Solar Energy 220 (2021) 473–490

Contents lists available at ScienceDirect

Solar Energy
journal homepage: www.elsevier.com/locate/solener

Design and development of realistic PV emulator adaptable to the

maximum power point tracking algorithm and battery charging controller
A. Nazar Ali a, K. Premkumar a, *, M. Vishnupriya b, B.V. Manikandan c, T. Thamizhselvan a
a
Rajalakshmi Engineering College, Tamil Nadu, India
b
Saveetha School of Engineering, Tamil Nadu, India
c
Mepco Schlenk Engineering College, Tamil Nadu, India

A R T I C L E I N F O A B S T R A C T

Keywords: In this paper, PV emulator is developed using double loop PI controlled DC-DC buck converter. Ambient envi­
Battery charging control ronmental conditions such as irradiance, temperature and wind speed are considered for creating PV reference
Genetic algorithm model in the PV emulator. The simulation model of the proposed PV emulator is created and tested using
MPPT
MATLAB Simulink package. To validate the effectiveness of the proposed PV emulator, it is tested with different
PI control
PV array
operating conditions such as varying irradiance, temperature and wind speed. The current–voltage and power-
PV emulator voltage characteristic of the PV emulator is compared with PV reference model. The suitability of the PV
emulator is tested with MPPT algorithm and battery charging controller in addition to number of objectives. Real
time testing of the proposed PV emulator is done experimentally and also corresponding results are discussed.
The proposed Solar PV emulator has the following properties: (1) higher bandwidth DC-DC buck converter, (2)
extremely reliable performance with lower response time, and (3) lower output voltage and current ripple.

1. Introduction PV simulator are as follows: In Cirrincione et al. (2008), real time

implementation of PV simulator using Growing Neural Gas Network
Renewable generation, integration and their associated control have (GNGN) based DC-DC buck converter was presented. Input parameter
emerged as promising and fertile area in the research field. Solar PV for GNGN is irradiance, temperature and output current of the buck
energy generation, wind energy generation, tidal energy generation and converter and output is reference voltage and it is compared with actual
bio mass energy generation has become popular in the renewable energy output voltage and it is processed through controller to emulate the PV
power generation (Akbari-Dibavar et al., 2020). As per statistics in characteristics. Normally IV characteristics of the PV cell are divided
(http://www.fi-powerweb.com/Renewable-Energy.html), Solar PV into three regions such as constant voltage source region (A-B), variable
power generation is the third most prominent energy supply in the voltage and current source region (B-C) and constant current source
world. More research is ongoing in maximum power point tracking of region (C-D and it is shown in Fig. 1. The developed model in Cirrincione
solar PV system, solar powered battery charging, solar powered electric et al. (2008), only operates in two regions i.e., constant voltage region
vehicle applications but researchers are facing quite a number of prob­ and variable voltage and current region. The simulator was tested using
lems in testing maximum power tracking algorithm, battery charging MATLAB simulink package and developed control system was imple­
controller and integration of all these in electric vehicle control in real mented using DSP processor. In Veerachary (2006), PSIM simulation
time. Since solar PV power generation normally depends on environ­ software has been used for realizing the PV simulator and it was
mental conditions such as irradiance level, temperature and wind speed, developed based with circuit-oriented simulation considering only
researchers are facing an uphill task with regard to testing their devel­ changing irradiance conditions. The designed system has been tested
oped algorithm under all possible varying environmental conditions with MPPT algorithm for various irradiance level conditions.
(Prasanth Ram et al., 2018). In order to overcome this problem, PV In Piazza and Vitale (2010), design and development of PV simulator
simulator has been developed for testing the developed algorithms. The under partial shaded conditions were presented. The controller imple­
PV simulator was developed and tested with two modes i.e., simulation mentation was created using MATLAB software and implemented in the
model testing and real time model testing. The literature reviews on such dSPACE controller. The system was working efficiently under variable

* Corresponding author.
E-mail address: [email protected] (K. Premkumar).

https://doi.org/10.1016/j.solener.2021.03.077
Received 5 October 2020; Received in revised form 10 March 2021; Accepted 27 March 2021
Available online 10 April 2021
0038-092X/© 2021 International Solar Energy Society. Published by Elsevier Ltd. All rights reserved.
A. Nazar Ali et al. Solar Energy 220 (2021) 473–490

Chavarría et al. (2014). The current control scheme with buck converter
was developed to mimic the behavior for the real time PV cell. Look up
table concept was used in this work for generating reference signal for
the controller and converter and this method has not considered the real
time variation of irradiance and temperature.
The PV panel parallel with programmed DC source based PV
emulator was presented in Zhou et al. (2014). In this method, actual PV
panel has been connected with DC current source to emulate the PV
panel characteristics. It is only useful when rating of PV panel is small
and cost of the PV panel becomes exorbitant in case researcher wish to
emulate and test the algorithm for large rating PV panel. Field pro­
grammable analog array was used to implement the PV emulator control
algorithm in Balato et al. (2016). The PV reference model output de­
pends on short circuit current, open circuit voltage and environmental
temperature and this method has not considered variations in irradiance
Fig. 1. IV characteristics of typical PV cell. level. The MATLAB simulink model based PV emulator was developed
and tested for grid integration in Iqbal and Tariq (2015). The voltage
voltage and current source region of the IV characteristics. In Can mode controlled buck converter was used to simulate the PV panel
(2013), MATLAB simulink model of the PV simulator was developed. characteristics. The look up table concept was used for generating
The developed system used the PI based voltage controlled buck con­ reference characteristics of the actual PV panel and this model was
verter to emulate the characteristics of the PV panel. While investigating working in variable voltage and current source region only. The digital
the results, it was found that the PV emulator emulated the variable signal processor was used to implement the PV emulator control algo­
voltage and current region of the PV panel only. The simulation study on rithm in Shah and Rana (2015). The reference PV model output depends
PV emulator for testing the grid connected inverter was proposed in

