2011 2nd Power Electronics, Drive Systems and Technologies Conference

Grid Connected DG Systems Based on Z-Source NPC

Inverters
S.M. Dehghan

E. Seifi
M. Mohamadian
R Gharehkhani
Electrical & Computer Engineering Faculty, Tarbiat Modares University, Tehran, Iran
[email protected], [email protected], [email protected], [email protected]

Abstract-- Z-source neutral-point-clamped (NPC) inverters have

been recently presented as a buck-boost three-level inverter. In
this paper five new grid-connected distributed generation (DG)
systems are proposed using z-source NPC inverters. With the
proposed systems, output power of two separate DG systems such
as photovoltaic (PV) arrays and fuel cell (FC) stacks are
extracted and delivered to the grid. Since NPC inverter consumes
same power from both DC inputs, different control systems are
used for each of five proposed topologies. The performance of the
proposed topologies is verified by simulation results.
Keywords-component; Multilevel inverters, grid-connected
distribution generation, fuel cells, photovoltaic cells, power
conversion.

I. INTRODUCTION

Figure 1. Dual z-source network three-level NPC inverter (DZSN-NPCI).

With the increasing concern about global environmental

protection and also growing demand for energy, distributed
generation (DG) systems based on renewable energy sources
have significantly increased throughout the world. Most of DG
systems such as photovoltaic (PV), wind turbine (WT) and
fuel cell (FC) stack, require a two-stage power electronic
converter. An inverter is used to interface with the utility grid.
This converter transfers the active power from DG to the grid
and also controls the reactive power. Besides, another
converter is required to extract maximum power from
renewable source or keep the output voltage constant.
However two-stage converters are expensive, less efficient and
difficult to control [1]. Recently the z-source inverter has been
used as replacement of two-stage converters in the DG
systems [2, 3 & 4]. A z-source network is used as a boost
converter in the z-source inverter. Thus this inverter can be
considered as a buck-boost converter that can exactly satisfy
requirement of the grid connected DG systems.

Figure 2. Single z-source network three-level NPC inverter (SZSN-NPCI).

In this paper, z-source NPC inverters are employed to

connect two separate renewable sources to the utility grid. As
a result, five different topologies based on DZSN-NPCI and
SZSN-NPCI inverters are proposed. The proposed topologies
present several advantages such as connecting two DG
systems to the grid with only one converter that reduces
volume and cost, increases efficiency and facilitates the
control. The power quality of current injected to the grid can
be improved because of three-level structure. In addition, a
hybrid energy system can be developed via the proposed
topology. While hybrid DG system normally use several
converters [10 & 11], using the proposed topologies, only one
converter is required.

Three-level neutral-point-clamped (NPC) inverters, having

many inherent advantages, are commonly used as the
preferred topology for medium voltage ac drives. These
inverters have recently been explored for other low-voltage
applications including grid-interfacing power converters to
improve the power quality [5, 6 & 7]. Z-source network is also
used in NPC inverters. Fig. 1 and Fig. 2 show two types of zsource NPC inverters: dual z-source network NPC inverter
(DZSN-NPCI) [8] and single z-source network NPC inverter
(SZSN-NPCI) [9], respectively.

978-1-61284-421-3/11/$26.00 2011 IEEE

104

II. Z-SOURCE NPC INVERTER TOPOLOGY

DZSN-NPCI includes two z-source networks connected to
upper and lower DC links of NPC inverter (Fig. 1). The zsource networks are supplied by two separate DC sources.
Normally, the z-source network can boost voltage of DC
inputs. Consequently two boost factors can be defined for zsource NPC inverter as:

ViU = BU VoU =

VoU
1 2(TSCOU + TSCB ) / T

(1)

ViL = B L VoL =

VoL
1 2(TSCOL + TSCB ) / T

(2)

where VoU, VoL are DC input voltages and ViU, ViL are DC
link voltages. TSCOU, TSCOL and TSCB are the shoot-through time
intervals of only upper z-source network output, only lower
output and both outputs, respectively. For low output THD,
boost factors should be determined such that voltage of both
DC links of inverter would be the same. In NPC inverter, if
both DC links are the same, equal powers are extracted from
both DC links.

Figure 3. Simple carried-based switching method for z-source NPC

inverters.

