Design of MPPT Controller using ANFIS and HOMER based sensitivity analysis for MXS 60 PV module

Transcription

1 Design of MPPT Controller using ANFIS and HOMER based sensitivity analysis for MXS 6 PV module A.Padmaja Asst.Professor, EEE, JNTUK-UCEV Vizianagaram, AP, India M.Srikanth PG Student, EEE, JNTUK-UCEV Vizianagaram, AP, India Abstract This paper presents an intelligent control technique for the Maximum Power Point Tracking (MPPT) of a photovoltaic (PV) system using adaptive neuro-fuzzy inference system (ANFIS) under variable solar irradiation conditions. The MXS 6 PV Module specifications is considered for the analysis and models of solar PV module and a DC/DC Boost converter are developed in MATLA/SIMULINK environment. Initially, an MPPT controller is designed using Perturb and Observe (P&O) method. However, this conventional method cannot track rapid changes in the solar irradiation effectively. Hence, an intelligent controller is designed using ANFIS which draws much energy and fast response under continuously changing operating conditions. The PV module with proposed MPPT controller is analyzed in stand-alone mode. The major disadvantage with PV system is its uncertain and intermittent power output which depends on weather conditions. PV module alone cannot supply reliable power to the isolated load effectively. To overcome this, PV module can be connected to the grid. It serves two purposes; in case of deficit solar irradiation, power can be taken from the grid and when there is surplus irradiation, power can be fed to the grid. In order to predict the power supplied to the load and grid under different operating conditions sensitivity analysis has been carried out for the PV system with designed MPPT controller is simulated using HOMER Pro Software. Keywords Maximum power point tracking (MPPT), Adaptive neuro fuzzy inference system (ANFIS), photovoltaic module(pv), Perturb and Observation (P&O) technique, PI controller. I. INTRODUCTION Solar energy is one of the most promising renewable resources that can be used to produce electric energy through photovoltaic (PV) system [], which directly converts the energy in photons of light to electrical energy. The output power of solar panel depends on solar irradiance and temperature. As the load impedance depends on application, a dc-dc converter is used for improving the performance of solar panel. A typical solar PV module converts only 3 to 4 percent of the incident solar irradiation into electrical energy. To improve the efficiency of solar PV system and enhance the power delivered to the load, maximum obtainable power from the solar PV module has to be extracted. Maximum power point tracking (MPPT) techniques are used to extract the maximum power from the solar PV module and transferring that power to the load which in turn improves the efficiency of the overall PV system [2]. Due to irregular patterns of solar irradiance, temperature and the load impedance, the current-voltage-power (I-V-P) characteristic of PV systems is nonlinear. In general, there is a unique point on the V-P curve, called the Maximum Power Point (MPP), at which the entire PV systems (array, converter, etc.) operates with maximum efficiency and produces its maximum output power [3]. Maximum power from the solar PV system is obtained only when irradiance is present. Absence of irradiance makes difficulty to supply required load demand. Thus to overcome this drawback, solar PV system supplying isolated load is connected to the grid. Thus gird connected systems, sensitive analysis forecasts the power flow through solar PV system to the Grid during surplus generation and vice-versa during deficient PV generation. This paper is organized in the follows: Section 2 presents the PV module and its characteristics and Section 3 describes the MPPT technique using P&O and design of an intelligent MPPT controller using adaptive neuro-fuzzy inference system (ANFIS). Simulation results for the case study and sensitivity analysis for grid connection by HOMER pro is presented in the section 4. II. MODELING OF SOLAR PV MODULE In this section PV equation modeling is presented. The key specification of the MXS 6 PV module considered for this study is shown in Table. In addition to that, series resistance (Rs) as.2ω, band gap energy (Eg) of the semiconductor is taken as. ev and ideality factor (A) of semiconductor is taken as.6 are considered for the modeling [4]. A. Equivalent circuit of PV model To know the electronic behavior of a solar cell, the ideal equivalent electrical circuit of a solar cell can be represented by a light generated current source, I ph, in parallel with a single-diode as shown in Fig. In practice no solar cell is ideal, so a shunt and series resistance component is added to the model. In practice R sh is very large and that of R s is very small. The output power of a solar cell is very less hence to increase the output power of solar PV systems, the solar cells are connected in series and parallel configurations to form solar PV modules and array [5]. 24-5, IJIRAE- All Rights Reserved Page -4

