Design and Simulation of a Photovoltaic and Fuel

Cell Based Micro-grid System

Amit Kumer Podder Kawser Ahmed Naruttam Kumar Roy
Dept. of Electrical and Electronic Dept. of Electrical and Electronic Dept. of Electrical and Electronic
Engineering Engineering Engineering
Khulna University of Engineering & Khulna University of Engineering & Khulna University of Engineering &
Technology, Technology, Technology,
Khulna, Bangladesh Khulna, Bangladesh Khulna, Bangladesh
E-mail : [email protected] E-mail : [email protected] E-mail : [email protected]

Md. Habibullah
Dept. of Electrical and Electronic
Engineering
Khulna University of Engineering &
Technology,
Khulna, Bangladesh
E-mail : [email protected]

Abstract— Energy is one of the primary needs for survival of all harmful by-products are released in the generation of electricity.
organisms in the universe. Energy has dependably been a In the same manner, the hydrogen fuel cell produces pure water
noteworthy need for the everyday human life. For socio-economic as a by-product. This is high time to seek an alternative way to
advancement, the non-stop supply of adequate energy is vital. The generate electricity using clean energy. One solution which is
instinctive assets are constrained and are being depleted at such a gaining popularity is a micro-grid system. A micro-grid is a
flustering rate, it is expected that the petroleum gas, which is the distributing system which provides low voltage or medium
foremost energy source of Bangladesh, will be exhausted by the year voltage supply from a bunch of micro sources in a controlled and
2031. To solve this problem, the micro-grid is one of the best
coordinated way by being connected to the main power grid,
solutions. This paper presents the design and simulation of a hybrid
micro-grid system based on photovoltaic and fuel cell. The battery operated in islanded mode or be completely off-grid [2].
system is also incorporated in order to use the system in an A lot of studies have been done on the hybrid micro-grid
emergency case. The performance of the system is evaluated at system. The performance and cost viability of a PV, fuel cell
different loading conditions. Furthermore, on-grid stability of the (FC) and battery based hybrid grid-tied system with intermittent
system is tested by connecting it to the IEEE 13 bus system. Load load conditions are analyzed in [3]. Four progressive modeling
flow analysis is done in order to ensure the power penetration from stages are designed and five key metrics are utilized in this paper
the proposed system to the grid. The design and simulation work is in order to determine the most effective micro-grid system. A
performed in the MATLAB/SIMULINK platform and found viable cost-effective and environment-friendly system is also
satisfactory performance.
proposed. Integration of PV array along with FC and ultra-
Keywords—Micro-grid, Photovoltaic, Fuel cell, Load flow analysis. capacitor system for energy production using short and long-term
storage capacity and appropriate control strategy is proposed in
I. INTRODUCTION [4]. The performance and modeling of a smart grid monitoring
system with separate solar plants (with a rated capacity of
Considering the generation of electricity, first thing comes 25.5kW) and loads at different locations is analyzed in [5]. The
up to our mind is fossil fuel as the traditional power grid is based transmission and distribution model with realistic impedance
upon coal, oil, gas etc. But these sources are questioned because values of a 25km long transmission line using SCADA and PLC
of their not being sustainable towards the environment. Besides, is also designed for the system. A two-level (control level and
in upcoming years, the limited fossil fuel will be finished, system level) strategy for energy management of DC micro-grid
increasing pollution in the environment and causes global is presented in [6] for PV/FC/battery based system. Dynamic
warming. To prevent these consequences, the world has to shift management of power for PV and proton exchange membrane
towards renewable energy like solar energy, wind power, FC based AC/DC micro-grid system is presented in [7]. In the
hydroelectric energy, geothermal power etc. Among these, solar aforementioned studies, the behavior of the systems has not been
energy and hydrogen fuel cell is highly acceptable as they are analyzed if those are connected to an IEEE 13 bus system.
efficient, reliable, eco-friendly, zero combustion and having high This paper presents a model consists of a solar array, battery
power density [1]. The key feature of solar energy is that no storage, and fuel cell stack. These are connected with DC/DC

