Professional Documents
Culture Documents
*Faculty of Electrical Engineering, Universiti Teknologi MARA, Pasir Gudang, Johor, Malaysia
**Faculty of Electrical Engineering, Universiti Teknologi MARA, Shah Alam, Malaysia
***Faculty of Applied Science, Universiti Teknologi MARA, Shah Alam, Malaysia
atiqah.hamizah@gmail.com, ambomaliki@gmail.com, hedzlinzainuddin@yahoo.com
‡
Corresponding Author: Atiqah Hamizah Mohd Nordin; Universiti Teknologi MARA, Pasir Gudang, Johor,
Tel: +60 73818195, atiqah.hamizah@gmail.com
Abstract- This paper presents the development of inverter simulation model in Grid-Connected Photovoltaic System (GCPV)
in Matlab/Simulink software. This work is a part of the development of a complete GCPV system simulation model. The inverter
simulation model accepts solar irradiance and temperature parameters as inputs (at PV array) and outputs AC power. A phase-
shift control method is used to ensure that the inverter output is synchronized with unity power factor to the utility grid. The
developed model is validated against data obtained from an actual installed GCPV system at Malaysia Green Technology
Corporation (MGTC) in Bangi, Malaysia with system capacity of 45.35 kWp. The validation of model is through Root Mean
Square Error (RMSE) and Mean Absolute Percentage Error (MAPE). Results showed that inverter output power from simulation
is acceptable with small deviation from the actual data. This is due to inaccuracies of predicting de-rating factors listed in this
work.
Keywords—Photovoltaic; solar; grid inverter; simulation; de-rating factor.
simulation model [4]. The developed model is validated by at the Pack A1 PV system consists of three units of single-
comparing the simulation output with the data obtained phase inverter. This study however will concentrate only on
from an installed GCPV system at Malaysia Green one of the single-phase inverter units. Therefore the output
Technology Corporation (MGTC) in Bangi, Malaysia. The power of the inverter from the simulation will be converted
installed GCPV system which is also known as Pack A1 into the equivalent three-phase value, to compare with the
system has an installed capacity of 45.35 kWp. actual values from the actual data which is available in
three-phase values.
2. Methodology The inverter model in this study is designed by referring
to the specification of the Pack A1 system’s inverter which
2.1 Inverter Desgin is Fronius IG500 as listed in Table 1.
The centralized inverter topology is applied at the Pack
A1 PV system at MGTC. The centralized inverter topology
Table 1. Specification of Fronius IG500 Inverter
Subjects Descriptions
Power factor 1
950
INTERNATIONAL JOURNAL OF RENEWABLE ENERGY RESEARCH
Atiqah Hamizah Mohd Nordin et al. ,Vol. 4, No. 4, 2014
951
INTERNATIONAL JOURNAL OF RENEWABLE ENERGY RESEARCH
Atiqah Hamizah Mohd Nordin et al. ,Vol. 4, No. 4, 2014
Fig. 4. Equivalent circuit at AC side of the inverter Figure 6. Vinv, Igrid and Vgrid in phasor diagram
In order to achieve the unity power factor, the The phase angle of modulating signal (Vm) is varied
waveform of Iinv must be in phase with the waveform of with solar irradiance and temperature level in order to get
Vgrid as illustrated in Fig. 5. In the phasor diagram, it is the Iinv synchronized with the Vgrid. The appropriate phase
represented as in Fig. 6. As in Fig. 6, Vinv has to lead Vgrid angle for a certain range of temperature and irradiance level
has to be set. In the developed model, the phase angle of V m
with the angle in order to get the Iinv in phase with the
will automatically change to the appropriate value
Vgrid. This is achieved by adjusting the phase angle of the
according to the range of cell temperature level. For
modulating signal (Vm).
simplicity, since there is a quite linear relationship between
V, I
Vinv solar irradiance and cell temperature, the developed model
I inv
will only detect the solar irradiance value for controlling the
Vgrid selection criteria of phase angle value.
t
2.4 Simulation Work
The inverter is modeled in circuit-based approach as
shown in Fig. 7. The ‘inverter switching’ subsystem in Fig.
Fig. 5. Waveforms of Vinv, Iinv and Vgrid
7 contains the switching signal generation for the IGBTs
and also the subsystem of the grid synchronisation control
as depicted in Fig. 8. The repeating sequence block is set to
generate triangular carrier signal (Vcr) with the frequency of
10 kHz. The sinusoidal modulating signal (Vm) is generated
using the sinusoidal AC voltage source block with the
frequency of 50 Hz. The relational operator is used as the
comparator to compare Vcr and Vm to produce the Vg. The
‘NOT logical operator’ block is used to invert the Vg signal
for the inverter switching.
