322

IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS, VOL. 62, NO. 1, JANUARY 2015

Modulation Techniques to Reduce Leakage

Current in Three-Phase Transformerless
H7 Photovoltaic Inverter
Tan Kheng Suan Freddy, Nasrudin A. Rahim, Senior Member, IEEE, Wooi-Ping Hew, Member, IEEE, and
Hang Seng Che, Member, IEEE

AbstractRecently, reduced common-mode voltage

(CMV) pulsewidth modulation (RCMV-PWM) methods have
been proposed to reduce the leakage current in three-phase
transformerless photovoltaic (PV) systems. However, most
of these studies only focus on leakage current elimination
and neglect the overall performance of the PV systems on
issues such as cost, voltage linearity, dc-link current ripples, and harmonic distortion. In this paper, a three-phase
transformerless inverter, adapted from the single-phase H5
topology, is investigated. Since the H5 topology has been
conventionally developed for a single-phase system, its
adaptation to the three-phase system requires the development of corresponding three-phase modulation techniques.
Hence, modulation techniques are proposed based on conventional PWM. The performances of the proposed PWM,
in terms of CMV, leakage current, voltage linearity, output current ripples, dc-link current ripples, and harmonic
distortion are studied and discussed via simulation and
experiment. It is proven that the proposed topology is able
reduce the leakage current without sacricing the overall
performance of the system.
Index TermsCommon-mode voltage (CMV), leakage
current, photovoltaic (PV) system, transformerless.

I. I NTRODUCTION

HE rapid increase in human population and the fast growth

of industries have shifted the attention of the research
community toward photovoltaic (PV) energy. PV energy is
free, green, and inexhaustible. Recently, PV power systems
have become widespread due to the government incentives,
Manuscript received November 24, 2013; revised March 13, 2014;
accepted May 12, 2014. Date of publication June 2, 2014; date of current
version December 19, 2014. This work was supported in part by the
Campus Network Smart Grid for Energy Security under Grant H-1600100-D000032 and by the Peruntukan Penyelidikan Pascasiswazah (PPP)
under Grant PV105-2012A.
T. K. S. Freddy and W.-P. Hew are with the Power Energy Dedicated
Advanced Center (UMPEDAC) and the Department of Electrical Engineering, Faculty of Engineering, University of Malaya, Kuala Lumpur
59990, Malaysia (e-mail: freddytan.86@gmail.com; wphew@um.
edu.my).
N. A. Rahim is with the Power Energy Dedicated Advanced Center
(UMPEDAC), University of Malaya, Kuala Lumpur 59990, Malaysia,
and also with King Abdulaziz University, Jeddah 21589, Saudi Arabia
(e-mail: nasrudin@um.edu.my).
H. S. Che is with the Power Energy Dedicated Advanced Center
(UMPEDAC), University of Malaya, Kuala Lumpur 59990, Malaysia
(e-mail: hsche@um.edu.my).
Color versions of one or more of the figures in this paper are available
online at http://ieeexplore.ieee.org.
Digital Object Identifier 10.1109/TIE.2014.2327585

Fig. 1.

H7 topology for three-phase transformerless PV systems.

reduction in PV arrays prices, and the advancement of power

electronics and semiconductor technology [1][6].
Generally, there are two types of grid-connected power systems, i.e., with a transformer and without a transformer. The
transformer used can be a high-frequency transformer on the
dc side or a low-frequency transformer on the ac side [7]
[10]. In addition to stepping up the input voltage, it plays an
important role in safety by providing galvanic isolation, which
eliminates the leakage current and avoids dc injection into
the grid. Nevertheless, the transformers are bulky, heavy, and
expensive. Hence, transformerless PV systems are introduced
to overcome these issues. They are smaller, lighter, lower in
cost, and highly efficient [9][12].
However, safety is the main concern for the transformerless
PV systems due to high leakage current. Without galvanic
isolation, a direct path can be formed for the leakage current
to flow from the PV to the grid. When the PV is grounded,
stray capacitance is created. The fluctuating potential charges
and discharges the stray capacitance, which generates high
leakage current [13][15], [26]. In addition to the safety issue,
the high leakage current will degrade the performance of the
PV system by increasing the grid current ripples, losses, and
electromagnetic interference.
Conventional pulsewidth modulation (PWM), either spacevector PWM (SVPWM) or discontinuous PWM (DPWM), are
not suitable for three-phase transformerless PV applications
due to high leakage current. In order to reduce the leakage current to meet the requirement of the standard, several
conversion structures and modulation techniques have been
proposed recently. In [5], the connection between the neutral

0278-0046 2014 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.
See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.