978-1-7281-1126-1/19/$31.00 ©2019 IEEE

boost converter to gain optimum voltage level and to stop the number of panel, photo current ( ), and saturation current ( ),
fluctuation as they are connected with dc bus (400V). The dc ideality factor (a), series ( ) and parallel ( ) resistances can be
voltage is applied to a three-phase dc-ac converter to produce a presented as
380V ac voltage. A transformer is utilized for increasing voltage + × × ×
level for connecting it to the utility/main grid or it can also be = × − × −1
× ×
used in home appliances without increasing voltage level a bit. ×
×
This micro-grid system has been designed to generate − (3)
approximately 73kW in both islanded and grid-connected mode. ×
The performance of the proposed system is analyzed by
connecting it at different loading conditions both in standalone The specifications of the selected PV module named as Samsung
and grid-connected mode. Moreover, the load flow analysis is Sun Power T5-SER-235P are presented in Table I.
performed after the grid connection. It is anticipated that this
model will lay the basis for building up a very advanced small
B. Fuel Cell
scale system.
The most common fuel cell among other fuel cell is proton
II. SYSTEM MODELLING exchange membrane fuel cells (PEMFCs). It is very common
The system modelled in this work includes a PV array, a fuel because it can operate at a lower temperature (-20 to 1000C) and
cell, battery storage, dc-dc converter, three-phase dc-ac inverter, can start its operation very fast from its idle condition to full load
dc load, ac load, transformer, and the utility grid. The condition.
TABLE I
SPECIFICATIONS OF THE PV MODULE SAMSUNG SUN POWER T5-SER-
235P [8]
Parameters Value
Maximum power 235.0658W
Open circuit voltage, Voc 36.96V
Short-circuit current, Isc 8.48A
Voltage at maximum power point, Vmp 29.42V
Current at maximum power point, Imp 7.99A
Temperature coefficient of Voc -0.34391%/deg.C
Temperature coefficient of Isc 0.046958%/deg.C
Cells per module, (Ncell) 60
Diode ideality factor 0.97118
Series resistance, Rs 0.395020 ohm
Shunt resistance, Rsh 238.76 ohm

Fig. 1: Block diagram representation of the overall designed system

The fuel cell used in the hybrid system is 6kW. The stack fuel
representation of the designed system in a block diagram is cell model provided by Sim Power System of MATLAB/
shown in Fig. 1. The system components are delineated below. SIMULINK platform is utilized in the system. The parameters of
the fuel cell model can be obtained from the simple polarization
A. PV array curve or the datasheet provided by the manufacturer. Due to the
In a PV system, the solar energy is converted to electrical energy activation losses and losses caused by diffusion and resistivity,
by using one or more PV modules. Mainly, the system consists the output voltage of the cell can be expressed as
of panels, and various mechanical electrical connectors in order = − . − (4)
to produce the desired output. The panels are connected in series
and parallel connection to provide the desired amount of voltage where, and A are the empirical coefficients obtained from a
and current. The output current of a single PV module is non-linear regression method detailed in [9], and are the
expressed as [5] fuel cell and exchange current in ampere, denotes the
resistance of diffusion and combined cell and represents the
= − exp
( )
−1 −( + )/ ) (1) settling time of the cell due to a change in the current. The
ultimate output voltage of the stack full cell can be expressed as

where, q is the charge of the electron and is the Boltzmann = × (5)

constant. Since the value of irradiance and temperature is not where, the number of cells are represented by N. The fuel cell
constant over the time, the photocurrent ( ) also varies and stack model [9] is shown in Fig. 2.
expressed as [5]
C. Battery Model
=[ + ( − )] (2)
Due to high efficiency, high density of energy and environment -
friendly nature compared to lead acid or NiCd battery, Li-ion
where, the short circuit current is denoted by , and battery is chosen here for the proposed work. The available
represents the working and reference temperature of the cell, battery model in the Sim Power System is utilized in the work
is the co-efficient of temperature at and is the value of which is based on the modified Shepherd curve fitting model [9]
irradiance in kW/m2. Now, the expression of current derived as shown in Fig. 3. The expression of battery voltage ( ) for the
from the PV array in terms of series ( ) and parallel ( ) selected Li-ion battery can be presented as [9]
 to generate a constant 400V dc voltage. The output voltage
waveform from the dc-dc converter is shown in Fig. 4.