Table 1
Summary of literature review on PV emulator.
Reference No. as provided in Simulation Hardware platform used Controller used PV reference Region of Application Cost/
the section & Year Software used model used operation testing Complexity

MATLAB Digital signal processor PI voltage single diode VVCS – Medium

Buck Converter controller model
(Veerachary, 2006) PSIM – – single diode CVS& VVCS MPPT Medium
model
(Piazza and Vitale, 2010) MATLAB dSPACE Buck Converter PI voltage double diode VVCS Partial shaded Medium
controller model testing
(Can, 2013) MATLAB PI voltage single diode VVCS – Medium
controller model
(Chavarría et al., 2014) MATLAB – PI current Look uptable VVCS Inverter Medium
controller
(Zhou et al., 2014) – Solar PV Panel – – CVS& VVCS MPPT High
Programmable DC source
(Balato et al., 2016) – FPAA Buck Converter PI current Single diode VVCS – High
controller model
(Iqbal and Tariq, 2015) MATLAB – PI voltage Lookup table VVCS Grid High
controller Integration
(Shah and Rana, 2015) MATLAB – Bang- Bang control Single diode VVCS – Medium
model
(Ahmad Saudi Samosir, 2015) MATLAB – PI current single diode VVCS – Medium
controller model
(Ebrahim et al., 2015) MATLAB dSPACE Programmable DC – single diode VVCS – Medium
supply model
(Mohammed Azharuddin MATLAB dSPACE Buck Converter PI voltage single diode CVS & VVCS – Medium
et al., 2014; Azharuddin controller model
et al., 2016)
(Ickilli et al., 2012) Xilinx FPGA Buck Converter PI voltage single diode CCS & VVCS – High
controller model
(Lu and Nguyen, 2012) MATLAB PICAXE AXE-08 M – single diode CVS & VVCS MPPT High
Microcontroller, Boost model
converter
(Atoche et al., 2015) Python ARM Processor Forward – Look up table VVCS – High
Converter
(Moussa et al., 2019) MATLAB Linear regulator PI Current logarithmic CVS & VVCS – High
regulator amplifier
(Chalh et al., 2018a; Chalh – DC source Variable resistor – No reference VVCS MPPT Simple
et al., 2018b) model
(Jayawardana et al., 2019) RSCAD RTDS Simulator, Buck single diode CVS & VVCS Grid High
simulation Converter model Integration
software
(Carralero, 2019) MATLAB Four-Switch Buck-Boost DC- PI Voltage single diode CVS & VVCS – Medium
DC Converter Regulator model
(Ullah et al., 2020) MATLAB Hardware in the loop Buck fractional order single diode CVS & VVCS – Medium
Converter sliding mode model
control

474
A. Nazar Ali et al. Solar Energy 220 (2021) 473–490

Fig. 2. Block diagram of proposed PV Emulator.

on four parameters such as irradiance level, temperature, load current series connected switched DC source was used for varying the temper­
and load voltage. Bang-Bang control algorithm was developed to control ature level equivalent to real time environmental scenario. It has oper­
the buck converter in order to mimic the PV model characteristics. This ated only in the variable voltage and current source region and also large
model was only working in two regions i.e., constant voltage source deviation in the characteristics of the PV emulator was observed. Sum­
region and variable voltage and current source region. mary of the literature survey is shown in Table 1.
In Samosir (2015), simulation model of PV emulator using PID Operating regions are shown in Fig. 1 and most of the PV emulator is
controlled buck converter was implemented. Single diode model of the found to operate only on constant voltage source and variable voltage
PV panel was taken as reference model to supply the reference input for and current source regions and it seldom operated in constant current
the controller. The system has been tested for different irradiance level source region. Regarding the controller used to control the converter,
and temperature conditions. The system operates only in the variable either the PI voltage controller or PI current controller has been incor­
voltage and current source region. The dSPACE controller with pro­ porated and no attempt has been made in utilizing the combination of
grammable DC source was used to develop the PV emulator in Ebrahim both controllers. The double loop controller, one for the inner loop and
et al. (2015). Single diode model was used for PV panel reference model the other for the outer loop enables the PV emulator to operate in all the
in the dSPACE controller. The system was tested for different real time regions. Mostly single diode model was used as reference PV model and
environmental conditions. It has only emulated the variable voltage and rarely look up table concept was used. These models has the following
current source region. In Mohammed Azharuddin et al. (2014) and input parameters such as irradiance, temperature, open circuit voltage,
Azharuddin et al. (2016), dSPACE controller was used to implement the short circuit current, load current of the converter and load voltage of
PV emulator control algorithm. PI voltage controlled DC-DC buck con­ the converter and these models has not taken into consideration the
verter was used for emulating the behavior of the PV panel. Single diode ambient conditions like real time wind speed. The PV emulator control
model was used to generate reference signal for the controller. The algorithm was implemented using dSPACE, FPGA, DSP processor, Arm
system was operating only in two regions namely constant voltage processor and FPAA. These hardware controllers are costlier and also
source and variable voltage and current source region. PV emulator presents greater complexity in the practical implementation.
control algorithm was implemented using Field Programmable Gate From the above presented literature survey, it is clear that there is
Array (FPGA) in Ickilli et al. (2012). Single diode model was used to ample scope in the development of PV emulator considering number of
generate the reference single for the controller. PI voltage controller was realizable objectives. The list of objectives includes emulating the PV
used to control the DC-DC converter to emulate the characteristics of the emulator to operate in three operating region and to consider all
PV panel. The developed model was tested for different irradiance and ambient conditions for the reference PV model. Further, the objective
temperature conditions. The system exhibited the behavior of constant comprises achieving effective emulation of PV panel in all regions using
voltage source region and variable voltage and current source region. double loop controller implementation and to implement the PV
In Lu and Nguyen (2012), PV emulator control concept was imple­ emulator control algorithm using low cost microcontroller. Also the
mented in PICAXE AXE-08M microcontroller. The single diode model objectives include testing the PV emulator adaptability for MPPT and
was used to generate reference signal for the controller so as to control battery charging application under fast varying irradiance, temperature
the characteristics of the DC-DC converter in order to mimic the and wind speed conditions. The organization of paper is as follows:
behavior of the real time PV panel. Larger deviation in the PV charac­ Design and development of proposed PV emulator is explained in the
teristics of the developed PV emulator was observed while examining Section 2. MATLAB simulink model and corresponding results are dis­
the experimental results. The Arm processor was used to implement the cussed in Section 3 and Section 4 presents the hardware implementation
control algorithm of the PV emulator in Atoche et al. (2015). Look up details. Concluding remarks are outlined in Section 5.
table concept was used to generate the reference signal for the forward
converter to emulate the PV panel characteristics. The system was tested 2. Proposed PV emulator
for different irradiance and temperature conditions and system has
operated only in the variable voltage and current source region. In The block diagram of the proposed PV emulator is shown in Fig. 2.
Moussa et al. (2019), logarithmic amplifier concept was used to develop The proposed emulator consist of DC source, DC-DC buck converter,
the reference PV model for emulating the PV panel characteristics. The electrical load, PV reference model, PI current controller, PI voltage
system has operated only in two regions such as constant voltage source controller, PWM pulse generator, voltage and current measurement
region and variable voltage and current source region. In Chalh et al. system. The PV reference model have four inputs namely irradiance,
(2018a) and Chalh et al. (2018b), DC source with series resistance was temperature, wind speed and converter load voltage and has only one
used to emulate the characteristics of the PV panel. Variable series output i.e., PV reference current. The PV reference current is compared
resistance switched stack was used for varying the irradiance level and with the load current of the converter and processed through PI current

