Fuzzy embedded MPPT modeling and control of PV system in a hybrid power system Syed Zulqadar Hassan*1, IEEE Member, Hui Li1, Tariq Kamal2, Mithulananthan Nadarajah2, Faizan Mehmood3 1
State Key Laboratory of Power Transmission Equipment and System Security and New Technology, School of Electrical Engineering, Chongqing University, Chongqing, China 2 Power and Energy System Group, School of Information Technology and Electrical Engineering, University of Queensland, Brisbane, QLD 4072, Australia 3 Electrical Engineering Department, University of Engineering & Technology, Taxila, Pakistan
[email protected],
[email protected],
[email protected],
[email protected],
[email protected] Abstract—The literature is populated with different Maximum Power Point Tracking (MPPT) methods for Photovoltaic (PV) system to obtain maximum power from it. This piece of work provides an artificial intelligence-based fuzzy logic MPPT modeling and control of PV system in a grid connected hybrid power system under different weather patterns. The proposed technique uses seven fuzzy sets with seven linguistic variables applied to a DC–DC converter. Furthermore, a battery module is added as an energy storage system during surplus power and/or backup device during load demand. The overall operation of system is performed by classical logic power management switching algorithm. The performance of proposed method is compared with and without Proportional Integral Derivative (PID) MPPT controllers. MATLAB simulation results show better behavior of proposed method in terms of load tracking and reliability. Index Terms— Fuzzy logic, PV system, Battery, Hybrid power system, Load tracking.
I. INTRODUCTION According to the Ministry of Planning Pakistan, during fiscal year 2014-2015, the total energy supply is 24,830 MW [1]. Moreover, according to the ministry report, only 106 MW (0.43% of total generation) is produced by Renewable Energy Sources while remaining is generated from fossil fuels e.g., hydel, nuclear, oil, and gas, etc. After rapid depletion of fossil fuels and the high import cost of fuel, Pakistan government realized the important of RES, that it reduces the use of fossil fuels, which protect the environment from green house effects [2], [3]. Knowing the fact of RES, Pakistani Government starts building numerous solar parks [4]. Hence, Photovoltaic (PV) systems power generation has been increased in Pakistan [5]. The PV process is the technology, which directly converts light energy into direct current (DC) electricity. For PV system, the power converter interface is required to transfer DC power to AC. The main objectives of electronic interface are to extract maximum power from PV by tracking Maximum Power Point (MPP) and also convert DC to appropriate AC [6]. The locus of MPP has a non-linear relationship with cell temperature and solar irradiance. Therefore, in order to operate PV at MPP, the system must have MPP tracker [7]. To track MPP, several intelligent and conventional control methods are proposed [8]. Conventional methods contain
perturbation and observation (P&O), voltage-feedback methods, Hill Climbing (HC) and Incremental Conductance (IC) etc., while intelligent method includes Fuzzy Logic Control (FLC), Genetic Algorithm (GA), Ant Colony Optimization (ACO) and Neural Network (NN) [9]–[12]. Numerous studies have been conducted on MPPT based on conventional and intelligent control methods. In [13]–[16], the authors track the MPP with P&O, IC and HC techniques due to its simple implementation. However, the authors are unable to eliminate the power oscillation created on MPP and divergence caused due to weather changes. Therefore, it is needed to have an appropriate MPPT method which can reduce oscillation and coverage quickly. Hence, intelligent control methods are used to overcome drawbacks of conventional control techniques [17]. In last few years, different researcher’s proposed intelligent control based MPPT [18], [19]. In [20], GA is used to increase the accuracy of an ANN-based MPPT algorithm. In GA, population size, mutation rate and number of generations are great problems. In [21], PSO with the capability of direct duty cycle is successfully applied to track MPP of a PV system but PSO still has the dependency problem on initial values of the particles. In [22], a novel ACO is implemented to get the PIMPPT controller gains. ACO is used to improve both the design efficiency of PI control systems and its performance to get optimal PI parameters. In ACO, the optimal selection of number of ants, solution archive and locality of search space etc., are some thoughtful issues. In this paper, an Artificial Intelligent (AI) based MPPT controller is proposed. In AI, FLC based controller is used to extract maximum power from PV System. The proposed controller operates on two primary inputs i.e., PV voltage and power and generates an appropriate duty cycle. This study mainly focuses on investigating MPPT performance, which was created by studying P-I graph of PV array and determining rule base of fuzzy logic controller with different weather conditions. Diffident linguistic variables are used to develop fuzzy model and form FLC controller. The performance of proposed controller is checked under real world environmental conditions. The paper is organized into six different sections. Starting from introduction described in Section II. Next section briefly describes the test-bed is created for simulation. Similarly, Section III enlightens the PV array modeling. Section IV
*Corresponding Author:
[email protected]
978-1-5090-3552-6/16/$31.00 ©2016 IEEE 1
II. TEST BED FOR PROPOSED CONTROLLER In this section, the design of entire test bed for proposed controller is discussed. The test bed contains a PV system along with storage device (i.e., battery) which is connected to Utility Grid (UG) via three phase inverter. The PV is connected to inverter via DC-DC boost converter, which is controlled by FLC. Battery system is added to provide backup to the grid at night time. UG receive power from PV and battery during peak hours while delivering power to battery for charging during off peak hours. Power sharing between different sources is performed by Power Management System (PMS). The complete architecture of test-bed is shown in figure 1.
Eqs. (1) and (2) are used during the simulations to achieve the output characteristics of the PV system at different irradiance and temperature conditions. The output power of PV system varies according to atmospheric condition or load current as shown in Fig. 2. Array type: SunPower SPR-305-WHT; 13 series modules; 66 parallel strings
0.75 kW/m2
300
0.5 kW/m2
200
0.25 kW/m2
100 0
0
100
200
300
400 500 Voltage (V)
600
700
AkT ( N p I SC I PV N p I ) N s Ns I PV RS ln NpI Q N p
(1)
where VPV denotes the output voltage of a PV; NP and NS are the number of cells joined in parallel and in series, respectively; Q is the electron charge; RS stands for series resistance of the SC; ISC gives the light-generated current; IPV provides the output current of SC; 𝐼′ is the reverse saturation current; A represents the dimensionless junction material factor; k is the Boltzmann constant (1.38×10−23 J/K); T is the temperature (K). The output power delivered by the PV system into the DC bus could be written as PPV VPV I PV (2)
5
x 10
Power (W)
0.75 kW/m2
2
0.5 kW/m2
1
0.25 kW/m2 0
0
100
200
300
400 500 Voltage (V)
600
700
800
900
Fig. 2: P-V and I-V characteristic curves of PV
The different parameters of PV used in this research is given in table I. At 1000 W/m2, the maximum output power is 260 kW. Table I: PV Parameters
Parameter Model No of cells/module No of series modules/string No of parallel strings Voltage at MPP/module Current at MPP/module Open Circuit Voltage Short-Circuit Current
Value SunPower SPR-305-WHT 96 13 66 54.7 V 5.58 A 64.2 V 5.98 A
Inverter
Battery
Distribution Transformer
PV Array
Utility Grid
+
+ _
_
IPV
900
1 kW/m2
PV based power generation is becoming progressively important as a RES since they present many advantages such as, environmentally friendly, emitting no noise, non-depleted, incurring no fuel costs and requiring little maintenance among others. The structural unit of PV array is the solar cell (SC), which is fundamentally a p–n semiconductor junction. The current–voltage (I-V) characteristic of a solar PV is a non-linear nature and is given by [23]–[25].
VPV
800
3
III. MODELING OF PV SYSTEM
VPV
1 kW/m2
400 Current (A)
describes the proposed methodology for designing MPPT FLC. Section V explains the detailed simulation results followed by concluding in Section (VI).