E. Voltage Source Inverter

The output voltage of the dc-dc converter is utilized as the input
voltage of the designed three-phase voltage source inverter. For
designing it in SIMULINK platform, six insulated gate bipolar
transistors (IGBTs) are used along with an anti-parallel diode for
each IGBT. The six gating signals of six IGBTs are generated by
the PWM gating signal generator with a switching frequency of 2
KHz.

Fig. 2: Fuel cell stack model [9]

Battery
FC
= − − . + exp(− . ) − . (6)

Voltage (V)
where, and represents the constant battery voltage and
polarization constant. , and denotes the actual and filtered
current of the battery, respectively. Q is the battery capacity, and PV
is the battery internal resistance while and denotes the
amplitude of exponential zone and inverse time constant of the
exponential zone, respectively. To show the effect of the state of
Time (seconds)
charge of the battery on the performance of it, an additional term,
Fig. 4: Output voltage waveform from the dc/dc boost converter of PV, FC and
, referred as polarization voltage is added in the Eqn. Battery.
(6). Again, instead of actual current, , filter current, are
utilized for the resistance of polarization which is indicated in the F. Load
Eqn.(6) by the term, . The output of the voltage source inverter is connected to a 268V,
50Hz, 43kW three phase ac load in order to test the system.

G. Three Phase Transformer

Subsequent to executing all the above models in

MATLAB/SIMULINK, the models are joined together to make a
dc bus. The produced dc voltage from these sources is
synchronized to a specified dc voltage (400V approx.) so that
one inverter can be utilized for the entire system. The inverter
output voltage is then filtered and the filtered output ac voltage
is 268V (r.m.s). The three-phase output voltage is then connected
to a three-phase transformer having delta-star configuration for
stepping up the 268V ac to 4160V ac in order to connect the
system to the grid.

H. IEEE 13 Bus Test Feeder System

Fig. 3: The battery model [9].
The IEEE 13 bus test feeder is utilized to test the performance of
the proposed system. The test feeder shows some fascinating
features. The 13 bus test feeder contains shunt capacitor banks,
D. DC/DC Boost Converter distributed and unbalanced load and in–line transformers. The
single line diagram of the test feeder with the proposed designed
system is shown in Fig. 5. The nominal voltage of the test feeder
For converting renewable energy, a dc-dc converter is an
is 4.16kV and it contains two in-line transformer and shunt
essential component of the micro-grid system. The dc-dc boost
capacitor banks. The parametric value of the transformer and
converter is utilized to provide a constant dc voltage to the dc- ac
shunt capacitor bank are presented in Table II and III [10]. The
inverter. The output voltage of the solar cell and fuel cell varies
line segment data i.e. the lengths of the line from Node A to
with the irradiation and amount of hydrogen and oxygen
Node B are presented in Table IV. Table IV also shows a
respectively. Again, the output voltage of the storage battery is
configuration code which is utilized to present the spacing of the
low. Hence, to eliminate the variation and low output voltage,
model. The parametric values of different distributed and spot
three dc-dc converters with closed loop configuration are utilized
loads at different nodes are presented in Table IV.
 TABLE II
PARAMETRIC VALUE OF IN-LINE TRANSFORMERS
kVA kV-high kV-low R- X-%
%
XFM -1 500 4.16 – Gr.W 0.48 – Gr.W 1.1 2
XFM -2 500 4.16 – Gr.W 0.48 – Gr.W 1.1 2

Voltage (V)
Voltage (V)
Voltage (V)
Time (seconds)

Fig. 6: Waveform for the input dc voltage, output inverter voltage, filtered ac
voltage across the load.

Fig. 5: The single line diagram of the IEEE 13 bus test feeder with the proposed
designed system.
III. SIMULATION RESULT
TABLE III
PARAMETRIC VALUE OF SHUNT CAPACITOR BANK
In this work, a 77.5kW solar array is designed with 66 solar
Node Ph-A Ph-B Ph-C panels connected in parallel and 5 solar panels connected in
(kVAr) (kVAr (kVAr series. A 6kW PEM FC and a 12V battery with 42000Ah rated
675 200 200 200 capacity are also designed for the backup of the PV array. A
611 100 closed loop dc-dc boost converter is connected to all sources i.e.
Total 200 200 300 PV array, FC, and the battery. The combined dc output voltage
from the PV, FC and battery is shown in Fig. 4. This combined
output voltage constitutes a 400V dc bus, which can be utilized
for dc appliances. This dc output voltage is utilized as the input
of the three-phase dc/ac inverter.