475
A. Nazar Ali et al. Solar Energy 220 (2021) 473–490

Fig. 3. Simulink model of the proposed PV emulator.

where,
( )
( ) − VMP
IMP A2 ×VOC
A1 = 1 − ×e
ISC
( )
VMP (2)
− 1
V
A2 = ( OC )
IMP
ln 1 −
ISC
The parameter A1 and A2 depends on following parameters of the
solar panel such as short circuit current (ISC), open circuit voltage (VOC),
current at maximum power point (IMP) and voltage at maximum power
Fig. 4. Single diode model of the solar PV cell. point (VMP). The parameters ISC, VOC, IMP and VMP depends on the irra­
diance, temperature, wind speed, parameter of the PV panel datasheet.
These are expressed in the following Eqs. (3), (4) & (5) and corre­
controller and it generates the PV reference voltage. The PV reference
sponding simulink model is shown in Fig. 7,
voltage is compared with converter load voltage and processed through
PI voltage controller and it generates the duty cycle. The duty cycle G
ISC (G, TC ) = ISCS × × (1 + δ × (TC − TS ) )
processed through PWM pulse generator unit and it will generate the GS
pulse to control DC-DC buck converter for emulating the PV reference (3)
G
model characteristics. The simulnk model of the proposed PV emulator IMP (G, TC ) = IMPS ×
GS
× (1 + δ × (TC − TS ) )
is shown in Fig. 3. The design details of PV reference model, PI voltage
and current controller and DC-DC Buck converter are explained in the VOC (TC ) = VOCS + ε × (TC − TS )
(4)
proceeding section. VMP (TC ) = VMPS + ε × (TC − TS )

2.1. Modeling of PV reference model TC = 1.14 × (T − TS ) + 0.0175 × (G − 300) − Ar × W + 30 (5)

The equivalent circuit for PV reference model is shown in Fig. 4. The where, ISCS, IMPS, VOCS, VMPS are the short circuit current, current at
PV reference model output current (IPV) is derived as a function of the PV maximum power point, open circuit voltage and voltage at maximum
reference model output voltage (VPV). The current, voltage and power of power point at standard test conditions i.e., standard irradiance (GS) is
the solar PV reference model depends on the following parameters such 1000 W/m2 and temperature (TS) is 25 ◦ C. Current temperature coeffi­
as irradiance (G), cell temperature (TC) and wind speed (W). The cient and voltage temperature coefficient are denoted by δ and ε. Con­
equation relating PV current and PV voltage is expressed by the stant Ar depends on the solar cell types i.e., 1.509 for monocrystalline
following Eqs. (1) and (2) and corresponding simulink model are shown solar cell, 1.468 for polycrystalline solar cell, and 1.450 for thin-film
in Fig. 5 and Fig. 6, technology solar cell (Ayaz et al., 2020).
⎡ ⎛ ( ) ⎞⎤
VPV
⎢ ⎜
IPV = ISC × ⎣1 − A1 × ⎝e
A2 ×VOC
⎟⎥
− 1⎠⎦ (1) 2.2. Designing of PI current and voltage controller

In this section, PI current and voltage controller is developed for DC-

476
A. Nazar Ali et al. Solar Energy 220 (2021) 473–490

Fig. 5. Simulink model of the Eq. (1).

Fig. 6. Simulink model of the Eq. (2).

477
A. Nazar Ali et al. Solar Energy 220 (2021) 473–490

Fig. 7. Simulink model of the Eqs. (3), (4) and (5).

Fig. 8. Circuit diagram of DC-DC buck converter.

Fig. 9. Closed loop transfer function model of DC-DC buck converter.

DC buck converter to emulate the behavior of the PV reference model. the duty cycle (d) of the switching elements. The transfer function model
To design the PI controller, transfer function model of the DC- DC buck of the DC-DC buck converter is expressed in the following Eqs. (6) & (7)
converter is derived first. The circuit diagram of the DC-DC buck con­ as,
verter is shown in Fig. 8. The specifications of the DC-DC converter are:
Vo VS
inductance (L) is 22mH, capacitance is 4700 µF, switching frequency is = (6)
D 1 + S2 ∙LC + S∙RL
2 kHz and power rating of the converter is 350 W.
The DC-DC buck converter output voltage and current depends on

478
A. Nazar Ali et al. Solar Energy 220 (2021) 473–490

Fig. 10. Flowchart of Genetic algorithm optimization for PI controller.