Maximum Power Point Tracking, FLC and P&O Algorithm PPV
P*PV
Battery charging and discharging controller PB
TB
P*B
Voltage, Current and Frequency regulator
Vabc Iabc
Utility grid controller
PL PG
P* G
Power Management System (PMS) Fig 1: Architecture of proposed hybrid power system
2
A. Fuzzy Logic based MPPT Controller Design It is difficult to model a non-linear system with a conventional controller in terms of regulation, damping etc. The FLC is based on the set of fuzzy rules developed using expert knowledge. A FLC contains a fuzzifier, fuzzy inference engine, fuzzy rule base and defuzzifier as shown in figure 3. In this research, Mamdani fuzzy model is used. Crisp ∆V in U
Fuzzifier Crisp ∆P in U
Fuzzy set in U
as input scalar and Z3 as output. Initially, different types of MFs are selected i.e., Trapezium, Cauchy, Gaussian, Triangle etc. Using expert knowledge [26], trial and error method, Triangle and Trapezium methods are used together. MFs for ∆V, ∆P and ∆D are shown in figures 4, 5 and 6, respectively. NB
1 Degree of Membership
IV. PROPOSED METHODOLOGY
0.4 0.2
V [V (n) V (n 1)] Z1
(3)
P [ P(n) P(n 1)] Z 2
(4)
D(n 1) [ D D] Z3
(5)
Degree of Membership
Crisp ∆D in V
Degree of Membership
PS
-0.2
0
0.2
0.4
0 -1
PB
PM
-0.8
NM
NS
ZE
PS
-0.6
-0.4
-0.2
0
0.2
0.4
Fig 6: MF for output variable ∆D
0.5
0
0
0.2
V 0
0.2
1
0.4
0.6
0.8
Seven fuzzy sets are established using seven linguistic variables. These linguistic variables are associated on two input variables. The linguistic terms used in establishing a fuzzy model are Negative Big (NB), Negative Medium (NM), Negative Small (NS), Zero (ZE), Positive Small (PS), Positive Medium (PM) and Positive Big (PB). The Membership Functions (MFs) are defined in interval of [-1,-1] with Z1 and Z2
-1
0
-0.5
-1
0.5
1
P
Fig 7: I/O surface waveform of fuzzy logic controller
1
V
Fig 4: MF for input variable ∆V
0.6
The input-output control surface for fuzzy controller is shown in figure 7. After understanding system parameters and error manipulations, the rules are developed for FLC. The rule based developed for system is shown in Table II. The fuzzy inferences system is designed using Mamdani model.
1
-0.2
0.8
PB
PM
D
0.4
-0.4
1
0.2
0.6
-0.6
0.8
0.4
-0.5
-0.8
0.6
0.6
0.8
0 -1
-0.4
NB
D
ZE
-0.6
0.8
where n represents the time index, Z1, Z2 and Z3 are input and output scaling gains, V(n) and P(n) represents the instantaneous voltage and power of PV array. 1
-0.8
1
For designing FLC model, initially it is recommended to find out the variables for the proposed system. In this paper, two input parameters i.e., Change in array power (∆P) and array voltage (∆V), and one output (Change in duty cycle (∆D)) are used to develop FLC for MPPT. Input and output variables of FLC are defined as;
NS
PB
PM
Fig 5: MF for input variable ∆P
Fig 3: Fuzzy logic controller model
NM
PS
P
Fuzzy set in V
NB
ZE
0.6
Fuzzy Rule Base
Defuzzifier
NS
0.8
0 -1
Fuzzy Inference Engine
NM
Table II: Rule base of FLC
∆V
∆P NB NM NS ZE PS PM PB
NB
NM
NS
ZE
PS
PM
PB
NB NB NM ZE PS PM PB
NB NM NS ZE PS PM PM
NM NS NS ZE PS PS PS
ZE ZE ZE ZE ZE ZE ZE
PS PS PS ZE NS NS NM
PM PM PS ZE NS NM NB
PB PM PS ZE NM NB NB
3
FLCs are used in different applications. Generally, a base of 25 rules was used due to its low computational cost, but it does perform so well. Therefore, in this research a base of 49 rules is used for better performance. To find an optimal MPP, P-V characteristics of PV panel are taken. Based on input parameters (∆P and ∆V), appropriate output signal (∆D) is generated. The output is a pulse width modulator signal (PWM) signal. The PWM signal (known as D) is fed to DC/DC boost converter to operate inverter. The crisp output is obtained by center of gravity defuzzification process [27]. n
D
D j 1 n
j
A
Initially, PV array, battery and utility grid are modelled in Matlab/Simulink. PMS is developed to ensure sustainable power flow. Similarly, the FLC is developed in build-in fuzzy toolbox with corresponding inputs, rule bases and output. The FLC MPPT modelled is compared with conventional PID based P&O algorithm. Output power of PV array with different controller is shown in figure 9. The black dotted line represents the reference power (PREF) which theoretical calculates the power of PV array. Similarly, red, blue and brown line represents the output power generated with FLC controller (PFLC), PID controller (PPID) and without controller (PWC). From figure 9, sunset at 5.5 Hrs, the PV array starts to generate power. Comparing FLC with PID at 12 Hrs, it is revealed from zoomed figure that there is a power difference of about 1 kW, 8 kW and above 15 kW between PREF and PFLC, PPID, PWC.