2 . TABLE Key specification of MXS 6 PV module Parameter Variable Value Maximum power P m 6W Maximum Voltage V m 7.V Current at max power I m 3.5A Open circuit voltage V oc 2.6V Short circuit current I sc 3.74A Total No.of cells in Series N s 36 Total No.of cells in Parallel N p B. Basic Equations of PV Model Fig. Single Diode Equivalent Circuit Model for PV system. The non-linear relationship between current-voltage of the PV module can be described mathematically using basic equations from the theory of semiconductors and photovoltaic. ) PhotoCurrent: The irradiance incident on the solar PV panel results in the flow of current in the system called as Photocurrent. It depends on the irradiation and also the temperature. The linear relationship can be attributed as follows: I = I + K (T -T ) i * λ (2.) ph sc r where I sc is the cell short circuit current K i is the current temperature coefficient T is the cell working temperature T r is the reference temperature λ is the solar irradiance 2) Reverse Saturation Current: The reverse saturation current I rs of module at reference temperature is given as I sc Irs = (2.2) q*v oc exp - NsKAT where V oc is the open circuit voltage q is the electron charge=.6 x -9 C K is the Boltzman Constant=.38 x -23 J/K N s is the no.of cells connected in series. Reverse saturation current depends on the temperature variations. 3) Module Saturation Current: The module saturation current is defined as I s is depend up on temperature and is given by 3 - T Tr T I s = Irs exp q* Eg (2.3) Tr KA where I rs is the Reverse Saturation Current. E g is the Energy Band Gap of Semiconductor. 4) PV Output Current: The output current of the PV module is given by q( V IR s I N p * I ph N p * I s exp (2.4) N s * kat 24-5, IJIRAE- All Rights Reserved Page -4

3 where R s is the series resistance of the solar cell V is the output voltage of the cell. N p is the no.of cells in parallel. The solution of the equation (2.4) can be found out using iterative process and is need to solve an algebraic loop in the output current. C. Analysis of PV Module Characteristics Solar PV Module Characteristics Current Versus Voltage and Power Versus Voltage Curves of Solar PV Module characteristics are shown below: Current (A) W/sqm 5 W/sqm 7 W/sqm 9 W/sqm Voltage(V).5 Fig.2 I-V Characteristics of PV Module With Variable Irraidance Current (A) Voltage(V) Fig.3 I-V Characteristics of PV module with variable Temperature W/sqm 5 W/sqm 7 W/sqm 9 W/sqm Power (W) 3 2 Fig.4 P-V Characteristics of Solar PV Module The characteristics shown in Fig.2 revels the plot against Voltage Versus Current with Constant Temperature and variable Irradiance. In this, the current rapidly varied while the voltage varied moderately. This results net increase in output power with increase in the Irradiance level. In Fig.3 the plot shows Constant Irradiance and Variable Temperature. In this, as the temperature increases the output current increases marginally while the output voltage decreases drastically. There is a net reduction in the output power. The characteristics in Fig.4 show that for increased irradiance level the power generated by the solar PV module increases. A. Need for MPPT Voltage(V) III. CONTROL STRATEGY A solar panel converts only 3% to 4% of the incident solar irradiation into electrical energy. Maximum power point tracking technique is used to enhance the efficiency of the solar panel. According to Maximum Power Transfer theorem, the power output of a circuit is maximum when the source impedance matches with the load impedance. Hence problem of tracking the maximum power point transformed to an impedance matching problem. In the source side a DC-DC boost convertor is connected to a solar panel in order to enhance the output voltage, so that it can be used for different applications. By changing the duty cycle of the boost converter appropriately, the source impedance can be matched with the load impedance. There are many MPPT techniques available to track maximum power from a solar module [6]-[7]. In this paper P&O and ANFIS techniques are implemented and compared. B. Perturb & Observation The perturb-and-observe method is the simplest method. It is also known as perturbation method. This is most commonly used in commercial PV products. This is essentially a trial and error method. In this only voltage sensor is used to sense the PV array voltage. Thus the implementation cost is less and hence easy to apply. The time complexity of this algorithm is very less. 24-5, IJIRAE- All Rights Reserved Page -42