Io VS Normally, trial and error method or Ziegler-Nichols method are used

= R
(7) to find the parameter of the PI controller but it is less accurate and time
D 1 + S ∙LC + S∙RL
2

consuming process. Conventional PI control tuning approaches sug­

The closed loop transfer function model of DC-DC buck converter is gested by most researchers are model-dependent and involve models of
shown in Fig. 9. This model consist of two loops i.e., inner loop and outer a reduced order. The tuning rule proposed for a specific reduced order
loop. Inner loop is used to control the output voltage of the DC-DC buck model will not give a suitable solution, especially for unstable systems.
converter and outer loop is used to control the output current of the DC- In order to get the best possible controller parameters, most classical PI
DC buck converter. The tracking of the reference current and voltage tuning approaches require numerical computations. Due to these rea­
from the PV reference model depends on gain parameters (KI1, KI2, KV1 sons, Genetic Algorithm (GA) approaches have been used in recent de­
and KV2) of the PI current and voltage controller. These parameters are cades to optimize the PI gains in order to deal with nonlinear and
optimized using classical Genetic Algorithm (GA). complicated processes and it has following merits such as easy imple­
mentation, fast convergence and excellent global exploration ability.
Genetic algorithm is meta-heuristics optimization algorithm. It has
five steps i.e., initial population, selection, crossover, mutation and
reproduction. Initial population is generated randomly using upper and
lower bounds of the parameters of the given problem. After initialization
of the population, fitness value for each population set is determined
using objective function of the given problem and sort the fitness value
and corresponding population set in ascending (minimization problem)
or descending order (maximization problem). Next step is selection
process. In this process, fittest population is taken for further iteration
and remaining population sets are taken from the crossover and muta­
tion process. In crossover, two set or more than two sets of population
are taken for matting and it produces offspring and it is included in the
population sets. In mutation process, some of the offspring mutate with
Fig. 11. Convergence plot.
low random values. In the final reproduction stage, population from the

479
A. Nazar Ali et al. Solar Energy 220 (2021) 473–490

Table 2
Optimized parameter of the PI controller using GA.
KV1 KV2 KI1 KI2 Best fitness Worst Fitness Mean Standard deviation Computation time (sec)

11.2 12.5 296 145 0.21 0.54 0.42 0.142 415

(a) Testing of MPPT using PV Emulator

(b) Testing of solar MPPT battery charger controller using PV Emulator

Fig. 12. Testing of MPPT and solar MPPT battery charger controller using PV Emulator.

selection process, crossover and mutation process are ordered and send 2.3. Testing of MPPT and solar MPPT battery charger controller using PV
to next iterations for finding optimal values as per the given problem. emulator
This process is continued until stop criterion reached or maximum it­
erations reached (Protopopova and Kulik, 2020). The flow chart for The developed PV emulator is tested with maximum power tracking
optimization of the gain parameter of the PI voltage and current algorithm and solar battery charger controller to validate the effec­
controller is shown in Fig. 10. tiveness of the proposed PV emulator. Fig. 12(a) shows the block dia­
Objective function is important to find out the optimum gain pa­ gram of testing Perturb & Observe (P&O) MPPT algorithm with PV
rameters of PI voltage and current controller using GA. The integral of emulator. DC-DC boost converter and P&O method are used to extract
absolute of voltage error function is taken as objective function for this the maximum power from PV emulator. PV voltage and current is pro­
optimization and it is expressed in the following Eq. (8), posed through P&O method and at every instant, P&O method perturbs
∫ the PV voltage and current by controlling duty cycle of the DC-DC boost
f (KV1 , KV2 , KI1 , KI2 ) = |VPVREF − VCONVERTER | (8) converter to extract the maximum power from the PV emulator.
Simuilink model of testing MPPT algorithm using PV emulator is shown
The parameters used for the genetic algorithm are: number of pop­ in Fig. 13(a). The specifications of the boost converter are: inductance
ulations is 20, maximum number of iteration is 100, selection type is (L) is 13mH, input and output capacitance values are 1000µH, switching
roulette wheel selection, crossover type is Arithmetic crossover, cross­ frequency is 1 kHz and power rating of the converter is 400 W.
over rate is 0.7, mutation rate is 0.1 and number of trial is 100. Fig. 11 Fig. 12(b) shows the block diagram for testing the solar MPPT battery
shows the convergence plot for the PI controller optimization. The ge­ charger controller using PV emulator. DC-DC buck converter and P&O
netic algorithm converges to the global minima after 45 iterations and method is used to extract the maximum power from PV emulator. The
fitness value of the optimization is 0.21. The genetic algorithm has been duty cycle is passed to next stage based on the output from the battery
executed for 100 trials and the corresponding results are presented in charger controller. The battery charger controller checks the two con­
Table 2. ditions; State Of Charge (SOC) of the battery is compared with 100% and
battery voltage is compared with fully charged voltage settings and cut

480
A. Nazar Ali et al. Solar Energy 220 (2021) 473–490

Fig. 13. Simulink model of testing of MMPPT and solar MPPT battery charger controller using PV Emulator.

controller are provided in the proceeding section.

