Simulation and Experimental Verification of Intelligence MPPT Algorithms for Standalone Photovoltaic Systems

This study presents compared with Fuzzy Logic Control (FLC) and Adaptive Neuro-Fuzzy Inference System (ANFIS) Maximum Power Point Tracking (MPPT) algorithms, in terms of parameters like tracking speed, power extraction, efficiency and harmonic analysis under various irradiation and cell temperature conditions of Photovoltaic (PV) system. The performance of a PV array are affected by temperature and solar irradiation, In fact, in this system, the experimental implementation and the MATLAB based simulations are In this topology, each Cascaded H-Bridge Inverter (CHBI) unit is connected to PV module through an Interleaved Soft Switching Boost Converter (ISSBC). It also offers another advantage such as lower ripple current and switching loss compared to the conventional boost converter. The results are evaluated by simulation and experimental implemented on a 150 W PV panel prototype with the microcontroller platform. The simulation and hardware results show that ANFIS algorithm is more efficient than the FLC algorithm.


INTRODUCTION
Photovoltaic energy has increased interest in electrical power applications, since it is considered as a fundamentally endless and generally available energy resource.However, the output power induced in the photovoltaic modules depends on solar irradiance and temperature of the solar cells.Therefore, to maximize the efficiency of the renewable energy system, it is compulsory to track the maximum power point of the PV array.This point is called the Maximum Power Point (MPP).The locus of this point has a nonlinear variation with solar irradiance (G) and the cell Temperature (T).Thus, in order to operate the PV array at its MPP, the PV system must include a Maximum Power Point Tracking (MPPT) controller (Gao et al., 2013).
Many MPPT techniques have been reported in the literature such as perturb and observation, incremental conductance, (Hohm and Ropp, 2000;Esram and Chapman, 2007;Safari and Michele, 2011) artificial neural network, fuzzy logic based controller (Ben Salah and Ouali, 2011;Alajmi et al., 2011;Kottas et al., 2006;Salhi and El-Bachtri, 2011) ANFIS (Putri and Rifa, 2012;Mellit and Kalogeria, 2011) etc.In this study FLC and ANFIS MPPT algorithm being used to extract the maximum DC power from PV module.The generated DC power is converted into AC, in order to be used in a standalone system.In recent times, multilevel inverter topologies have received more attention to the use in PV applications (Beser et al., 2010).The output waveforms are much improved over conventional inverter.This leads to harmonics in the output voltage and current of the multilevel inverter (Leon et al., 2013).To overcome the difficulties, in this study, single phase selective harmonic elimination ANN integrated modulation technique is proposed and verified.

Stand alone PV system:
The block diagram of the proposed topology for ISSBC and CHBMLI based stand-alone PV system is shown in Fig. 1, the inverter is fed from photovoltaic module through a DC-DC converter integrated with FLC and ANFIS MPPT algorithm.The output of the single phase SHE trained ANN unit has been applied to the CHBI to achieve a balanced output with improved power quality even under non-ideal condition of PV cell.To analyze and compare FLC and ANFIS MPPT algorithms for different conditions such as changing solar irradiation and cell temperature.The simulation results are validated experimentally in a hardware setup with the 16F877A microcontroller platform.

PV array modeling and simulation:
The PV array used in the proposed system is 72 multi-crystalline where, I PV,Cell is current generated by the incident light (directly proportional to sun irradiation), I 0,Cell is leakage current of the diode, q is electron charge 1.6021×10-19 C, k is Boltzmann constant (1.38×10-23 J/K), T is Temperature of the PN junction and a is diode ideality constant.To develop embedded SIMULINK model based on current equation and manufacturer's data sheet parameter of BP SX 150S model as shown in Table 1.

MPPT control algorithms:
The MPPT algorithm is used for extracting the maximum power from the PV module and passes it on to the load.A converter serves the purpose of transferring maximum power from the solar PV module to the load.By changing the duty cycle the load impedance, as seen by the source, is varied and matched at the point of the peak power with the source so as to transfer the maximum power.

