CC BY-NC-ND
0Available online at www.sciencedirect.com
ScienceDirect
Energy Procedia 6 (2011) 633-642
MEDGREEN 2011-LB
Fuzzy Logic Control for the tracking of maximum power point of a PV system
F.Bouchafaaa*, I.Hamzaouia, A.Hadjammara
aLaboratory of Instrumentation, Faculty of Electronics and Computer, University of Sciences and Technology Houari Boumediene,
BP 32 El-Alia 16111 Bab-Ezzouar Algiers, Algeria.
Abstract
Tracking of the maximum power point (MPPT) plays an important role in photovoltaic (PV) power systems because they maximize the power output from a PV system for a given set of conditions, and therefore maximize they array efficiency. This work presents a comparative study between different control strategies used most conventional digital namely perturbation and observation (P & O) and incremental Conductance (INC) with digital control by fuzzy logic (FLC). The introduction of fuzzy controller as a solution has given very good performance and whatever the parametric variation of the system.
©2010 Published by Elsevier Ltd. Selection and/or peer-review under responsibility of [name organizer]
Keyword: Photovoltaic, Converter DC/DC, MPPT, Fuzzy logic control,Perturbation and Observationjincremental Conductance.
1. Introduction
A photovoltaic generator can operate over a wide range of voltage and current output, but in case you want to maximize the energy produced (connected to UPS, battery charger), it is interesting to include a search of maximum power point in converters [1].
In fact the I (V) depend on the solar irradiance and temperature. These climatic variations result in fluctuations in the maximum power point. Because of these fluctuations, it often inserts one or more controlled static converters for the furtherance of the maximum power point. These commands are known
* Corresponding author. Tel.: (213) 664 916 650; fax: 213 (0) 21 24 71 87. E-mail address: fbouchafa@gmail.com
1876-6102 © 2011 Published by Elsevier Ltd. doi:10.1016/j.egypro.2011.05.073
as MPPT (Maximum Power Point Tracking) associated with the choppers, which provides coupling between the PV array receivers, forcing the first to deliver its maximum power [1].
This area makes up our days the subject of extensive research to improve the command to the pursuit of maximum power point.
In this work we present both numerical methods used to learn classical perturbation and observation (P & O) and incremental conductance (INC) and the numerical method using fuzzy logic (FLC).
A comparative study between these different methods of MPPT control to determine the most efficient and most robust to changes in climate: temperature and irradiance.
2. Electrical model of photovoltaic cell
To find the model of the photovoltaic generator, we must first find the electrical equivalent to that source. Many mathematical models have been developed to represent their highly nonlinear behaviour resulting from that of semiconductor junctions that are the basis of their achievements. Found in the literature several models of different photovoltaic generators them through the procedure and the number of parameters involved in the calculation of voltage and current end-of photovoltaic generator. We will present our work in the model with two diodes; in fact this model takes into account the different internal resistance of the PV cell (Fig.1) [1]. It consists of a current source iph that models the conversion of light energy flow electrical resistance Rsh shunt is a consequence of leaks by the side effect on the photovoltaic cell, a series resistance Rs, representing the various resistance contact and connection and two diodes D1 and D2 in parallel model the PN junction [2].
tv iphQyy
■VWW
Fig.1. Model of a photovoltaic cell with two diodes
The current generated by the module is given by the following equation:
I = Iph - Id1 - Id2 - IRP (1)
ph d1 d2 aRP
I = Iph-IS1-|expq4^ -1]-IS2-i expq4^i) -1]-
Where V and I represent the output voltage and current of the PV; q is the electronic charge; Iph corresponds to the light-generated current of the solar array. IS1,2 represent the current saturation of the two diodes; Ab2is ideality factor of the junction of Di and D2, K the Boltzmann's constant, T the cell temperature.
From equation (1) we notice that the output current of the PV module depends on the photocurrent itself, which itself depends on the solar insulation and the junction temperature of the cells of the module,
consequently the power which a module can deliver depends on the solar insulation and the temperature of the junction [2,3].
We have implemented the model of photovoltaic generator consists of 36 cells in series. in the environment "MATLAB / SIMULINK".
For a temperature T=25°C and an irradiance E=1000W/m2, we obtained the characteristics of a PV cell of changes in current I = f (V) and power P = f (V) based on the voltage of the PV cell is shown in figure 2.
Fig.2. Power and current voltage characteristics
3. PV generator in terms of variables
The electrical characteristic of a PV array varies according to temperature, illumination, its internal parameters, and generally the nature of the connected load. We have simulated the behavior of the generator under various constraints. These concepts are indeed necessary to understand the behavior of a PV array.
We vary the illumination between 400W/m2 1000W/m2 and a constant temperature of 25°C. The influence of illumination on the I = f (V) and P = f (V) is shown in figure 3.
KA) P{W}
Fig.3.Influence of illumination on the characteristic I = f (V) and P = f (V).
The variation of the illumination, we note that for a temperature 25°C, the increase in irradiance leads to an increase in maximum power and a slight increase in open circuit voltage. The short circuit current increases dramatically with increasing illumination. This implies that the optimal power generator is almost proportional to the illumination.