Table 3
Parameter of the PV reference model.
3. Simulation results and discussion
S.No Parameter Values

1 Maximum Power (Watts) 248.9 In this section, simulation results of the proposed PV emulator for
2 Open circuit Voltage (Volts) 38.4 different operating conditions and testing of MPPT and solar MPPT
3 Short Circuit Current (A) 8.85
battery charger controller are outlined. Parameters used for the PV
4 Voltage at maximum power point (Volts) 30.7
5 Current at maximum power point (A) 8.11 reference model is shown in Table 3. The developed PV emulator is
6 Current temperature coefficient (% A/◦ C) 0.07 tested with variable load for the three cases. In case 1, temperature and
7 Voltage temperature coefficient (% (A/◦ C)) − 0.35599 wind speed are fixed at 25 ◦ C and 6 m/s and irradiance is varied and kept
at 400 W/m2, 600 W/m2, 800 W/m2 and 1000 W/m2. The corre­
sponding PV emulator voltage, current and power response are shown in
of voltage settings of the battery. The duty cycle is allowed to control the
Fig. 14(a).
DC-DC buck converter if both conditions are true otherwise duty cycle is
In case 2, irradiance and wind speed are fixed at 1000 W/m2 and 6
not allowed to control the DC-DC buck converter. Simulink model of the
m/s and temperature is varied and kept at 15 ◦ C, 25 ◦ C, 35 ◦ C and 45 ◦ C.
testing of solar MPPT battery charger controller using PV Emulator is
The corresponding PV emulator voltage, current and power response are
shown in Fig. 13(b). The specifications of the buck converter are:
shown in Fig. 14b). In case 3, irradiance and temperature are fixed at
Inductance (L) is 10mH, Input and output capacitances are 1000µH,
1000 W/m2 and 25 ◦ C and wind speed is varied and kept at 5 m/s, 7 m/s,
switching frequency is 1 kHz and power rating of the converter is 400 W.
9 m/s and 11 m/s. The corresponding PV emulator voltage, current and
The rating of battery is 12 V, and 40 Ah. The simulation results of PV
power response are shown in Fig. 14(c). From the results shown in
emulator operation for different irradiance, temperature and wind speed
Fig. 14(a), it is inferred that, PV emulator voltage decreases from open
conditions and testing of MPPT and solar MPPT battery charger
circuit voltage level to short circuit voltage level and it is due to increase

481
A. Nazar Ali et al. Solar Energy 220 (2021) 473–490

Fig. 14. Voltage, Current and Power response of the proposed PV Emulator with different irradiance, temperature and wind speed conditions.

of load from no load to short circuit current load. Power of the PV emulator. The PV emulator voltage, converter voltage, current and
emulator increases from no load and reaches maximum power condi­ power for the three cases are shown in Fig. 16(a–c). In case 1, irradiance
tion, from there it decreases sharply to short circuit power conditions. is varied and kept at 400 W/m2, 600 W/m2, 800 W/m2 and 1000 W/m2
PV emulator behaves as constant voltage source from 0 to 10 secs, as for every 0.5 sec (Temperature and Wind Speed fixed at 25 ◦ C and 6 m/s)
variable voltage and current source from 10 to 15 secs and as constant and corresponding results are measured. The PV emulator voltage and
current source from 15 secs to 35 secs. Replica of the characteristics current are maintained at par with the voltage and current corre­
exhibit for the case 2 and case 3 conditions also. sponding to the maximum power point condition for the varying irra­
The current–voltage (IV) and power-voltage (PV) characteristics of diance level so as to extract maximum power from PV emulator. From
the PV emulator and PV reference model are compared for the three the power response of PV emulator and converter, it is observed that
cases and depicted in Fig. 15(a–c). The solid line response indicates the MPPT algorithm effectively extracts the maximum power from the PV
IV and PV characteristics of PV reference model and circle line response emulator for different irradiance conditions. In case 2, temperature is
indicates the IV and PV characteristics of the proposed PV emulator. The varied and kept at 15 ◦ C, 25 ◦ C, 35 ◦ C and 45 ◦ C for every 0.5 sec
IV and PV characteristics of the PV reference model and proposed PV maintaining irradiance and wind Speed fixed at 1000 W/m2 and 6 m/s.
emulator are superimposing each other for all the considered cases. In case 3, wind speed is varied and kept at 5 m/s, 7 m/s, 9 m/s and 11 m/
From this result analysis, it is evident that the proposed PV emulator s for every 0.5 sec maintaining irradiance and temperature fixed at 1000
effectively emulates the characteristics PV reference model for all W/m2 and 25 ◦ C. For both case 2 and 3, response and characteristics
operating cases considered. The emulation efficiency of the PV emulator similar to case 1 are obtained. The actual power and theoretical
as calculated from the ratio of the power of the PV emulator to the power maximum power of the PV emulator and power of converter are pre­
of the PV reference model is 98.25%. sented in the Table 4 and Table 5. The average maximum power point
The developed PV emulator is tested with classical perturb and extraction ratio of the MPPT algorithm from the ratio of actual
observe method in order to observe the suitability of the proposed PV maximum power to the theoretical maximum power of PV emulator is

482
A. Nazar Ali et al. Solar Energy 220 (2021) 473–490

Fig. 15. Comparison of IV and PV characteristics of Proposed PV emulator with PV reference model.

97.45%. The average converter efficiency is 98.3% as determined form power of PV emulator and battery charging and charging efficiency is
the ratio of the output power of converter to the power of PV emulator. shown in Table 6. It is observed that, average charging efficiency of the
The developed PV emulator is tested with solar MPPT battery solar MPPT battery charger controller determined from the ratio of
charger controller to check the appropriateness of the proposed PV power of the converter to the power of the PV emulator is 93.9%.
emulator. The PV emulator and converter voltage, current and power for
three cases are shown in Fig. 17. For the case 1 shown in Fig. 17(a), 4. Experimental verification of the performance of proposed PV
irradiance is varied and kept at 400 W/m2, 600 W/m2, 800 W/m2 and emulator
1000 W/m2 for every 0.5 sec (Temperature and Wind Speed fixed at
25 ◦ C and 6 m/s) and corresponding results are measured. The battery In this section, experimental verification of the performance of
voltage is maintained from 11 to 13 V, i.e., between cut of voltage and proposed PV emulator is developed and tested in real time to check the
fully charged voltage settings. State of charge of the battery is increased suitability and effectiveness of the proposed system. The implementa­
due to charging current from the converter. Similar characteristics tion diagram of the proposed PV emulator is shown in Fig. 18(a).
exhibit Results are obtained and provided for the case 2 where the Snapshot of the real-time hardware setup of the proposed PV emulator is
temperature is varied at kept at 15 ◦ C, 25 ◦ C, 35 ◦ C and 45 ◦ C for every shown in Fig. 18(b). The experimental setup consists of step down
0.5 sec and maintaining irradiance and wind speed fixed at 1000 W/m2 transformer, bridge rectifier with filter circuit, DC-DC buck converter,
and 6 m/s. For the case 3 also results are obtained by varying the wind current sensor, voltage sensor, potentiometer, arduino mega control
speed and keeping at 5 m/s, 7 m/s, 9 m/s and 11 m/s for every 0.5 sec board, driver circuit and electrical load. The specification of the com­
maintaining irradiance and temperature at 1000 W/m2 and 25 ◦ C ponents used in the proposed PV emulator is shown in the Table 7. The
operating conditions. From the response of PV emulator voltage, current proposed PI voltage and current controller is implemented in the
and power, it is observed that the maximum power effectively extracted Arduino Mega board using MATLAB simulink environment and it is
from the PV emulator during charging conditions of the battery. The shown in Fig. 18(c). Irradiance, temperature and wind speed input