4 ) P&O Algorithm: The Perturb & Observe algorithm states that when the operating voltage sensed from PV panel is perturbed by a small increment, if the resulting change in power P is positive, then the direction of MPP perturbing in the same direction. If P is negative, the direction of MPP and the sign of perturbation supplied have to be changed TABLE 2 - P&O Technique Perturbation Change in Power Next Perturbation Positive Positive Positive Positive Negative Positive Negative Positive Negative Negative Negative Positive The Fig.5 shows the module output power versus module voltage for a solar panel at a given irradiation. The point marked as MPP is the Maximum Power Point, the theoretical maximum output obtainable from the PV panel. Consider A and B as two operating points. As shown in Fig.5 the point A is on the left hand side of the MPP. Move towards the MPP by providing a positive perturbation to the voltage. On the other hand, point B is on the right hand side of the MPP. When give a positive perturbation, the value of power becomes negative, thus it is imperative to change the direction of perturbation to achieve MPP. The flowchart for the P&O algorithm is shown below Fig.6. Power (W) A MPP B Voltage (V) Fig.5 Solar PV Characteristics Showing MPP and Operating Points A and B. Start P&O Algorithm Measure V(k), I(k) Calculate power P(k)= V(k)*I(k) P=P(k)-P(k-) No P> Yes Yes V(k)-V(k-)> No No V(k)-V(k-)< Yes Decrease Array Voltage Increase Array Voltage Decrease Array Voltage Increase Array Voltage Update History V(k)-V(k-) P(k)>P(k-) Fig. 6 Flow chart representing P&O technique 24-5, IJIRAE- All Rights Reserved Page -43

5 C. ANFIS (Adaptive neuro fuzzy inference system): ANFIS is a learning technique with data that uses Fuzzy Logic to transform given inputs into a desired output through highly interconnected Neural Network, which are weighted to map the numerical inputs into an output. ANFIS combines the benefits of the two machine learning techniques (Fuzzy Logic and Neural Network) into a single technique. An ANFIS works by applying Neural Network learning methods to tune the parameters of a Fuzzy Inference System (FIS) [8]. Using a given input/output mapping data, the ANFIS toolbox constructs a fuzzy inference system (FIS), whose membership function parameters are adjusted using either back-propagation algorithm or combination of back propagation algorithm and least squares type of method. This process of learning is called Hybrid Learning technique. This allows fuzzy systems to learn from the data they are modeling. ) Procedure for : To track the MPPT with ANFIS require input and output data sets [9]. These input and output data sets are taken from the system operating constraints. There are two possible ways to collect training data. One is collecting data from the real-time system, another one is from simulation by developing an accurate dynamic model for PV module []. Collecting data from the real-time system was very difficult due to the irregular nature of weather and the inability to control the weather conditions. Therefore, the training data were collected in this work from simulation after the development of dynamic PV module [4]. The proposed take the operating temperature and irradiance as input and predicate the maximum output power from PV module at that instant [3]. At the same operating temperature and irradiance, the actual output power from the PV module is calculated by sensing operating voltage and current [9].The predicated power and calculated powers are compared and the error is given to a proportional integral (PI) controller, to generate operating signals [9]. The operating signal generated by the PI controller is given to the PWM generator. The PWM signal is generated using high frequency of carrier signal as compared to the operating signal. The frequency of carrier signal used is 25 khz. The generated PWM signals manage the duty cycle of DC DC converter, in order to adjust the operating point of the PV module. IV. SIMULATION RESULTS The PV module considered is of MSX6 type. This module is used as test module to test the proposed ANFIS based MPPT Control Scheme with Conventional P&O technique for system supplying isolated load. ANFIS based control scheme is simulated using GUI (Graphical user interface) in the Matlab /Simulink for the test system. The operating temperature is varied from 5 C to 65 C in a step change of 5 C and the solar irradiance is varied from W/m 2 to W/m 2 in a step change of 5W/m 2. These data is used to train the ANFIS controller. Total 29 training sample data sets and 2 epochs are used to train the ANFIS. By using given input/output sample data set, the ANFIS constructs a fuzzy inference system (FIS). The membership function parameters are adjusted using the hybrid optimization method of training the FIS [2]. Fig.7 shows the training data and ANFIS output. The structure of ANFIS, generated by the Matlab code is a five layer network as shown in Fig.8. It has two inputs (irradiance level and operating temperature), one output and three member- ship functions for each input. Nine fuzzy rules are derived from six input membership functions. These rules are derived according to the input and output mapping, so as to construct maximum output power for each value of input temperature and irradiance level. Fig.9 shows output of fuzzy rule for a specific value of operating temperature and irradiance level. The ANFIS generated surface is shown in Fi.. It is three-dimensional plot between temperature, irradiance and maximum power. The ANFIS surface indicates that the maximum available power from PV module increases with increase in Irradiance level and moderate temperature []. Fig. 7 Training Data and ANFIS Output. Fig.8 Structure of ANFIS. 24-5, IJIRAE- All Rights Reserved Page -44