FLC MPPT controller:
The FLC MPPT algorithm is used to ISSBC to compensate the output voltage of PV system to keep the voltage at the value which maximizes the output power.The fuzzy logic controller consists of three basic elements, namely fuzzification, rule base inference engine and defuzzification.
Fuzzification comprises the process of transfosrming numerical crisp inputs into linguistic variables based on the degree of membership to certain sets.The fuzzification variables are logical decision framed in inference engine block and deliver linguistic where, P (k) and V (k) are the instant power and voltage of the photovoltaic system, respectively E (k) is zero at the maximum power point of PV array.The input E (k) shows if the operation point at the instant k is located on the left or on the right of the MPPT on the PV Characteristic while the input CE (k) expresses the moving direction of this point Eq. ( 3).

ANFIS MPPT algorithm:
The ANFIS system is used to formulate the neural network architecture in the inference engine of a Fuzzy controller.The functional block diagram and structure of ANFIS is shown in Fig. 4. The structure comprises of three distinct layers namely input layer, hidden layer and output layer.The ANFIS controller implemented in this study is of the model described as above whose fuzzifier section comprises of the input signals error (e) and change in error signal (ce) whose membership functions are selected as Gaussian membership function.The defuzzifier of the ANFIS is the output function that is the modulation index (d).
The input membership functions are mapped to the output membership function by 49 rules through grid partitioning method using the FIS generator in MATLAB SIMULINK.The 2500 data sets to train Soft switching boost converter: It serves as a suitable interface for PV cells to convert low voltage, high current input into a high voltage low current output.The interleaved boost inverter consists of two single phase boost inverters that are connected in parallel and inverters operating 180° out of phase with 30 kHz switching frequency (Jung et al., 2011).It is pointed out that in interleaved inverter mode 60 kHz effect is achieved by phase shifting of the two 30 kHz switching signals.Because the inductor ripple currents are out of phase, they cancel each other and the input-ripple current reduce to 12% of that of a conventional boost inverter.The best input-inductor-ripple-current cancellation occurs at 50% duty cycle.

Single phase CHB inverter:
The cascaded multilevel inverter is composed of a number of H-bridge inverter units with separate DC source for each unit and can be connected in cascade to produce a near sinusoidal output voltage waveform.There are different switching strategies implemented for minimization and elimination of harmonics (Tsang and Chan, 2013;Ravi et al., 2011).This fact leads to distortion in the output voltage waveform of the multilevel power inverter: In this study selective harmonic elimination pulse with modulation technique is implemented to generate the switching duty cycle for CHB inverter.The Eq. ( 4) shows the contents of the output voltage at infinite frequencies, the module voltage V pv1 -V pv2 are associated to their respective switching angle α 1 -α 2 .These trigonometric transcendental equations can be solved by GA and implemented to find the switching angle (offline) for a set of predetermined modulation indices to get the required fundamental output voltage in a nine level cascaded multi level inverter.The switching angles (α 1 , α 2 ) lie in between 0 and π/2.

RESULTS AND DISCUSSION
Simulation results: The SIMULINK software validates the performance of the MPPT techniques under different operating conditions.The parameters are considered in the Standard Test Condition (STC): 1000 W/m 2 and cell temperature of 25°C.The simulation circuit diagram is shown in Fig. 6.The Simulations are carried out for the two techniques under dynamically changing solar irradiations at temperature of 25°C.Figure 9 shows output power of sudden changes in solar irradiation from 600 to 1000 W/m 2 .In this analysis, the two techniques are able to extract the MPP, the detailed simulation result, voltage, power and corresponding efficiency tabulated in Table 2. Observed result, the ANFIS more power extracted and fast response to reach the new MPP, after solar irradiation changes compared to FLC also gives a fast steady state response with less oscillation.

Dynamic variation of the solar irradiation:
Simulations are carried out for the two techniques under dynamically changing solar irradiations at temperature of 25°C.Figure 9 shows output power of sudden changes in solar irradiation from 600 to 1000 W/m 2 .In this analysis, the two techniques are able to extract the MPP, the detailed simulation result, voltage, power and corresponding efficiency tabulated in Table 2.  Observed result, the ANFIS more power extracted and fast response to reach the new MPP, after solar irradiation changes compared to FLC also gives a fast steady state response with less oscillation.