483
A. Nazar Ali et al. Solar Energy 220 (2021) 473–490

Fig. 16. Testing of MPPT algorithm with PV emulator.

Table 4
Actual and theoretical Power of PV emulator & MPPE ratio.
G (W/m2) Ppve (W) Pth (W) %η T (◦ C) Ppve (W) Pth (W) %η W (m/s) Ppve (W) Pth (W) %η

400 101.9 103.21 98.73 15 263.5 265.3 99.32 5 238.9 252.8 94.50
800 150.7 152.8 98.63 25 252.3 253.3 99.61 7 242 251.8 96.11
600 197 200.8 98.11 35 240.4 247.6 97.09 9 245.2 250.6 97.85
1000 240.4 252.3 95.28 45 228.4 241.2 94.69 11 248.3 249.6 99.48

command are given from potentiometer to analog port 0, 2 and 4 of voltage of the converter. The PV reference model is used to provide the
Arduino board and then it is converted into suitable form of PV reference PV reference current based on the potentiometers settings. The current
model. and voltage of the actual and reference value is processed through PI
The converter voltage and current is sensed through voltage divider controller to generate the control signal (Duty cycle). The control signal
circuit and ACS 712 model. Then it is processed through analog port of 6 is processed via PWM block to generate PWM pulses for buck converter
and 8 of the Arduino Board and it is converted into actual current and to emulate the behavior of the PV reference model. The PWM signal out

484
A. Nazar Ali et al. Solar Energy 220 (2021) 473–490

Table 5
Power of PV emulator & converter and converter efficiency.
G (W/m2) Ppve (W) Pcon (W) %η T (◦ C) Ppve (W) Pcon (W) %η W (m/s) Ppve (W) Pcon (W) %η

400 101.9 99.7 97.84 15 263.5 259.1 98.33 5 238.9 234.9 98.33
800 150.7 147.7 98.01 25 252.3 248.1 98.34 7 242 238.1 98.39
600 197 193.5 98.22 35 240.4 236.5 98.38 9 245.2 241 98.29
1000 240.4 236.5 98.38 45 228.4 224.4 98.25 11 248.3 244.2 98.35

Fig. 17. Testing of Solar MPPT battery charger controller using PV emulator.

via digital pin4 of the Arduino Board. the load test of PV emulator. In Case 1, PV emulator is loaded with
Load test is carried out to find the IV and PV characteristics of the resistive load with various irradiance conditions (T and W fixed at 25 ◦ C
developed PV emulator hardware model. Three cases are considered for and 6 m/s) and its corresponding results such as PV emulator voltage,

485
A. Nazar Ali et al. Solar Energy 220 (2021) 473–490

Table 6
Power of PV emulator and battery charging & charging efficiency.
G (W/m2) Ppve (W) Pcon (W) %η T (◦ C) Ppve (W) Pcon (W) %η W (m/s) Ppve (W) Pcon (W) %η

400 101.6 96.12 94.61 15 259 243.1 93.86 5 235.1 220.3 93.70
800 149.7 141.2 94.32 25 247.9 232.7 93.87 7 238 223.2 93.78
600 194.8 183.2 94.05 35 236.5 221.8 93.78 9 241.1 226.2 93.82
1000 236.5 221.8 93.78 45 225 210.7 93.64 11 244.1 229.1 93.85

Fig. 18. Hardware arrangement of the proposed PV emulator.

current and power are measured. The IV and PV characteristic for case 1 model.
is shown in Fig. 19(a). In case 2, PV emulator connected with resistive The developed PV emulator is tested with MPPT controller for the
load is tested under various temperature conditions (G and W fixed at notified three cases. In case1, irradiance is varied from 500 w/m2 to
1000 W/m2 and 6 m/s) and the corresponding results such as PV 1000 w/m2 (Temperature and wind speed fixed at 25◦ and 6 m/s) and
emulator voltage, current and power are measured. The IV and PV the corresponding results such as PV emulator voltage, PV emulator
characteristic for case 2 is shown in Fig. 19(b). In case 3, PV emulator current, converter voltage and converter current are shown in Fig. 20(a).
loaded with resistive load is subjected to various wind speed conditions Power level of the PV emulator and converter at 500 W/m2 is 128.7 W
(G and T fixed at 1000 W/m2 and 25 ◦ C) and corresponding results such and 123.7 W respectively. Power level of the PV emulator and converter
as PV emulator voltage, current and power are measured. The IV and PV at 1000 W/m2 is 241.9 W and 232.7 W respectively. The efficiency of the
characteristic for case 3 is shown in Fig. 19(c). In real time, the devel­ system for this case is around 96.15%. In case2, Temperature is varied
oped PV emulator effectively emulates the behavior of the PV reference from 25◦ to 45◦ (Irradiance and wind speed fixed at 1000 W/m2 and 6