952
INTERNATIONAL JOURNAL OF RENEWABLE ENERGY RESEARCH
Atiqah Hamizah Mohd Nordin et al. ,Vol. 4, No. 4, 2014
Out 1 [G1]
[Irradiance] In1
Out 2 [G2]
Inverter switching
Decoupling [ Igrid]
inductor
3 i
+
-
Conn3
[ G1] [ G2]
C
C
g
g
IGBT1_2 IGBT2_3
E
E
+
v [ Vinv ] +
-
Load -
v [Vgrid]
C
g
IGBT2_2 IGBT2_ 4
E
Conn 4
4
Inverter LC filter
1 In1 Out1
Vm,inv
Irradiance
Synchronization Comparator
control
> NOT 1
G1
2
G2
Vcr,inv
953
INTERNATIONAL JOURNAL OF RENEWABLE ENERGY RESEARCH
Atiqah Hamizah Mohd Nordin et al. ,Vol. 4, No. 4, 2014
Decoupling [ iinv ]
Inductor
+ i
-
25+273.15 T + + +
Temperature
[K] + v Load +
- [vinv] - v [ vgrid]
1000 G - - -
Irradiance
[Wm-2]
PV Array MPPT Inverter
Low-pass Grid
Filter
Fig. 9. Measurement positions of Iinv, Vinv and Vgrid
software. It shows that the THD of the Iinv at T = 69.51 oC
Fig. 10 shows the sinusoidal waveforms of the Vgrid and and G = 808 Wm-2 is found to be 0.51 %. The THD of the
Iinv of the simulated model for the solar irradiance of 808 Iinv for several levels of solar irradiance input are also listed
Wm-2 and temperature of 69.51oC. It is observed that the Iinv in Table 3. It is observed that the THD values of I inv from
waveform is in phase with the Vgrid waveform, which means the simulation for varying solar irradiance level are less than
that the inverter has synchronized to the grid. Fig. 11 shows the maximum THD allowed according to the IEC 61727
the frequency spectrum of Iinv until 40th order of harmonics, standard which is 5 % [7].
-2 o
at G = 808 Wm and T = 69.51 C which is obtained from
Fast Fourier Transfer Analysis tool in Matlab Simulink
400
300
Vgrid
200
Iinv
100
Voltage [v],
Current [I]
-100
-200
-300
-400
1.44 1.45 1.46 1.47 1.48 1.49 1.5
Time [s]
954
INTERNATIONAL JOURNAL OF RENEWABLE ENERGY RESEARCH
Atiqah Hamizah Mohd Nordin et al. ,Vol. 4, No. 4, 2014
50
-50
0 0.5 1 1.5
Time (s)
100
Mag (% of Fundamental)
80
60
40
20
0
0 200 400 600 800 1000 1200 1400 1600 1800 2000
Frequency (Hz)
955
INTERNATIONAL JOURNAL OF RENEWABLE ENERGY RESEARCH
Atiqah Hamizah Mohd Nordin et al. ,Vol. 4, No. 4, 2014
(RMSE) and Mean Absolute Percentage Error (MAPE). factors such as manufacturer’s tolerances and module
The RMSE is frequently used to measure the difference mismatch losses (fmm), losses in cables (fcable), dirt on the
between the values simulated from the model and the actual surface of PV module (fdirt) and inverter efficiency (inv).
values obtained from the real system that is being modeled Table IV summarized the value of inv, fmm, fdirt, and fcable
and is defined in Eq. (4). The calculated RMSE value has which are used in the simulation of the developed model.
unit in Watt (W). MAPE is a measure of accuracy in a fitted- The values of fmm and inv can be obtained from the PV
time series value and usually is expressed as a percentage. module and inverter datasheets. The values of fdirt and fcable
In MAPE, the absolute values of all percentage errors are are assumed based on Malaysia’s climate, frequency of
summed up and the average is computed as represented in rainfall, and has been used by several researchers [8].
