system with MPPT is al-ready pumping water. Similarly, in the afternoon it goes idle nearlyearlier than the system with MPPT. The flow rate of water is alsolower throughout the operating period. The figures show thatMPPT offers significant performance improvement.

28 M.M. Algazar et al. / Electrical Power and Energy Systems 39 (2012) 21–28

6. Conclusion

The main purpose of this paper was to propose a new techniquefor MPPT of stand alone PV system using fuzzy logic control. Thispurpose leads us to achieve many goals, which can be summariesas follow:

� Design and simulate a PV array in MATLAB/SIMULINK usingdata sheet for commercial PV array, which give us the perfor-mance of the PV array (I–V curve and P–V curve) under anyatmospheric conditions (irradiance and Temperature) and canbe used in any research related with photovoltaic. The resultshows that the PV model using the equivalent circuit providesgood matching with the real PV module.� Design and simulate DC–DC Cúk converter and Pulse Width

Modulator in MATLAB/SIMULINK (SimPowerSystems) whichcan be use in any power electronics research.� Design and simulate a MPPT controller using FLC in MATLAB/

FUZZY TOOL BOX/ SIMULINK.

Finally, this paper presents a simple but efficient photovoltaicwater pumping system. It models each component and simulatesthe system using MATLAB/SIULINK. Simulations use SimPower-Systems in SIMULINK to model a DC pump motor. It performssimulations of the whole system and verifies functionality andbenefits of MPPT. Simulations also make another comparison withthe system without MPPT in terms of energy produced and flowrate of water pumped using atmospheric conditions for one day.The results validate that MPPT can significantly increase the effi-ciency of energy production from PV and the performance of thePV water pumping system compared to the system withoutMPPT. This increasing could bring large savings if the system islarge.

References

[1] Hohm DP, Ropp ME. Comparative study of maximum power point trackingalgorithms. Progress in Photovoltaics: Research and Applications; November2003. p. 47–62.

[2] Hussein KH, Muta I, Hoshino T, Osakada M. Maximum photovoltaic powertracking: an algorithm for rapidly changing atmospheric conditions. In: IEEproceedings – generation, transmission and distribution. vol. 142; January,1995. p. 59–4.

[3] Aït Cheikh MS�, Larbes C, Tchoketch Kebir GF, Zerguerras A. Maximum powerpoint tracking using a fuzzy logic control scheme. Revue des EnergiesRenouvelables 2007; 10(3): 387-395.

[4] Knopf H. Analysis, simulation and evaluation of maximum power pointtracking (MPPT) methods for a solar powered vehicle, Master thesis. Portland,State University; 1999.

[5] Akihiro Oi. Design and simulation of photovoltaic water pumping system,Master’s thesis. California Polytechnic State University, San Luis Obispo; 2005.

[6] Luque Antonio, Hegedus Steven. Handbook of photovoltaic science andengineering. John Wiley & Sons Ltd.; 2003.

[7] Roger Messenger, Ventre Jerry. Photovoltaic systems engineering. 2nd ed. CRCPress; 2003.

[8] Veerachary M, Senjyu T, Uezato K. Voltage-based maximum power pointtracking control of PV systems. IEEE Trans Aerosp Electron Syst 2002;38:262–70.

[9] Passino KM, Yurkovich S. Fuzzy control. Addison Wesley; 1998.[10] Masoum MAS, Sarvi M. A new fuzzy-based maximum power point tracker for

photovoltaic applications. Iran J Electrical Electron Eng 2005;1(January).[11] Nopporn Patcharaprakitia, 1, Suttichai Premrudeepreechacharnb,�, Yosanai

Sriuthaisiriwongb, 2 Maximum power point tracking using adaptive fuzzylogic control for grid-connected photovoltaic system. aDepartment of ElectricalEngineering, Rajamagala Institute of Technology, Chiang Rai 57120, Thailand.bDepartment of Electrical Engineering, Chiang Mai University, Chiang Mai50200, Thailand.

[12] National Renewable Energy Laboratory (NREL). Daily plots and raw data files;May 8, 2010 and July 4, 2010 <http://www.nrel.gov/solar_radiation/data. html>.

[13] Kyocera Solar Inc. Solar water pump applications guide; 2001 <www.kyocerasolar.com>.

[14] BP Solar BP SX150 – 150W. Multi-crystalline photovoltaic module datasheet;2001.

[15] Mohan, Undeland, Robbins. Power electronics – converters, applications, anddesign. 2nd ed. John Wiley & Sons Ltd.; 1995.

[16] Kyocera Solar Inc. Solar water pump applications guide; 2001 <http://www.?tul=0?>kyocerasolar.com>.