( Di ) (6)
A ( Di ) j 1
The error between the reference signal and power generated by FLC, PID and without controller is illustrated in figure 10. From figure 10, the error of FLC is about 1 kW for entire 24Hrs. Some spikes are also generated by FLC due to rapid change in irradiance level, but they are recovered very quickly. For PID controller, the maximum error difference is 8kW. Similarly, for without controller error difference is nearly constant at 18kW.
V. SIMULATION RESULTS AND DISCUSSION In this research, Islamabad, Pakistan region is taken as a case study. The weather data (solar irradiance and temperature) are recorded by Pakistan Metrological Department (PMD) for a typical summer day i.e., 22nd June, 2016 [28], [29]. The temperature and irradiance for a complete day is shown in figure 8.
40
40
30
37
600
34 Irradiance Temperature
400
31
200
20 10 Error
800
0 -10 -20
28
PREF-PFLC
-30 0
0
4
8
12 Time (Hrs)
16
20
25 24
-40
0
Fig 8: Weather data of a typical summer day (22 June 2016)
4
PREF-PPID
8
PREF-PWC
12 Time (Hrs)
16
20
24
Fig 10: Error in PV power with respect to reference power
250
PREF PFLC PPID
200
PWC Power (kW)
Irradiance (W/m2)
1000
42
Temperature (oC)
1100
150
255 250 245 240 235
100
11
11.5
12
12.5
50
0 0
4
8
12 Time (Hrs)
16
20
24
Fig 9: PV array output power with different controllers
4
concludes that that MPP of any PV systems find with fuzzy logic in shorter time runs compared to traditional control methods.
Average efficiency of proposed controller is calculated as;
ACKNOWLEDGMENT
T
ei
1 PREF (t ) Pi (t ) dt i FLC , PID,WC T 0
(7)
PREF ei % PREF
(8)
i
where 𝑃̅𝑅𝐸𝐹 is average reference power having value 112.82 kW. Using (7), average error (𝑒̅𝑖 ) for FLC, PID and WC are calculated 0.933 kW, 3.83 kW and 9.56 kW, respectively. Using average in (8), the efficiency of FLC, PID and WC is calculated as 99.18%, 96.6% and 91.52%, respectively. Hence, concluded that the performance of DC-DC boost converter is sufficiently improved by using FLC controller. According to IEEE Standard for Interconnecting Distributed Resources with Electric Power Systems [30], the maximum deviation allowed in grid voltage frequency, RMS and Total Harmonic Distortion (THD) level are 0.8%, 6% and 5%, respectively. In this research, the grid line-line voltage is 440V and frequency is 50Hz. Applying maximum deviation percentage, critical limits for voltage RMS and frequency are found as 466V, 424V, 50.4 Hz and 49.6 Hz, respectively. Figure 11 shows the grid voltage RMS, frequency and THD level. It is clearly stated from figure 11, that the power quality and system stability parameter (i.e., voltage, frequency and THD) are below the critical level. Voltage (V)
460
REFERENCES [1]
“Ministry of Planning Pakistan, fiscal year 2014-2015 Report,” 2015. [Online]. Available: www.pc.gov.pk/wpcontent/uploads/2015/06/Ch18-Energy.pdf.