For SZSN-NPCI, STU and STL are related to R and can be

derived as:

SZSN-NPCI uses a dual-input dual-output z-source

network (Fig. 2). Outputs of this z-source network are
connected to upper and lower DC links of NPC inverter.
Output voltages of DIDO-ZSN can be expressed by:

ViU = VC VoL

(3)

ViL = VC VoU

(4)

Therefore voltage of both DC inputs should be same for

low output THD. Consequently DC link voltage can be
expressed by [12]:

Vi = BVo =

T TSCOU

T
Vo
TSCOL 2TSCB

PU T + TSCOU TSCOL
=
PL T TSCOU + TSCOL

1 + BU
2 BU
1 + BL
STL =
2 BL

(9)

STL =

2 BR + R + 1
2 B (R + 1)

(10)

In regard to features of DZSN-NPCI and SZSN-NPCI

inverters, this paper proposes five topologies for gridconnected DG systems: 1) FC-FC system based on DZSNNPCI with constant power control (CPC) (Fig. 4) 2) Hybrid
PV-FC system based on DZSN-NPCI with maximum power
control (MPC) (Fig. 5) 3) Hybrid PV-FC system based on
DZSN-NPCI with CPC (Fig. 6) 4) FC-FC system based on
SZSN-NPCI with CPC (Fig. 7) 5) Hybrid PV-FC system
based on SZSN-NPCI with MPC (Fig. 8)

(5)

The control system in all proposed configurations consists

of two control bocks: 1) grid power control block 2) DG
source control block. The first control block generates the
reference signals (Vq* and Vd*) for the PWM generator. Using
the reference frame oriented along the grid voltage, the active
and reactive power may be expressed as [13]:

(6)

Fig. 3 shows a simple carried-based switching method for

z-source NPC inverters [12]. Conventional phase opposite
disposition level-shifted PWM (POD-LSPWM) method has
been modified for z-source NPC inverters using two extra
shoot-through signals STU and STL. When upper carrier is
greater than STU, an upper shoot-through can be implemented
if three switches SA1, SB1 and SC1 are turned ON. In a
similar manner, when lower carrier is less than STL, a lower
shoot-through is created by truing three switches SA2, SB2
and SC2 ON. For DZSN-NPCI, STU and STL are as:

STU =

2B + R + 1
2 B (R + 1)

III. CONFIGURATION AND SYSTEM CONTROL

Where Vo = VoU = VoL. Although voltage DC inputs of

SZSN-NPCI should be the same; power of them can be
different as ratio of powers (R) can be determined by [12]:

R=

STU =

P = (3 2) v d i d

(11)

Q = (3 2) v d i q

(12)

where v is grid voltage and i is the current to the grid. The

subscripts d and q stand for direct and quadrature
components, respectively. As suggested by (11) and (12),
active and reactive power control can be achieved by
controlling direct and quadrature current components,
respectively. For a reference reactive power, the reference qaxis current (iq*) is determined. To obtain unity power factor,
the reference q-axis current is set to 0. The reference d-axis

(7)
(8)

105

current (id*) is determined by the reference active power. The

proposed configurations use different methods to determine
the reference d-axis current.
Second control block determines the shoot-through signals
(STU and STL). In the continue the proposed configurations
are described in detail.

Figure 7. Fuel cell fuel cell (FC-FC) system based on SZSN-NPCI with
constant power control.

Figure 4. Fuel cell fuel cell (FC-FC) system based on DZSN-NPCI with
constant power control.

Figure 8. Hybrid photovoltaic fuel cell (PV-FC) system based on SZSNNPCI with maximum power control.

A. FC-FC system based on DZSN-NPCI with CPC

In this configuration, two FC stacks which can be different
are connected to z-source NPC inverter. A reference value (P*)
is set as desired delivered power to the grid. The reference daxis current (id*) is determined by (9). In this manner each of
FCs supplies half of P*. While FCs have a unique polarization
curve and their output voltage drops dramatically when the
output current increases [14], to maintain DC link voltages
constant, shoot-trough time intervals are controlled using
following equations:

Figure 5. Hybrid photovoltaic fuel cell (PV-FC) system based on DZSNNPCI with maximum power control.

STU = VCU Vi*

(13)

STL = VCL Vi*

(14)

where reference DC link voltage (Vi*) is a design parameter.

B. Hybrid PV-FC system based on DZSN-NPCI with MPC
In this configuration, PV array and FC stacks are
connected to upper and lower z-source networks, respectively.
The maximum power of PV is extracted and delivered to the
grid. In this topology, the output power of FC stack is equal to
output power of PV arrays. To extract maximum power from
PV array, the upper DC input voltage should be regulated to
the reference DC input voltage (VoU*) which is determined by

Figure 6. Hybrid photovoltaic fuel cell (PV-FC) system based on DZSNNPCI with constant power control.