FREDDY et al.: MODULATION TECHNIQUES TO REDUCE LEAKAGE CURRENT IN H7 PV INVERTER

323

TABLE I
P ULSE PATTERNS FOR VARIOUS PWM M ETHODS

point of the grid to the middle point of the dc link ensures

constant voltage across the stray capacitance, which leads to
zero leakage current. However, this is not practical because
such connection creates the inductance in the neutral line. This
inductance generates high-frequency potential between the PV
array and the ground, which leads to leakage current higher
than the permissible level recommended in the standard [17].
A new conversion topology has been proposed in [19] with
extra two switches and 12 diodes. These excessive additional
components significantly increase the cost, the losses, and the
size of the entire PV system. Moreover, the leakage current is
still relatively high (280 mA) [19].
Reduced CMV PWM (RCMV-PWM) methods such as active
zero-state PWM (AZPWM) [20], near-state PWM (NSPWM)
[21], and remote-state PWM (RSPWM) [22], have been proposed recently. Without zero vectors, RCMV-PWM methods
are able to reduce the CMV to reduce the leakage current.
Nevertheless, the overall performances in terms of voltage
linearity, dc-link and output current ripples, line-to-line output
voltage pattern, losses, and total harmonic distortion (THD)
of the output current are degraded. Considering all criteria for
the transformerless PV systems, this paper investigates a threephase transformerless inverter, as shown in Fig. 1, by adapting
the popular single-phase H5 topology [28] for three-phase operation. As shown in Fig. 1, an additional switch is added to the
conventional three-phase inverter structure to provide galvanic
isolation. Given that a total of seven switches are utilized, this
topology is referred to as the H7 topology hereafter.
H5 topology was originally proposed in [28] to reduce the
leakage current in single-phase transformerless PV systems.
Although H7 is a simple extension from the H5 inverter, study
on this topology is yet to be reported and is thus attempted here
in this paper. Since the H5 topology has been conventionally
developed for a single-phase system, its adaptation to the threephase system (i.e., the H7 topology) requires the development
of a corresponding three-phase modulation technique. Hence,
modified DPWM (MDPWM) based on conventional DPWM
is proposed here. The performance of the H7 topology with
the proposed modulation techniques is compared with other
available RCMV-PWM methods [20][22] in terms of CMV,
leakage current, dc-link current ripples, output current ripples,
and the THD of the output current. Discussions are done
based on Matlab/Simulink simulations and further validated
via experimental works. It is proven that the H7 topology
with the proposed modulation techniques gives the best overall

Fig. 2. Voltage vector states for the RCMV-PWM methods with different
ways of partitioning the space-vector sectors. (a) Type A. (b) Type B.

performance and is suitable for transformerless PV applications

for 230-V (RMS) grid system.
This paper is organized as follows. Leakage current reduction
methods via RCMV-PWM are first discussed in Section II. The
H7 conversion structure with the proposed modulation techniques and the operation principles are presented in Section III.
Simulation and experimental results are shown in Sections IV
and V, respectively, to validate the performance of the various
topologies. Finally, conclusion is made in Section VI to summarize the results and findings.
II. C OMMON -M ODE B EHAVIOR AND L EAKAGE
C URRENT R EDUCTION M ETHODS
Leakage current minimization is one of the most important
considerations in transformerless PV inverters. In the past,
various transformerless PV inverter topologies have been introduced, with leakage current minimized by the means of
modulation techniques and conversion structure, which will be
discussed in this section.
A. Modulation Techniques
Several modulation techniques, which are termed as RCMVPWM, have been proposed recently as listed in Table I. The
modulation techniques can be classified into two types: A [20]
and B [21], [22], based on the way the space-vector sectors are
partitioned, as shown in Fig. 2.
In AZPWM, the active vectors are complemented with two
opposing active vectors to create zero vectors with equal time.
In RSPWM, a group of odd active vectors (V1 , V3 , and V5 ) or
even active vectors (V2 , V4 , and V6 ) is used to generate desired
output voltage with constant CMV. NSPWM employs only
three adjacent active vectors. Without using any zero vectors,

324

IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS, VOL. 62, NO. 1, JANUARY 2015

RCMV-PWM reduces the CMV to reduce the leakage current.