[2]
“Renewable energy sources are provided to overcome energy crises,” Renewables International, 2015. [Online]. Available: http://www.renewablesinternational.net/renewable-energy-sourcesare-provided-to-overcome-energy-crisespart-1/150/435/84917/.
[3]
M. A. Sheikh, “Energy and renewable energy scenario of Pakistan,” Renew. Sustain. Energy Rev., vol. 14, no. 1, pp. 354–363, 2010.
[4]
“Quaid-E-Azam Solar Park,” 2016. [Online]. Available: http://www.qasolar.com/app/download/10612778599/Brief+on+100 MW+Solar+Power+Plant+Bahawalpur+%28Main+Body%29.pdf?t =1442084921.
[5]
“RENEWABLES 2016 GLOBAL STATUS REPORT,” 2016.
[6]
Y. Du and D. D.-C. Lu, “Battery-integrated boost converter utilizing distributed MPPT configuration for photovoltaic systems,” Sol. Energy, vol. 85, no. 9, pp. 1992–2002, 2011.
[7]
S. Lalouni, D. Rekioua, T. Rekioua, and E. Matagne, “Fuzzy logic control of stand-alone photovoltaic system with battery storage,” J. Power Sources, vol. 193, no. 2, pp. 899–907, 2009.
[8]
I. Purnama, Y.-K. Lo, and H.-J. Chiu, “A fuzzy control maximum power point tracking photovoltaic system,” in fuzzy systems (FUZZ), 2011 IEEE international conference on, 2011, pp. 2432–2439.
[9]
H. Patel and V. Agarwal, “MPPT scheme for a PV-fed single-phase single-stage grid-connected inverter operating in CCM with only one current sensor,” IEEE Trans. Energy Convers., vol. 24, no. 1, pp. 256–263, 2009.
[10]
F. Fangrui Liu, S. Shanxu Duan, F. Fei Liu, B. Bangyin Liu, and Y. Yong Kang, “A Variable Step Size INC MPPT Method for PV Systems,” IEEE Trans. Ind. Electron., vol. 55, no. 7, pp. 2622–2628, Jul. 2008.
[11]
G. Petrone, G. Spagnuolo, R. Teodorescu, M. Veerachary, and M. Vitelli, “Reliability Issues in Photovoltaic Power Processing Systems,” IEEE Trans. Ind. Electron., vol. 55, no. 7, pp. 2569–2580, Jul. 2008.
[12]
T. Esram and P. L. Chapman, “Comparison of Photovoltaic Array Maximum Power Point Tracking Techniques,” IEEE Trans. Energy Convers., vol. 22, no. 2, pp. 439–449, Jun. 2007.
[13]
G. M. S. Azevedo, M. C. Cavalcanti, K. C. Oliveira, F. A. S. Neves, and Z. D. Lins, “Comparative evaluation of maximum power point tracking methods for photovoltaic systems,” J. Sol. Energy Eng., vol. 131, no. 3, p. 31006, 2009.
[14]
R. Faranda, S. Leva, and V. Maugeri, “MPPT techniques for PV Systems: Energetic and cost comparison,” in 2008 IEEE Power and Energy Society General Meeting - Conversion and Delivery of Electrical Energy in the 21st Century, 2008, pp. 1–6.
[15]
E. M. Ahmed and M. Shoyama, “Scaling factor design issues in variable step size incremental resistance MPPT in PV systems,” in 2011 IEEE Ninth International Conference on Power Electronics and Drive Systems, 2011, pp. 889–894.
[16]
W. Weidong Xiao and W. G. Dunford, “A modified adaptive hill climbing MPPT method for photovoltaic power systems,” in 2004
440
420
0
2
4
6
8
10
12
14
16
18
20
22
24
50.2
50
49.8
THD %
Frequency (Hz)
The research work is supported by National Natural Science Foundation of China (No.51377184), International Science & Technology Cooperation Program of China (No.2013DF G61520) and Fundamental Research Funds for the Central Universities (No.CDJZR12150074).
0
2
4
6
8
10
12
14
16
18
20
22
24
4 2 0
0
2
4
6
8
10
12
14
16
18
20
22
24
Time (Hrs)
Fig 11: Utility grid voltage RMS, frequency and THD level.