By varying the temperature between -10°C and 60°C under an irradiance of 1000W/m2, we can see the influence of temperature on the characteristics I = f (V) and P = f (V).
V(V) V(V)
Fig.4. Influence of temperature on the characteristic I = f (V) and P = f (V).
The open circuit voltage decreases significantly with increasing temperature as the maximum power. By constant, we notice a slight increase in short circuit current with increasing temperature.
For a temperature change, we deduce that the voltage changes significantly while the current remains constant. To get a maximum return, it is essential to work in the area of maximum power of the PV generator. For this, we used a research strategy places in this area [4]. Hence the need to introduce a power converter which will play the role of load-source adapter.
4. Principle of tracking the point of maximum power (MPPT)
The intercalation of a static converter DC/DC, as shown in figure 5, changes the operating point of the panel through an external control law in order to maximize the energy transferred permanently.
Fig.5. Line of the photovoltaic conversion
Most methods of tracking maximum power point based on the power-voltage characteristic of photovoltaic energy [6].
Different control algorithms exist, we present in this paper a comparative study between different conventional MPPT most used, namely perturbation and observation (P & O) and Incremental Conductance (INC) and the numerical method by fuzzy logic.
4.1. Method 'perturb & Observe'
It is the continuation of the MPP algorithm most commonly used. And as its name suggests, it is based on the perturbation of the system by the increase or decrease of the cycle, then observing the effect on the output panel [6].
If the value of the actual power P(k) of the panel is higher than the previous value P(k-1), so we keep the same direction of previous disturbance if they reverse the disruption of the previous cycle.
The flowchart of the algorithm of P & O is given in figure 6.
The main drawback of this method is the loss of power depends on the width of no disruption C.
If the width is not large, the algorithm of MPPT respond quickly to sudden and rapid changes in operating conditions but will incur losses in slowly changing conditions and the stable states.
If the width of the pitch is very small losses in stable states or slowly changing conditions will be reduced, but the system has a slow response to rapid changes in temperature or sunshine.
The value for the ideal width of the system can not be determined experimentally or by simulation, thus satisfying a compromise between fast response and power loss in steady states [5].
A disadvantage of the method of P & O is described by [7]. If a sudden increase in sunlight is produced there will be an increase of the power panel, the previous algorithm behaves as if this increase is produced by the effect of previous disturbance, then it continues in the same direction which is a wrong direction, away from what the real point of maximum power. This process continued until the stability of the sun or it is the true point of maximum power. This causes a response delay when sudden changes in operating conditions and power losses.
4.2. Method'incrementalconductance'
The algorithm, the incremental conductance, resulting from the partial derivative of the relative power of the output voltage of the photovoltaic panel to achieve the maximum power point characterized by a value dP/dV=0. The following equations characterize this method:
dL- d(VI) _ I + v^L_ o
dV dV dV (2)
_ _L _
, V " dV
The term I/V represent the conductance and dI/ dV the incremental conductance. At maximum power point these two terms are equal but different signs [6]. The flowchart of this algorithm is shown in figure 7.
The main advantage of this algorithm is its speed relative to P & O. but it is more complex to implement for simulation.
Fig.6. The flowchart of the algorithm P & O. Fig.7. The flowchart of the algorithm Incremental Conductance.
4.3. MPPT fuzzy logic-based
Conventional methods of tracking the optimal point of operation have shown their limits to sudden changes of weather and the load connected to the panel, several methods have emerged to try to alleviate these shortcomings and improve the operation of these generators.
The approach of Artificial Intelligence in the case of fuzzy logic is implemented to improve control performance and the pursuit of maximum power point by simulation and modeling of a controller based on fuzzy logic [8].
The advent of microcontrollers has enabled the spread of fuzzy control in the pursuit of optimal point during the last decade [4].
The fuzzy controller has the following three blocks:
Fuzzification of input variables by using the trapezoidal and triangular functions, then these variables fuzzification inference or are compared with pre-defined packages to determine the appropriate response. And finally the defuzzification to convert the subset fuzzification in values using the centroid defuzzification.
The five linguistic variables used are: NB (Negative Big), NS (Negative Small), ZE (Zero Approximately), PS (Positive Small), PB (Positive Big) [8].
The two FLC input variables are the error E and change of error CE at sampled times k defined by [3]:
= P(k) - P(k - 1)
( ) V(k) - V(k -1) (3)
CE(k) = E(k) - E(k - 1)
Where P(k) is the instantaneous power of the photovoltaic generator.
The input E(k) shows if the load operation point at the instant k is located on the left or on the right of the maximum power point on the PV characteristic, while the input CE(k) expresses the moving direction of this point.
The fuzzy inference is carried out by using Mamdani's method, (Table 1) [4], and the defuzzification uses the centre of gravity to compute the output of this FLC which is the duty cycle:
¿(d^)- daj (4)
da = j-1 n-
£ V(daj)
The control rules are indicated in Table 1 with E and CE as inputs and da as the output.