106

block determine reference voltage of PV (VoU*) Since both dc

input voltages of SZSN-NPCI should be the same, like
previous topology a PI controller is used to determine suitable
power ratio. In this topology, boost factor is determined by:

a maximum power point tracking (MPPT) block [15]. For this

purpose, the upper shoot-trough signal is calculated by:
*
STU = VCU
Vi*

(15)

where the reference capacitor voltage (VCU*) is determined by:

B = Vi* (VC* Vi* )

*
*
VCU
= 0.5 (Vi* + VoU
)

where VC* is reference voltage of z-source capacitor and is

equal to:
*
(20)
VC* = VoU
+ Vi*
Shoot-through signals can be obtained by (9) and (10). Just
like second topology, a capacitor voltage close loop control is
used to determine the reference d-axis current (id*).

(16)

On other hand, to keep constant lower DC link voltage,

lower shoot-through time interval is controlled using (14).
Since the power injected to grid is not a specified value, a
capacitor voltage close loop control (using capacitor voltage of
upper z-source network) is used to determine the reference daxis current (id*) as shown in Fig. 5. In this manner, the
capacitors voltages of the upper z-source networks are
maintained constant. This assures that all the power generated
by the PV arrays is transferred to the grid.

IV. SIMULATION
To verify the performance of the proposed topologies,
several simulations have been performed. Table I shows
simulation parameters.

C. Hybrid PV-FC system based on DZSN-NPCI with CPC

This topology is similar to previous topology; however a
given constant power is delivered to the grid. This power is
twice the average generated power of PV array. Consequently
the reference d-axis current (id*) is determined by (11). While
inverter extracts a constant power from the upper DC link,
output power of PV array is variable depended on temperature
and irradiation (that may be more or less than the required
power). A battery bank is used to compensate this power
difference. One of capacitors of upper z-source network is
replaced by the battery bank. The upper shoot-trough signal is
calculated by (15) to track maximum power from PV array.
The reference capacitor voltage (VCU*) is equal to output
voltage of the battery bank (Vbatt). In this topology, the
reference DC link voltage (Vi*) is not given and determined
by:
*
Vi* = 2Vbatt VoU

The topology A (Fig. 4) was simulated using FC1 and FC2

(Table I). The reference DC link voltage (Vi*) was set to 450
V. A step function (which changes from 15 kW to 5 kW at t =
1S) was used as reference power signal. Fig. 9 shows power
waveforms in this simulation. The active power delivered to
the grid tracks the reference signals. The same powers are
extracted from both FC stacks. The injected current to the grid
is shown in Fig. 10. The total harmonic distortion (THD) of
the current is less than 3%. It can be seen in Fig. 11 that
voltage of both DC links are the same and equal to the
reference value (450 V). The capacitor voltage of z-source
networks are shown in Fig. 12. Due to reduction of output
power of FCs after t=1S, the output voltage of FCs are
increased and consequently capacitor voltages rise.
The topology B (Fig. 5) was simulated using PV and FC2
(Table I). The reference DC link voltage was set to 450 V. It is
assumed that maximum power of PV changes from 7.5 kW to
2.5 kW at t = 1S. As shown in Fig. 13, the maximum power of
PV has been extracted. The output power of FC is almost
equal to power of PV. Fig. 14 shows that THD of the current
injected to the grid is always less than 5%. The voltages of
both DC links are equal to the reference value (450 V) as
shown in Fig. 15. While maximum power of PV decreases, the
output voltage at maximum power point reduces. With
reduction of output voltage of PV, capacitor voltage of upper
z-source network decreases too, as shown in Fig. 16. On the
other hand while output power of FC decreases after t=1S, the
output power of FC and consequently capacitor voltages of
lower z-source network rise.