VI. CONCLUSION The fuzzy controlled PV/Battery hybrid power system is simulated to supply power to the grid. Clearly, the proposed fuzzy controller is performed better in transitional states than with/without perturbation and observation PID based MPPT methods. The performance of the proposed controller is validated by the simulations done in Matlab/Simulink under real-world record of weather conditions. The proposed method
5
IEEE 35th Annual Power Electronics Specialists Conference (IEEE Cat. No.04CH37551), 2004, vol. 3, pp. 1957–1963. [17]
A. M. Z. Alabedin, E. F. El-Saadany, and M. M. A. Salama, “Maximum power point tracking for Photovoltaic systems using fuzzy logic and artificial neural networks,” in 2011 IEEE Power and Energy Society General Meeting, 2011, pp. 1–9.
[18]
H. Jafarnejadsani, J. Pieper, and J. Ehlers, “Adaptive Control of a Variable-Speed Variable-Pitch Wind Turbine Using Radial-Basis Function Neural Network,” IEEE Trans. Control Syst. Technol., vol. 21, no. 6, pp. 2264–2272, Nov. 2013.
[19]
R.-E. Precup, M. L. Tomescu, M.-B. Rădac, E. M. Petriu, S. Preitl, and C.-A. Dragoş, “Iterative performance improvement of fuzzy control systems for three tank systems,” Expert Syst. Appl., vol. 39, no. 9, pp. 8288–8299, 2012.
[20]
A. A. Kulaksız and R. Akkaya, “A genetic algorithm optimized ANN-based MPPT algorithm for a stand-alone PV system with induction motor drive,” Sol. Energy, vol. 86, no. 9, pp. 2366–2375, Sep. 2012.
[21]
K. Ishaque, Z. Salam, M. Amjad, and S. Mekhilef, “An Improved Particle Swarm Optimization (PSO)–Based MPPT for PV With Reduced Steady-State Oscillation,” IEEE Trans. Power Electron., vol. 27, no. 8, pp. 3627–3638, Aug. 2012.
[22]
A. H. Besheer and M. Adly, “Ant colony system based PI maximum power point tracking for stand alone photovoltaic system,” in 2012 IEEE International Conference on Industrial Technology, 2012, pp. 693–698.
[23]
T. Kamal and S. Z. Hassan, “Energy Management and Simulation of Photovoltaic/Hydrogen/Battery Hybrid Power System,” Adv. Sci. Technol. Eng. Syst. J., vol. 1, no. 2, pp. 11–18, 2016.
[24]
S. Z. Hassan, H. Li, T. Kamal, S. Mumtaz, L. Khan, and I. Ullah, “Control and Energy Management Scheme for a PV/SC/Battery Hybrid Renwable Power System,” Sci. Int., vol. 28, no. 2, pp. 955– 964, 2016.
[25]
S. Z. Hassan, S. Mumtaz, T. Kamal, and L. Khan, “Performance of grid-integrated photovoltaic/fuel cell/ electrolyzer/battery hybrid power system,” in 2015 Power Generation Systems and Renewable Energy Technologies, PGSRET 2015, 2015, pp. 1–8.
[26]
K. ERDAL, Y. ŞABAN, and R. Ö. HASAN, “Maximum power point tracking using fuzzy logic control,” Turkish J. Electr. Eng. Comput. Sci., vol. 24, no. 6, pp. 1–28, 2016.
[27]
S. Venkatanarayanan and M. Saravanan, “Fuzzy logic based PV energy system with SEPIC converter,” J. Theor. Appl. Inf. Technol, vol. 59, pp. 89–95, 2014.
[28]
“Weather Data Islamabad,” 2016. https://www.wunderground.com/history/.
[29]
“Solar Irradiance Data Islamabad,” 2016. [Online]. Available: http://www.pveducation.org/pvcdrom/properties-ofsunlight/calculation-of-solar-insolation.
[30]
“IEEE Standard for Interconnecting Distributed Resources with Electric Power Systems,” IEEE Std 1547-2003. pp. 1–28, 2003.
[Online].
Available:
6