The CMV (VCM )of a three-phase PV inverter is expressed as
[5], [18], [26]
VCM =

VAN + VBN + VCN

.
3

TABLE II
V ECTORS C OMBINATION AND C ORRESPONDING CMV FOR MDPWM

(1)

However, issues such as voltage linearity, harmonic distortion, and output voltage arise due to the elimination of
the zero vectors. One common problem for RCMV-PWM
methods is its bipolar line-to-line output voltage. In every
switching transition, the voltage changes across the inductors
from +VDC to VDC . This doubles the voltage stress across
the inductors to twice of the input voltage. Moreover, such
output voltage pattern generates overvoltage transients [21],
[23], [24], large current ripples across the inductors (eventually
requires larger filters), and high switching losses [1], which
reduce the overall efficiency of the entire system. Moreover,
NSPWM and RSPWM are only applicable in limited modulation range. NSPWM operates linearly only for 0.61 < m <
0.907, whereas RSPWM is only applicable for m < 0.52 (m is
modulation index). Out of this range, leakage current generated
is higher than that of conventional SVPWM. For RSPWM, very
high input voltage is required to inject the real power to the grid.
As a result, high operational losses and high THD occur, which
require the use of large filters [22][25].
B. Conversion Structure
When the transformer is removed, the galvanic connection
between the PV and the grid allows the leakage current to
flow. Thus, many studies have been done in [28][36] to
design conversion structure, which provides galvanic isolation
to reduce the leakage current. Either dc decoupling or ac
decoupling methods are utilized to disconnect the PV arrays
from the grid during the freewheeling period. The modulation
techniques are also carefully designed with the corresponding
conversion structure in order to generate the desired output
voltage and to reduce the leakage current. Such topologies
yield satisfactory overall performance. Nevertheless, most of
the galvanic isolation topologies are found in single-phase PV
inverters. For three-phase PV inverters, modulation techniques
are much more complicated, and galvanic isolation methods are
therefore difficult to be implemented.
III. P ROPOSED TOPOLOGY AND THE
O PERATION P RINCIPLES
Based on the given analysis, a simple modified H7 inverter
with a proposed PWM method (known as MDPWM) is investigated to reduce the leakage current without compromising the
overall performance of the PV systems.
A. Proposed MDPWM
Conventional DWPM yields outstanding performances in
terms of voltage linearity, switching losses, and THD of the
output current [37]. However, such optimum performances of
the DPWM are degraded in transformerless PV applications.
DPWM generates high leakage current, which causes high

Fig. 3. Switching pattern, corresponding line-to-line output voltages

and CMV for MDPWM in (a) A1 B1 and (b) A1 B2.

current ripples. As a result, the THD of the output current

also increases. In order to maintain the unique characteristic of
DPWM, a MDPWM based on conventional DPWM is proposed
here to reduce the leakage current.
Unlike RCMV-PWM, the proposed MDPWM uses both
active vectors and zero vectors to program the output voltage.
However, zero vectors (V0 and V7 ) produce high CMV (0 and
VDC ). Instead of two zero vectors, proposed MDPWM uses
only one zero vector (V7 ). This ensures that the CMV takes only
values of VDC /3, 2VDC /3, or VDC , as indicated in Table II.
In every odd regions (B1, B3, and B5), the pulse patterns of
MDPWM remains the same as in the conventional DPWM. The
difference appears only in even regions where V0 is replaced by
V7 in MDPWM. For example, as shown in Fig. 3(a), MDPWM
uses V1 , V2 , and V7 in region A1 B1. In region A1 B2, V1 ,
V2 , and V7 , (instead of V1 , V2 , and V0 ) are used as shown in
Fig. 3(b). Obviously, V7 has replaced V0 . The same approach
is applied to the rest of the regions, and the pulse patterns
are listed in Table I. Thus, the CMV is reduced, varying from
VDC /3 to VDC in odd sectors and VDC /3 to 2VDC /3 in even
sectors, as shown in Fig. 3.
MDPWM is proposed to reduce the CMV by reserving
the use of zero vectors. Zero vectors are important as they
determine the output voltage patterns. The elimination of the
zero vectors in RCMV-PWM reduces the CMV at the cost
of generating bipolar output voltage, which gives overvoltage