Dynamic variation of cell temperature:
This simulation is carried out to illustrate the performance of the MPPT methods under constant solar irradiation of 900 W/m 2 and changes in temperature from 25 to 50°C.The temperature has a slight effect on the cutoff circuit current.However, the open circuit voltage decays quickly as the temperature increases.Figure 10 shows the corresponding PV voltage.Power and efficiency, during slowly occurring as well as sudden changes in temperature, respectively, the FLC and ANFIS the maximum extracted power and response time tabulated in Table 3.When compared to ANFIS the FLC does not converge to the globally maximum power point.The ANFIS exhibits fast response and also converges to the globally maximum power point with slight fluctuations and highest PV output power in the change in dynamics of temperature.
Simulation with converter and inverter: Finally, in order to verify the performances of the FLC and ANFIS algorithm, the CHBI is connected to an RL load (R = 100 ohm and L = 20 mH) using switching frequency of 30 kHz in the ISSBC.The PV array receives constant solar irradiation of 1000 W/m 2 cell temperature at 25°C.The converter output voltage injected to the Hbridge inverter, the output of the stepped modulated inverter voltage, along with their harmonic spectrum (7.5 KHz) of the FLC and ANFIS algorithms are shown in Fig. 11 and 12, respectively.The Total Harmonic Distortion (THD) of the output voltage with the FLC model of control is 29.54% and with the ANFIS model they are 25.62%, respectively.It can be observed from the simulation results the percentage THD is less in ANFIS algorithm as compared to the FLC algorithm.

Experimental validation:
The simulation results were verified experimentally in the using the appropriate hardware built around the PIC 16F877A microcontroller as shown in Fig. 13.The solar panels are not shown in photograph.The controller program is downloaded into microcontroller and generates gating signals to the ISSBC and CHBI.For the validation of maximum power point tracking control, the developed system is tested at 9.00 h, 13:15 and 16.15 h.The irradiation and temperature were measured as 350 W/m 2 35.6°C 1050 W/m 2 35°C and 625 W/m 2 36°C,  respectively, experimental pattern also shown in Table 4.The experimentation with FLC and ANFIS algorithm collected detailed result tabulated in Table 5.
During test period the highest irradiation 1050 W/m 2 at cell temperature 35°C (case 2) analyzed both algorithm by (model: UNI-T) four channels DSO and take the voltage waveform with their harmonic spectrum.The (peak to peak) rms value of output voltage is found as 70.76 and 72.14 Volts, respectively.Figure 14 shows the voltage and harmonic spectrum and the corresponding THDs which are found to be 12.3 and 10.3%, respectively.Hence, in different operation modes, ANFIS algorithm improves the voltage quality, power extraction, harmonics elimination as compared to the FLC algorithm.

CONCLUSION
This study analyzes the simulation and experimental performance of FLC and ANFIS MPPT algorithms by stand-alone PV system.The configuration for the proposed system is designed and simulated using MATLAB/SIMULINK and implemented in 16F877A micro controlled prototype.The ANFIS MPPT algorithm improves the voltage quality, power extraction, harmonics elimination as compared to the FLC MPPT controller.The results obtained from ANFIS MPPT algorithm can gain importance in high performance applications such as PV standalone generation system.

Fig. 1 :
Fig. 1: General diagram of load connected photovoltaic system

Fig. 3 :
Fig. 3: FLC MPPT control system and FLC surface output output.Defuzzifier is used to convert linguistic fuzzy sets to actual value.In this study the fuzzy inference rule is carried out by using Mamdani's method and the defuzzification use the centre of gravity to compute the output of this FLC which is the duty cycle.The SIMULINK controller and surface output shown in Fig. 3.The two FLC input variables are the error E (k) and change of error CE (k) at sampled times k defined in Eq. (2): 1

Table 1 :
PV module parameter

Table 3 :
Dynamic behavior of temperature