6 Fig. 9 Output from Fuzzy Rules for specific values of temperature and irradiance. Fig. ANFIS Surface A. Case Study: The solar PV system modeled simulated by varying irradiances keeping temperature constant. The test cases applied to the system are as follows: Case (i): Proposed System supplying isolated load by varying irradiances. Case (ii): Proposed System Connected to the Grid with varying Irradiances and varying load. Case (i):stand-alone PV System: In this section the simulation results for PV module tracking maximum power is compared with the conventional (P&O) and ANFIS techniques. The analysis is made for constant and varying irradiances. The Comparisons plots for Power, Voltage and Current with respect to time for constant irradiations are shown in below Fig. to Fig.6 5 Power(W) Fig. Comparison of power outputs for 5 W/m 2 with proposed MPPT Controller. 5 P o w e r ( w ) Fig.4 Comparison of power outputs for 7W/m 2 with proposed MPPT Controller. 2 V o l t a g e ( V ) 5 V o lt a g e ( v ) Fig. 2 Voltage Comparison for Irradiance 5 W/m Fig.5 Voltage comparison for Irradiance 7 W/m , IJIRAE- All Rights Reserved Page -45

7 C u r r e n t ( A ) Fig. 3 Current comparison for Irradiance 5 W/m 2 C u r r e n t ( A ) Fig.6 Current comparison for Irradiance 7 W/m 2 iance The Comparisons plots for Power, Voltage and Current with respect to time for variable irradiations are shown in below Fig.8 to Fig.2 I rrad ia n c e(k w / m 2 ) Fig.7 Variable Irradiance. P o w e r( W ) Without Controllwer Fig.8 Comparison of power outputs for different controllers V o lt a g e (v ) 5 Wituout Controller C u rre n t (A ) Without Controllwer Fig.9 Voltage Comparison for variable irradiance Fig.2 Current Comparison for variable irradiance TABLE 2 Numerical analysis for stand-alone system S.NO Irradiance o/p power (Kw/m 2 ) From Fig. to Fig.2 and Table 2, it can be observed that both P&O and can track the maximum power nearly equal but settling time and response of ANFIS faster than conventional technique. 24-5, IJIRAE- All Rights Reserved Page -46

8 Case (ii): Grid Connected Solar PV system The irradiances is not available throughout the day. Hence in order to meet the load demand the system is connected to the grid. The proposed system capacity is improving by taking more such modules to interconnect with the grid by varying the load and having varying irradiances. The characteristics plot obtained for the system is show below: I r r a d i a n c e ( K W / m 2 ) Fig 2 Varying Irradiance P o w e r(w ) Fig 22 Power Comparison with different controllers 5 45 C u rre n t (A ) Fig.23 Current Comparison of PV Module.5 x 4 V o l t a g e ( v ) Fig.24 Voltage Comparison of PV Module p o w e r ( W ) Fig.25 Daily varying load From the plots Fig.2 to Fig.25 it is clear that solar output is varying with respect to incident irradiance even though the laod is varies the MPPT controllers can track the maximum power at the instant. By analyse the grid connected model the irradiance is not present through out the day but load is present. For continuity in supply the PV system is colabarate with grid and the power transitions is estimated by sensitivity analysis by HOMER pro. B. Sensitivity analysis for grid connected system Sensitivity analysis of the solar based system connected to the Grid is simulated using HOMER PRO Software. The analysis made in this system is related to predicting of power transitions under all conditions. The specifications of the system considered for the analysis is shown in Table 3 and Table , IJIRAE- All Rights Reserved Page -47

9 TABLE 3 Meteorological Data at JNTUK UCE Vizianagaram. Months Irradiance (Kwh/m 2 /d) January 4.75 February 5.57 March 5.88 April 6.64 May 6.5 June 5.2 July 4.53 August 4.6 September 4.67 October 5.24 November 4.59 December 4.59 Annual Mean 5.9 TABLE 4 Solar PV system Parameters Maximum Power (KW) 7.9 Total Capital Cost /Kw ($) 3 Replacement cost ($) 3 Annual maintenance cost ($) Life time (yrs) 25 ) Grid connected solar PV system The model considered for the Grid connected PV system supplying a load demand as shown in Fig.26. Fig.26 Grid connected PV system This analysis shows the various operational characteristics such as the load demand met by the solar PV system and the amount of the load deficient that has to be met by the grid connected system. These operational characteristics will help in predicting the safe operation of the power systems. 24-5, IJIRAE- All Rights Reserved Page -48