(17)

Similar to previous configurations, the lower shootthrough signal is determined by (14) to control the lower DC
link voltage.
D. FC-FC system based on SZSN-NPCI with CPC
This configuration is similar to first topology; however
DZSN-NPCI is replaced with a SZSN-NPCI. Although in this
topology a given power is delivered to the grid, the supplied
powers from FCs may be unequal (especially when FCs are
different) because dc input voltages of SZSN-NPCI should be
the same. Therefore a PI controller is used to determine power
ratio (R). On the other hand for desired DC link voltage, boost
factor is determined by:
*

(19)

The topology B (Fig. 6) was simulated using PV and FC2

(Table I). A battery bank with rated voltage of 350 V was
used. It is assumed that the average power of PV is 5 kW.
Thus the reference power was set to 10 kW. It is also assumed
that maximum power of PV changes from 7.5 kW to 2.5 kW at
t = 1S. As shown in Fig. 17, the maximum power of PV is
extracted. While inverter always extracts 5kW from both DC
links, the battery saves extra power of PV before t = 1S, and
supplies shortage of power after t = 1S. It can be seen in Fig.
17 that power of FC is always 5kW (half of the reference
power). As shown in Fig. 18 THD of the current injected to

B = Vi (VC Vi )
(18)
where VC is measured voltage of z-source capacitor. Shortthrough signal can be calculated by (9) and (10).
E. Hybrid PV-FC system based on SZSN-NPCI with MPC
In this configuration, PV array and FC stacks are
connected to the grid SZSN-NPCI. Maximum power of PV is
extracted and delivered to grid. For this purpose, A MPPT

107

the grid is always less than 2%. When maximum power of PV

decreases the output voltage of PV should be reduce for
MPPT. While the battery is used in upper z-source network
(constant voltage), the upper DC link voltage should increase
due to reduction of PV voltage as shown in Fig. 19. Since
voltages of both DC link should be the equal, the lower DC
link should also increase. On the other hand voltage of FC is
constant (due to its constant power). Therefore the capacitor
voltage of z-source network increases (Fig. 20).
The topology D (Fig. 5) was simulated using FC1 and FC2
(Table V). Vi* was set to 450 V. A step function (which
changes from 7.5 kW to 5 kW at t = 1S) was used as reference
power signal. As shown in Fig. 21, the active power delivered
to the grid tracks the reference power. However the powers of
FCs are not the same because of their different specifications.
The current injected to the grid is the almost sine waveform as
shown in Fig. 22. It can be seen in Fig. 23 that voltage of both
DC links are the same and equal to the reference value (450
V). The capacitor voltage of z-source networks and voltage of
FCs are shown in Fig. 24. It can be seen that voltage of FCs
are always the same. Due to reduction of output power of FCs
after t=1S, the output voltage of FCs are increased and
consequently capacitor voltages rise.

TABLE I.
Element
Grid

PV

FC1
FC2

Figure 10. Topology A (Fig.4): The current injected to the grid.

SIMULATION PARAMETERS
Parameter
Line Voltage
Frequency
Inductor (Li)
Rated Power
Rated Voltage
MPP Range
Rated Power
Output Voltage Range
Rated Power
Output Voltage Range

Value
380 V
50 Hz
5 mH
10 kW
280 V
150-300 V
10 kW
200-380 V
10 kW
150-300 V

Figure 11. Topology A (Fig.4): Voltages of DC links.

Figure 12. Topology A (Fig. 4): Capacitor voltages of z-source networks.

Figure 9. Topology A (Fig.4): The reference power, the active power

delivered to the grid, the output power of FCs.
Figure 13. Topology B (Fig. 5): The active power delivered to the grid, the
maximum and extracted power of PV, the output power of FC.

108

Figure 14. Topology B (Fig. 5): The current injected to the grid.

Figure 18. Topology C (Fig. 6): The current injected to the grid.

Figure 15. Topology B (Fig. 5): Voltages of DC links.

Figure 19. Topology C (Fig. 6): Voltages of DC links.

Figure 16. Topology B (Fig. 5): Capacitor voltages of z-source networks.

Figure 20. Topology C (Fig. 6): Capacitor voltages of z-source networks.

Figure 17. Topology C (Fig. 6): The active power delivered to the grid, the
maximum and extracted power of PV, the output power of FC and battery.

Figure 21. Topology D (Fig. 7): The reference power, the active power
delivered to the grid, the output power of FCs.

109

VI. REFERENCES
[1]

[2]
[3]

[4]

[5]

Figure 22. Topology D (Fig. 7): The current injected to the grid.

[6]

[7]

[8]

Figure 23. Topology D (Fig. 7): Voltages of DC links.

[9]

[10]

[11]

[12]

[13]
Figure 24. Topology D (Fig. 7): Capacitor voltages of z-source networks and
voltages of FCs.