486
A. Nazar Ali et al. Solar Energy 220 (2021) 473–490

Table 7 and 234 W respectively. Power level of the PV emulator and converter at
Specification of components used in the proposed PV emulator. 11 m/s is 248.1 W and 240.3 W respectively. The efficiency of the sys­
S. Components Ratings tem for this case is around 96.68%.
No The PV emulator is tested with market available luminous battery
1 Transformer 230/35 V, 400 VA charger controller and corresponding results are presented in Fig. 21.
2 Bridge rectifier - RI 8BU 30 400–1800 V, 30A Three cases are considered which are same as the test conditions
3 Filter Capacitor 1000 µF, 63 V considered for MPPT controller testing. The results for case1 condition
4 Power MOSFET – SEMIKRON (SKM 200 V, 120A such as PV emulator voltage & current and charging voltage & current is
120B 020)
5 Converter Inductor 22 mH, 30A
shown in Fig. 21(a). Power level of the PV emulator and converter at
6 Converter Capacitor 4700 µF, 63 V 500 W/m2 is 118.5 W and 110.6 W respectively. Power level of the PV
7 Arduino Mega Board ATmega2560, 16 MHZ, Flash emulator and converter at 1000 W/m2 is 239.8 W and 224.9 W
Memory 256 KB respectively. The efficiency of the system for this case is around 93.6%.
8 Current sensor – ACS 712 20 A
The results for case2 condition such as PV emulator voltage & current
9 Voltage Sensor- Voltage divider 45 V to 5 V
10 Potentiometer – Pankaj model 0–5 V, 1 KΩ and charging voltage & current is shown in Fig. 21(b). Power level of the
PV emulator and converter at 25◦ is 239.3 W and 223.5 W respectively.
Power level of the PV emulator and converter at 25◦ is 217.1 W and
m/s) and the corresponding results similar to case 1 for PV emulator and 203.2 W respectively. The efficiency of the system for this case is around
converter are obtained and shown in Fig. 20(b). Power level of the PV 93.5%. The results for case 3 condition such as PV emulator voltage &
emulator and converter at 25◦ is 242.9 W and 234.5 W respectively. current and charging voltage & current is shown in Fig. 21(c). Power
Power level of the PV emulator and converter at 25◦ is 225.2 W and level of the PV emulator and converter at 6 m/s is 239.3 W and 223.5 W
217.7 W respectively. The efficiency of the system for this case is around respectively. Power level of the PV emulator and converter at 11 m/s is
96.6%. In case3, wind speed is varied from 6 m/s to 11 m/s (Irradiance 246.8 W and 230.7 W respectively. The efficiency of the system for this
and Temperature fixed at 1000 W/m2 and 25◦ ) and the results similar to case is around 93.45%. From this analysis, it is ascertained that PV
case 1 for PV emulator and converter are obtained and shown in Fig. 20 emulator performs well under all operating conditions considered in
(c). Power level of the PV emulator and converter at 6 m/s is 242.5 W simulation as well as in real time.

Fig. 19. IV and PV characteristics of proposed PV emulator in real time.

487
A. Nazar Ali et al. Solar Energy 220 (2021) 473–490

(a) Case1

(b) Case2

(c) Case3
Fig. 20. Voltage and current response of PV emulator and converter with MPPT controller.

The cost comparisons of the proposed PV emulator with already environmental conditions such as irradiance, temperature and wind
existing literature are presented in the Table 8. From this analysis, the speed environmental condition of the PV cell. The double loop controller
proposed PV Emulator cost is around only $ 201.3 but cost of the PV was implemented for DC-DC buck converter and parameters of the PI
emulator with dSPACE controller and RTDS Simulator is around $ controller were optimized using classical genetic algorithm to operate
5800.90 to $ 6902.3. Therefore, the proposed PV emulator can easily the PV emulator in three region constant voltage source region, constant
implement with low cost. current source region and variable voltage and current source region.
The simulink model of the proposed PV emulator was created using
5. Conclusion MATLAB software. The developed model was tested and IV and PV
characteristics are compared with PV reference model and observed that
The simulation and real time investigation of PV emulator perfor­ PV emulator has an emulation efficiency of 98.2%. The PV emulator was
mance has been analyzed and presented in this paper. The DC-DC buck also tested with MPPT algorithm and solar MPPT battery charging
converter was used to emulate the behavior of the real time PV panel. control algorithm and it has maximum power point extraction ratio of
The PV reference model was developed considering ambient 97.45% and charging efficiency of 93.9%. Finally, experimental

488
A. Nazar Ali et al. Solar Energy 220 (2021) 473–490

(a) Case1

(b) Case2

(c) Case3
Fig. 21. Voltage and current response of PV emulator and battery with battery charger controller.

validation of the proposed PV emulator performance has also been done

Table 8 and proved that the proposed PV emulator performed well under all
Cost comparison of PV emulator. considered operating conditions.
Reference, Year Material Used Cost
Declaration of Competing Interest
(Azharuddin et al., Power supply, Buck Converter, dSPACE $
2016) 6902.90
(Jayawardana et al., Power supply, Buck Converter, RTDS $ The authors declare that they have no known competing financial
2019) Simulator, 5800.90 interests or personal relationships that could have appeared to influence
Proposed Work Power supply, Buck Converter, Ardunio $ 201.3 the work reported in this paper.
Mega Board