Tablel. Fuzzy rule table
eCE NG NP ZE PP PG
NG ZE ZE PG PG PG
NP ZE ZE PP PP PP
ZE PP ZE ZE ZE NP
PP NP NP NP ZE ZE
PG NG NG NG ZE ZE
These two variables and the control action a for the tracking of the maximum power point are illustrated in figure 8 [4].
loo -er> a 4D SD 1 DO
Mfîhûerslqt fttnrii-tlt
0-Q3I? .£.016 .O.DOS O D.oce O.Olö 0.032
Fig.8. Membership for inputs and outputs
4.4. Simulation results and interpretations
In this section, a simulation with MATLAB/ SIMULINK MPPT of a photovoltaic panel of 36 cells with two exponential connected to a storage battery through a chopper is used with P & O, INC. and FLC.
The comparative study between the three methods for tracking MPPT standard conditions (E=1000W/m2 and T=25°C) and variable climatic conditions.
Figure 9 shows that the FLC is faster than the controller based on classical numerical algorithms (P & O and INC) even if the INC has shown good speed.
The fuzzy controller has been very good improvements against the ripples in steady, he can eliminate them. While the NCI has shown that it provides less power loss.
The MPPT fuzzy logic control has better performance compared to each other at the time of response and stability.
Fig.9. Variation of different sizes of weather generator in constant
To test the temperature variation, we perform a rapid increase of 25°C to 60°C and irradiance with 1000W/m2 during 2s, as shown in figure 10.
Increasing the temperature always involves a decrease in power. The fuzzy controller has a response almost perfect continuation algorithm while P & O and INC are late and they present some fluctuations. On found that both control strategies in MPPT P & O and INC have ripples against the method by fuzzy logic presents better results and without undulations reflecting the non-sensitivity to temperature variations.
We also note that the fuzzy MPPT controller is faster. Losses due to the oscillations are very small when using this controller.
Figure 11 shows the behavior of the system, a variation of insulation of 800W/ m2 at 1000W/m2 over an interval of two seconds with a constant temperature T = 25°C.
The results show that the fuzzy controller following the deposit with less fluctuation, while the INC gives good speed performance. As an algorithm P & O, the transitional regime is characterized by a distance of PPM which explains the delay to reach the new value of power.
On found that both strategies MPPT control at P & O and INC are very affected by this variation they exhibit significant differences to achieve the MPPT by cons method by fuzzy logic presents better results and without undulations reflecting the not sensitive to changes in illumination.
Fig. 11. Response of the system in a variation of the illumination
5. Conclusion
The different results with different robustness test confirms the proper fonctionnement of fuzzy controller with good performance in the atmospheric variations of illumination and temperature thereby reducing power losses, with better dynamics than conventional numerical methods.
The following fuzzy controller with satisfaction at the sharp variations of temperature and illumination and a fast response time and less than that of conventional algorithms (P & O and INC). This eliminates the fluctuations in the power, voltage and duty ratio in steady state.
The controllers by fuzzy logic can provide an order more effective than the traditional controllers for the nonlinear systems, because there is more flexibility.
A fast and steady fuzzy logic MPPT controller was obtained. It makes it possible indeed to find the point of maximum power in a shorter time runs.
6. References
[1] F.Belhachat, C. Larbes, L. Barazane, S. Kharzi, "Commande neuro-floue d'un hacheur MPPT", 4éme conférence internationale "Computer Integrated Manufacturing", CIP'07, 03-04 Novembre 2007.
[2] Y. Pankow, "Étude de l'intégration de la production décentralisée dans un réseau basse tension. Application aux générateurs photovoltaïques", Thèse de doctorat Centre national de recherche technologique de Lille 2005.
[3] M. Azab, "A New Maximum Power Point Tracking for Photovoltaic Systems", Procedings of World Academy of Science, Engineering and Technology Volume 34 October 2008 ISSN 2070-3740.
[4] C. Ali, "Étude de la Poursuite du Point de Fonctionnement Optimal du Générateur Photovoltaïque", 3rd International Conference Sciences of Electronic, Technologies of Information and Telecommunications March 27-31, 2005 - TUNISIA.
[5]M. Hatti, "Contrôleur Flou pour la Poursuite du Point de Puissance Maximum d'un Système Photovoltaïque", JCGE'08 LYON, 16 et 17 décembre 2008.
[6] C. Liu, B. Wu and R. Cheung, "Advanced Algorithm for MPPT Control of Photovoltaic Systems", Canadian Solar Buildings Conference Montreal, August 20-24, 2004 Refereed Paper.
[7] R.W. Erickson, Fundamentals of Power Electronics, Chapman & Hall, 115 Fifth Avenue, New York, NY, 1997.
[8] N. Patcharaprakitia, and al. "Maximum power point tracking using adaptive fuzzy logic control for grid-connected photovoltaic system", in IEEE Power Eng. Society Winter Meeting,2002, pp. 372-377.
[9] M. A. S. Masoum, M. Sarvi, "Design, Simulation and Implementation of A Fuzzy-Based MPP Tracker under Variable Insolation and Temperature Conditions", Iranian Journal of Science & Technology, Transaction B, Engineering, Vol. 29, No. B1, Shiraz University 2005.