V. CONCLUSION
In this paper a single stage converter was used for
connection of dual DG system to the grid using Z-Source NPC
inverters (DZSN-NPCI and SZSN-NPCI) which improves the
efficiency, power quality and cost of total system. Five
different topologies were proposed: A) dual fuel cell stacks
system based on DZSN-NPCI with constant power control, B)
hybrid photovoltaic - fuel cell system based on DZSN-NPCI
with maximum power control, C) hybrid photovoltaic - fuel
cell system based on DZSN-NPCI with maximum power
control, D) dual fuel cell stacks based on SZSN-NPCI with
constant power control and E) photovoltaic - fuel cell hybrid
system based on SZSN-NPCI with maximum power control.
The control system using maximum delivered power control
and constant power were designed and discussed. The
proposed topologies were realized and validated by
simulation.

[14]

[15]

110

C. J. Gajanayake, D. M. Vilathgamuwa, and Poh Chiang Loh,

Development of a Comprehensive Model and a Multiloop Controller
for Z-Source Inverter DG Systems, IEEE Transactions on Industrial
Electronics, vol. 54, no. 4, pp. 2352 2359, 2007.
F. Z. Peng, Z-Source Inverter, IEEE Transaction on Industry
Applications, vol. 39, no. 2, pp. 504-510, 2003.
Y. Huang, M. Shen and F. Z. Peng, A Z-Source Inverter for Residential
Photovoltaic Systems, IEEE Transactions on Power Electronics, vol.
21, no. 6, pp. 1776 1782, 2006.
S. M. Dehghan, M. Mohamadian and A. Yazdian, A New Variable
Speed Wind Energy Conversion System Using Permanent Magnet
Synchronous Generator and Z-Source Inverter, IEEE Transactions on
Energy Conversion, vol. 24, no. 2, 2009.
R. Teichmann and S. Bernet, A comparison of three-level converters
versus two-level converters for low-voltage drives, traction and utility
applications, IEEE Transaction on Industry Application., vol. 41, no. 3,
pp. 855865, 2005.
Mariusz Malinowski and Steffen Bernet, Simple Control Scheme of
Three-Level PWM Converter Connecting Wind Turbine with Grid,
Conference on Renewable Energies and Power, 2009.
Salvador Alepuz, Sergio Busquets-Monge, Josep Bordonau, Javier
Gago, David Gonzlez and Josep Balcells, Member, Interfacing
Renewable Energy Sources to the Utility Grid Using a Three-Level
Inverter, IEEE Transaction On Industrial Electronic, vol. 53, no. 5,
2006.
P. C. Loh, S. Y. Feng, F. Blaabjerg, and K. N. Soon, Pulse-Width
Modulated Z-Source Neutral-Point-Clamped Inverter, IEEE
Transactions on Industry Applications, vol. 43, No.5, pp.1054-1061,
2007.
Poh Chiang Loh, , Sok Wei Lim, Feng Gao, and Frede Blaabjerg,
Three-Level Z-Source Inverters Using a Single LC Impedance
Network, IEEE Transactions on Power Electronics, vol. 22, no. 2, pp.
706-711, 2007.
Debosmita Das, Reza Esmaili1, Longya Xu1 and Dave Nichols, An
Optimal Design of a Grid Connected Hybrid Wind/Photovoltaic/Fuel
Cell System for Distributed Energy Production, Industrial Electronics
Society (IECON), 2005.
Hui Li, Zhong Du, Kaiyu Wang1, Leon M. Tolbert and Danwei Liu, A
Hybrid Energy System Using Cascaded H-bridge Converter, Industry
Applications Conference (IAS), vol. 1, pp. 198 203, 2006.
S.M. Dehghan, M. Mohammadian, A. Yazdian, R. Gharekhani,
Analysis and Carrier-Based Modulation of Z-Source NPC Inverters,
Submitted to International Journal of Electronic.
M. Chinchilla, S. Arnaltes and J.C. Burgos, Control of PermanentMagnet Generators Applied to Variable-Speed Wind-Energy Systems
Connected to the Grid, IEEE Transaction on Energy Conversion, vol.
21, no. 1, pp. 130-135, 2006.
F. Z. Peng, M. Shen and K. Holland, Application of Z-Source Inverter
for Traction Drive of Fuel CellBattery Hybrid Electric Vehicles,
IEEE Transactions on Power Electronics, vol. 22, no. 3, pp. 1054-1061,
2007.
Wu Libo, Zhao Zhengming, and Liu Jianzheng, A Single-Stage ThreePhase Grid-Connected Photovoltaic System With Modified MPPT
Method and Reactive Power Compensation, IEEE Transaction on
Energy Conversion, vol. 22, no. 4, pp. 881-886, 2007