489
A. Nazar Ali et al. Solar Energy 220 (2021) 473–490

References Ickilli, D., Can, H., Parlak, K.S., 2012. Development of a FPGA-based photovoltaic panel
emulator based on a DC/DC converter. In: Proc. Of 2012 38th IEEE Photovoltaic
Specialists Conference, Austin, TX, pp. 001417–001421.
Akbari-Dibavar, Alireza, Nojavan, Sayyad, Mohammadi-Ivatloo, Behnam, Zare, Kazem,
Iqbal, Muhammad Taqur, Tariq, Mohd, 2015. Analytical study of different DC-DC
2020. Smart home energy management using hybrid robust-stochastic optimization.
converter topologies for photo voltaic emulator. J. Automat. Syst. Eng. 9 (4),
Comput. Ind. Eng. 143.
245–254.
Atoche, Alejandro Castillo, Castillo, Javier Vázquez, Ortegón-Aguilar, Jaime, Carrasco-
Jayawardana, I., Man Ho, C.N., Pokharel, M., 2019. Design and implementation of
Alvarez, Roberto, Gío, Jesús Sandoval, Colli-Menchi, Adrian, 2015. A high-accuracy
switch-mode solar photovoltaic emulator using power-hardware-in-the-loop
photovoltaic emulator system using ARM processors. Solar Energy 120, 389–398.
simulations for grid integration studies. In: Proc. Of 2019 IEEE Energy Conversion
Ayaz, R., Nakir, I., Tanrioven, M., 2020. An improved matlab-simulink model of PV
Congress and Exposition (ECCE), Baltimore, MD, USA, pp. 889–894.
module considering ambient conditions. In: Int. J. Photoenergy, 2014, pp. 1–6.
Lu, Dylan D.C., Nguyen, Quang Ngoc, 2012. A photovoltaic panel emulator using a buck-
Azharuddin, Mohammed, Babu, Thanikanti Sudhakar, Bilakanti, Nishant,
boost DC/DC converter and a low cost micro-controller. Solar Energy 86 (5),
Rajasekar, Natarajan, 2016. A nearly accurate solar photovoltaic emulator using a
1477–1484.
dSPACE controller for real-time control. Electric Power Components Syst. 44,
Mohammed Azharuddin, S., Vysakh, M., Thakur, Harshal Vilas, Nishant, B., Sudhakar
774–782.
Babu, T., Muralidhar, K., Paul, Don, Jacob, Basil, Balasubramanian, Karthik,
Balato, Marco, Costanzo, Luigi, Gallo, Daniele, Landi, Carmine, Luiso, Mario,
Rajasekar, N., 2014. A Near Accurate Solar PV Emulator Using dSPACE Controller
Vitelli, Massimo, 2016. Design and implementation of a dynamic FPAA based
for Real-time Control. Energy Procedia 61, 2640–2648.
photovoltaic emulator. Sol. Energy 123, 102–115.
Moussa, I., Khedher, A., Bouallegue, A., 2019. Design of a low-cost PV emulator applied
Can, H., 2013. Model of a photovoltaic panel emulator in MATLAB-Simulink. Turkish J.
for PVECS. Electronics 8 (2), 232.
Electr. Eng. Comput. Sci. 21 (2), 301–308.
Piazza, Maria Carmela Di, Vitale, Gianpaolo, 2010. Photovoltaic field emulation
Carralero, L.L.O., e Silva, G.S. Barbara da S., Costa, F.F., Tahim, A.P.N., 2019. PV
including dynamic and partial shadow conditions. Appl. Energy 87 (3), 814–823.
Emulator Based on a Four-Switch Buck-Boost DC-DC Converter. In: Proc. Of 2019
Prasanth Ram, J., Manghani, Himanshu, Pillai, Dhanup S., Sudhakar Babu, T.,
IEEE 15th Brazilian Power Electronics Conference and 5th IEEE Southern Power
Miyatake, Masafumi, Rajasekar, N., 2018. Analysis on solar PV emulators: A review.
Electronics Conference (COBEP/SPEC), Santos, Brazil, pp. 1–5.
Renew. Sustainable Energy Rev. 81 (1), 149–160.
Chalh, A., Motahhir, S., El Hammoumi, A., El Ghzizal, A., Derouich, A., 2018a. A low-
Protopopova, Julia, Kulik, Sergey, 2020. Educational intelligent system using genetic
cost PV emulator for testing MPPT ALGORITHM. IOP Conference Series: Earth
algorithm. Procedia Comput. Sci. 169, 168–172.
Environ. Sci. 161.
Samosir, Ahmad Saudi, 2015. Modeling and evaluation of solar photovoltaic emulator
Chalh, A., Motahhir, S., El Hammoumi, A., 2018b. Study of a low-cost PV emulator for
based on simulink model. ARPN J. Eng. Appl. Sci. 10 (20), 10288–10295.
testing MPPT algorithm under fast irradiation and temperature change. Technol.
Shah, R., Rana, A., 2015. Voltage controlled buck converter based photovoltaic emulator.
Econ. Smart Grids Sustain. Energy 3 (11).
In: Proc. Of 2015 International Conference on Electrical, Electronics, Signals,
Chavarría, J., Biel, D., Guinjoan, F., Poveda, A., Masana, F., Alarcón, E., 2014. Low cost
Communication and Optimization (EESCO), Visakhapatnam, pp. 1–4.
photovoltaic array emulator design for the test of PV grid-connected inverters. In:
Ullah, N., Nisar, F., Alahmadi, A.A., 2020. Closed loop control of photo voltaic emulator
Proc. Of 2014 IEEE 11th International Multi-Conference on Systems, Signals &
using fractional calculus. IEEE Access 8, 28880–28887.
Devices (SSD14), Barcelona, pp. 1–6.
Veerachary, M., 2006. PSIM circuit-oriented simulator model for the nonlinear
Cirrincione, M., Di Piazza, M.C., Pucci, M., Vitale, G., 2008. Real-time simulation of
photovoltaic sources. IEEE Trans. Aerosp. Electron. Syst. 42 (2), 735–740.
photovoltaic arrays by growing neural gas controlled DC-DC converter. In: Proc. Of
Zhou, Z., Holland, P.M., Igic, P., 2014. MPPT algorithm test on a photovoltaic emulating
2008 IEEE Power Electronics Specialists Conference, Rhodes, pp. 2004–2010.
system constructed by a DC power supply and an indoor solar panel. Energy Convers.
Ebrahim, A.F., Ahmed, S.M.W., Elmasry, S.E., Mohammed, O.A., 2015. Implementation
Manage. 85, 460–469.
of a PV emulator using programmable DC power supply. In: South east Con 2015,
Fort Lauderdale, FL, pp. 1–7.

490