Eq. (5). MAPE is also often useful for purposes of reporting, Table 4. De-rating factors
because it is expressed in generic percentage terms that will De-rating factors Values
be understandable to a wide range of users. The RMSE of inv, 0.943
Pac_sim prior to Pac_actual is found to be 5920.646 W and the
fdirt 0.970
MAPE is 20.744 %.
fcable 0.980
2
∑𝑛
𝑖=1(𝑃𝑎𝑐𝑡𝑢𝑎𝑙_𝑖 −𝑃𝑠𝑖𝑚_𝑖 ) fmm 0.950
𝑅𝑀𝑆𝐸 = √ (4)
𝑛
The simulation output power need to be multiplied by
Where,
fmm, fdirt, fcable and also inv as in Table 4. In reality, some of
Pactual_i = actual power,
those parameters are difficult to estimate accurately such as
Psim_i = simulated power,
dirt factor, module mismatch and cable loss. The graph in
n = number of samples.
1 𝑃𝑑𝑎𝑡𝑎_𝑖 −𝑃𝑠𝑖𝑚_𝑖
Fig. 13 shows the comparison between the values of the
𝑀𝐴𝑃𝐸 = ∑𝑛𝑖=1 | | (5) simulation result with consideration of the de-rating factors
𝑛 𝑃𝑑𝑎𝑡𝑎_𝑖
Where, (Pac_sim_derate) and the actual values from the data (P ac_actual),
Pactual_i = actual power, for varying solar irradiance. It is observed that with the
Psim_i = simulated power, consideration of the de-rating factors, Pac_sim_derate is
n = number of samples. approaching to Pac_actual. The RMSE between Pac_sim_derate and
Pac_actual is 1790.783 W and the MAPE is 5.480 %. It is
3.2 With Consideration of De-rating Factors observed that the RMSE and MAPE are reduced after taking
In reality, it is important to note that in the actual grid- into accounts the de-rating factors.
connected PV system, not all power generated from the PV
array could be transferred into the grid. There are power
losses that occur in the PV system due to some de-rating
Fig. 13. AC output power with consideration of de-rating factors against solar irradiance
connected PV power system has been developed and
4. Conclusion demonstrated. The THD of the inverter output current I inv
for varying solar irradiance input is below the maximum
The single-phase full-bridge inverter for grid- THD allowed based on the IEC 61727 standard which is 5
956
INTERNATIONAL JOURNAL OF RENEWABLE ENERGY RESEARCH
Atiqah Hamizah Mohd Nordin et al. ,Vol. 4, No. 4, 2014
%. The inverter output power is also compared to the AC acceptance test of grid-connected systems: Ministry of
power obtained from the actual data using RMSE and energy, green technology and water Malaysia / MBIPV
MAPE. It is also found that the inverter output current and project, 2010.
the grid voltage is synchronized. The consideration of fmm,
fdirt, fcable and inv at the simulation output is also important
in order to get more accurate results. The RMSE and MAPE
for both DC and AC sides are decreased with the
consideration of these de-rating factors.
Referencence
[1] S. Shaari, A. M. Omar, A. H. Haris, and S. I. Sulaiman,
Solar Photovoltaic Power: Fundamentals: Ministry of
Energy, Green Technology and Water,
Malaysia/Malaysia Building Integrated Photovoltaic
Project, 2010.
[2] S. B. Kjaer, J. K. Pedersen, and F. Blaabjerg, "A
review of single-phase grid-connected inverters for
photovoltaic modules," Industry Applications, IEEE
Transactions on, vol. 41, pp. 1292-1306, 2005.
[3] M. Calais, J. Myrzik, T. Spooner, and V. G. Agelidis,
"Inverters for single-phase grid connected photovoltaic
systems-an overview," in Power Electronics Specialists
Conference, 2002. pesc 02. 2002 IEEE 33rd Annual,
2002, pp. 1995-2000.
[4] A. H. M. Nordin and A. M. Omar, "Modeling and
simulation of Photovoltaic (PV) array and maximum
power point tracker (MPPT) for grid-connected PV
system," in Sustainable Energy & Environment
(ISESEE), 2011 3rd International Symposium &
Exhibition in, pp. 114-119.
[5] A. Malfait, R. Reekmans, and R. Belmans, "Audible
noise and losses in variable speed induction motor
drives with IGBT inverters-influence of the squirrel
cage design and the switching frequency," in Industry
Applications Society Annual Meeting, 1994.,
Conference Record of the 1994 IEEE, 1994, pp. 693-
700 vol.1.
[6] M. H. Nehrir and C. Wang, Modeling and control of
fuel cells: distributed generated applications: John
Wiley & Sons, 2009.
[7] "Photovoltaic (PV) systems - Characteristics of the
utility interface," in IEC 61727, 2004.
[8] S. Shaari, A. M. Omar, A. H. Haris, and S. I. Sulaiman,
Solar photovoltaic power: testing, commisioning and
957