TABLE IV: TABLE FOR THE LINE SEGMENT, DISTRIBUTED LOAD AND SPOT LOAD DATA FOR THE IEEE 13 BUS TEST FEEDER [10].

Line Segment Data Distributed Load Data

Node Node Length Config. Node A Node B Load Ph-1 Ph-1 Ph-2 Ph-2 Ph-3 Ph-3
A B (ft.) Model (kW) (kVAr) (kW) (kVAr) (kW) (kVAr)
632 645 500 603 632 671 Y-PQ 17 10 66 38 117 68
632 633 500 602 Spot Load Data
633 634 0 XFM-1 Node Load Model Ph-1 Ph-1 Ph-2 Ph-2 Ph-3 Ph-3
(kW) (kVAr) (kW) (kVAr) (kW) (kVAr)
645 646 300 603 634 Y-PQ 160 110 120 90 120 90
650 632 2000 601 645 Y-PQ 0 0 170 125 0 0
684 652 800 607 646 D-Z 0 0 230 132 0 0
632 671 2000 601 652 Y-Z 128 86 0 0 0 0
671 684 300 604 671 D-PQ 385 220 385 220 385 220
671 680 1000 601 675 Y-PQ 485 190 68 60 290 212
671 692 0 Switch 692 D-I 0 0 0 0 170 151
684 611 300 605 611 Y-I 0 0 0 0 170 80
692 675 500 606 Total 1158 606 973 627 1135 753
Fig. 7: MATLAB/SIMULINK diagram for the proposed PV/FC based micro-grid system.

The input dc voltage waveform, the output voltage of the inverter

and voltage across the ac load are shown in Fig. 6. The overall
proposed system designed in MATLAB/SIMULINK platform is
shown in Fig. 7. A 43kW load is connected to the proposed
Power (W)

system in order to determine the response of the system. The

voltage and current waveform across the load is shown in Fig. 8.
The power consumed across the 43kW load is shown in Fig. 9. It
can be seen that the load consumed a stable power of 43kW after
a delay of 2.1s. This time delay happens because all three sources
contribute to meet the load demand at the same time which
causes a time delay for synchronizing themselves.

Time (seconds)

Fig. 9: Power consumed by the load of 43kW.

Current (A)

During the loading condition, the battery unit discharges in order

to supply the load demand. The state of charge of the battery
decreases as shown in Fig. 10, which indicates the discharging of
it. In order to test the stability of the system in grid connected
mode, the proposed system is connected to the 680 bus of the
IEEE 13 bus system. This IEEE 13 bus test feeder is available in
the MATLAB as a default bus system. Therefore, the load flow
Voltage (V)

analysis of this default bus system is done without and with the
proposed hybrid system. The load voltage and current while
connecting the system with the IEEE 13 bus system are shown in
Fig. 11. The load flow data before and after connecting

Time (seconds)

Fig. 8: Load current and load voltage waveform across the 43kW load.
 to the IEEE 13 bus, the system yields stable voltage and current
as shown in Fig. 11.

IV. CONCLUSIONS
SOC (%)

This paper presents an approximate 73kW hybrid micro-grid

system designed with PV, FC, and battery storage system. In
order to test the system’s performance in off-grid mode, a 43kW
ac load is connected to the proposed system via a dc/ac inverter.
It is shown that the system yields stable output voltage and
current. Moreover, the stability of the system is validated by
connecting it to the IEEE 13 bus system, and an approximate of
Time (seconds) 73kW real power is added to the grid. Under this on-grid mode,
the proposed system also remains stable. Hence, the proposed
Fig. 10: State of charge of the battery unit. hybrid micro-grid system may be the future solution for the
ongoing power crisis in the universe.