10 Load Demand Solar Irradiance Load Demand Solar PV output Fig 27 Sensitive Analysis Representing Solar Irradiance Vs Load Demand Fig 28 Sensitive Analysis Representing Solar PV output Vs Load Demand. The part of the solar irradiance incident on the solar PV module is converted to electricity for which the load demand is met is shown in Fig.27. There is a part of the load demand which has not met by the solar system is shown in Fig 28. This is because of the irradiance not constant throughout the day. Obliviously if solar system alone used, it will result in deficient in serving the load. Hence a grid connected system is essential such systems. Load Demand Grid Purchase Load Demand Grid Sale Fig 29 Sensitive Analysis Representing Grid Power Vs Load Demand. Fig 3 Sensitive Analysis Representing Grid Sale Vs Load Demand The unmet part of the load during all throughout is taken from the grid. Fig.29 represents the amount of supply drawn from the grid throughout the year. The part of the load met form the solar PV system and the excess power generated during the day is supplied to the grid for sale is shown in Fig 3. This will have an economic sense during the production of excess of energy supplying to the gird and during the deficiency period taking from the gird. This will eliminate the cost of the batteries needed to store the excess of energy generated for future use during lack of irradiance in the day. V. CONCLUSION The conventional MPPT control techniques, Perturb and Observe is compared with the advanced ANFIS control technique. Both of them were applied on a chain of energy conversion supplied using a DC/DC Boost converter. The results simulated are subjected to same environmental conditions. It can be realized that MPPT ANFIS controller, which is initially based on the experience of the operator during the training stage, has a very good transient performance. It improves the responses of the photovoltaic system it reduces the time response to the track the maximum power point. This proves the effectiveness of the ANFIS control for photovoltaic systems under varying environmental conditions. Sensitivity analysis explores the power delivered from the solar PV system to the continously varying load demand and the defeceint that has to be met from grid. The excess power that can be transferred to the grid which will eliminate the necessity of storage systems such as batteries to store the excess of energy generated. This analysis predicts the power flow all throughout the day and all throughout the year. REFERENCES [] Purohit, CO 2 emissions mitigation potential of solar home systems under clean development mechanism, India. Energy Resources, vol.34,pp.4 23,29. [2] Gabler H. Autonomous power supply with photovoltaics: photovoltaics for rural electrification reality and vision. Renew Energy,vol.5, pp.2-8, , IJIRAE- All Rights Reserved Page -49

11 [3] DeSoto W, Klein SA, Beckman WA. Improvement and validation of a model for photovoltaic array performance. Sol Energy, vol.8, pp.78-88, 26. [4] Li-qun L, Zhi-xin WA. Rapid MPPT algorithm based on the research of solar cell's diode factor and reverse saturation current. WSEAS Trans Syst, vol.7,issue.5,pp ,28. [5] Mahmodian MS, Rahmani R, Taslimi E, Mekhilef S. Step by step analyzing, odeling and simulation of single and double array PV system in different environmental ariability,proceedings of the international conference on future environment and energy IPCBEE, vol.28, issue.22,pp.32-34,singapoore,22. [6] Esram T, chapman PI, Comparison of photovoltic array maximum power point tracking techniques, IEEE Trans Energy Convers, vol.22, issue.2, 27. [7] Moubayed N, EI-Ali A, Outbib R, A comparison of two MPPT techniques for PV system, WSEAS Trans Environ Dev vol.5,issue.2,29. [8] Ansari MF. Sharma BC, Saini P. Maximum power point tracking of a solar PV module using ANFIS. Proceedings of the 3 rd IEEE international conference on sustainable energy technologies. Nepal: September 22. [9] Reisi AR, Moradi MH, Jamasb S, Classification and comporison of maximum power point techniques for photovoltaic system : a review. Renew Sustain Energy Rev, vol.9,pp ,23. [] Bahgat ABG, Hewa NH, Ahmad GE, EI Shenawy ET, Maximum power point tracking controller for PV system using neural networks.renewenergy, vol.3, pp , 25. [] Moubayed N, EI-Ali A, Outbib R, A comparison of two MPPT techniques for PV system, WSEAS Trans Environ Dev vol.5,issue.2,29. [2] Esram T, chapman PI, Comparison of photovoltic array maximum power point tracking techniques,ieee Trans Energy Convers, vol. 22, issue.2, pp. 27 [3] Abdulaziz M, Aldovbhani S, John R, Maximum power point tracking of PV system using ANFIS prediction and fuzzy logic tracking Proceeding of the international multi conference of engineers and computer scientists, vol.2, Hong Kong 9-2 March 28. [4] Sedaghati F. Nahavandi A, Badamchizadeh MA, Ghaemi S, Fallal MA, PV maximum power-point tracking by using artificial neural networks. Math Prob Eng 22:-(Article ID 5679). 24-5, IJIRAE- All Rights Reserved Page -5