REFERENCES
Current (A)

[1] A. K. Podder, K. Ahmed, N. K. Roy, and P.C. Biswas, “Design and

simulation of an independent solar home system with battery backup,”
Proc. in IEEE 4th ICAEE, Dhaka Bangladesh, 28-30 September, 2017, pp.
427-431.
[2] N. Hatzaiargyriou, H. Asano, R Iravani, and C . Marnay, “ Micro-grids:
An overview of ongoing research, development and demonstration
projects,” IEEE Power and Energy magazine, pp. 78 -94, July/August
2007.
Voltage (V)

[3] M. Patterson, N. F. Macia, and A. M. Kannan, "Hybrid microgrid model

based on solar photovoltaic battery fuel cell system for intermittent load
applications,” IEEE Transactions on Energy Conversion, vol. 30, no. 1,
pp. 359-366, March 2015.
[4] M. Uzunoglua, O.C. Onar, and, M.S. Alam, “Modeling, control and
simulation of a PV/FC/UC based hybrid power generation system for
stand-alone applications,” Renewable Energy, vol. 34, pp. 509-520, 2009.
Time (seconds)
[5] Y. Kabalci, and E. Kabalci, “Modeling and analysis of a smart grid
Fig. 11: Current and voltage waveform after connecting the proposed system to
monitoring system for renewable energy sources,” Solar Energy, vol.153,
the IEEE 13 bus system.
pp. 262–275, 2017.
[6] Y. Han, W. Chen, Q. Li, H. Yang, F. Zare, and Y. Zheng, “Two-level
the proposed system to the IEEE 13 bus test feeder is shown in energy management strategy for PV-Fuel cell-battery-based DC micro-
Table V. It can be seen that that the real power before and grid,” International Journal of Hydrogen Energy, 2018
https://doi.org/10.1016/j.ijhydene.2018.04.013.
[7] R. K. Sharma and S. Mishra, “Dynamic power management and control of
TABLE V: PARAMETRIC VALUE OF THE LOAD FLOW ANALYSIS a PV PEM fuel-cell-based standalone ac/dc micro-grid using hybrid energy
BEFORE AND AFTER CONNECTION OF THE PROPOSED SYSTEM TO storage,” IEEE Transactions on Industry Applications, vol. 54, no. 1, pp.
THE IEEE 13 BUS SYSTEM 526-538, Jan.-Feb. 2018.
Parameters Before connection After connection [8] The SolarDesignTool webpage [online], Available: http://www.
solardesigntool.com/components/module-panelsolar/Sunpower/1079/SER-
Total generation P = 3518.74 kW P = 3591.39 kW 235P/specification-data-sheet.html.[ Accessed on May, 2018].
Q = 1540.14 kvar Q = 1525.74 kvar [9] S. N. Motapon, L. Dessaint and K. Al-Haddad, “A comparative study of
Total PQ load P = 3101.90 kW P = 3101.87 kW energy management schemes for a fuel-cell hybrid emergency power
Q = 1880.42 kvar Q = 1880.39 kvar system of more-electric aircraft,” IEEE Transactions on Industrial
Total Z shunt load P = 363.47 kW P = 434.73 kW Electronics, vol. 61, no. 3, pp. 1320-1334, March, 2014.
Q = -479.42 kvar Q = -498.59 kvar [10] IEEE Distribution Planning Working Group Report, “Radial distribution
Total losses P = 53.36 kW P = 54.79 kW test feeders,” IEEE Transactions on Power Systems, vol. 6, no.3, pp 975-
Q = 139.14 kvar Q = 143.93 kvar 985, August, 1991.

after connection of the proposed system is 3518.74 and

3591.39kW, respectively. Therefore, 72.65kW power is added to
the grid because of the proposed system. The total PQ load
power remains the same but the real power consumed by the total
Z shunt load is increased from 363.47kW to 434.73kW as shown
in Table V. Therefore, from the generated 72.65kW power due to
the proposed system, 71.26kW power is consumed by the Z
shunt load. The rest power of 1.43kW is lost by the system which
is indicated by the value of total power losses before and after
connection of the proposed system to the test feeder as shown in
Table V. Note that, after connecting the proposed hybrid system