FREDDY et al.: MODULATION TECHNIQUES TO REDUCE LEAKAGE CURRENT IN H7 PV INVERTER

325

equivalent circuit of the H7 inverter when both active vectors

and zero vector are applied in region A1 B1 where Sa and
Sb are the equivalent single-pole double-throw switches for the
half-bridge of each phase. Phase A of the H7 inverter is clamped
to the positive dc-link voltage.
As shown in Fig. 4(a), when active vectors (V1 , V2 ) are
applied, S7 is turned on to generate the desired output voltage,
and the corresponding CMV becomes VDC /3 and 2VDC /3, respectively. On the other hand, when zero vector V7 is applied, all
the upper switches S1 , S3 , and S5 are turned on and connected
to the positive (P) of the dc-link voltage. At this moment, S7
is turned off to disconnect the PV from the grid. Therefore,
leakage current finds no path to flow, as shown in Fig. 4(b).
VAN and VBN decrease, and VCN increases until their values
reach the common point, i.e., VDC /3. The CMV becomes
Fig. 4. Simplified equivalent circuit of the H7 inverter during (a) active
vectors and (b) zero vectors in region A1 B1.

transients and increased current ripples. On the other hand,

MDPWM does not suffer from such problem as it generates
unipolar line-to-line output voltage, as shown in Fig. 3. The
main drawback of MDPWM is the simultaneous switching of
two inverter legs, which leads to increased switching losses.
However, applying MDPWM is the only condition to reduce
the leakage current while maintaining the unique characteristic
of DPWM.
It is worth noting that, ideally, the CMV should be kept at
a constant value to eliminate leakage current, as in RSPWM.
However, RSPWM suffers from the problem of a limited modulation region, which increases its dc-link voltage requirement
and the corresponding losses. Here, the MDPWM does not
maintain the CMV at a constant value as in RSPWM. Instead,
it merely reduces the CMV, in a similar manner as the NSPWM
and AZPWM. This reduction in CMV, while being able to
reduce the leakage current, may not be sufficient to lower
the leakage current enough to satisfy the current standard of
300 mArms [17]. Hence, in the H7 topology, additional switch
S7 is utilized to provide galvanic isolation to the leakage current
path to further reduce the leakage current. This is explained in
the succeeding section.
B. Operation of the H7 Conversion Structure
When the transformer is removed from the inverter, it forms
a direct path for the leakage current to flow between the PV
and the grid. This happens during the freewheeling period
when all the upper switches turns on or turns off at the same
time. In other words, the freewheeling period is the moment
when the zero vector is applied. Hence, an additional switch
S7 is added to the three-phase full-bridge inverter to provide
galvanic isolation between the PV array and the grid during
the freewheeling period. The circuit configuration of the H7
inverter is shown in Fig. 1.
MDPWM clamps each phase leg of the inverter to the positive of the dc-link voltage, whereas the other two phases are
modulated at switching frequency. Fig. 4 shows the simplified

VAN + VBN + VCN

3


VDC
VDC
1 VDC
VDC
+
+
.
=
=
3
3
3
3
3

VCM =

(2)

The switching pattern and the corresponding CMV for both

region A1 B1 and A1 B2 is shown in Fig. 3. In region A1
B2, S7 is turned on all the time to generate the desired output
voltage. Voltage vectors V1 , V2 , and V7 are used, and the CMV
is VDC /3, 2VDC /3, and VDC , respectively.
The same principles are applied to all the other regions.
Obviously, CMV reduces and varies between VDC /3 and VDC .
In addition to the CMV reduction, galvanic isolation further
reduces the leakage current. As a result, the H7 topology with
the proposed MDPWM combines the optimum behavior for
both DPWM and RCMV-PWM methods.
It is worth noting that voltage VAN , VBN , and VCN are
impossible to reach the common point VDC /3 during the disconnection of the S7 . In practical applications, they are floating
and oscillating due to the switches junction capacitance and
resonant circuit effects. This effect will be seen clearer in the
simulation results shown in Fig. 8. In fact, this effect can be
avoided using the CMV clamping branch, which have been
proposed in [10], [29], and [30]. However, the CMV clamping
for the three-phase system requires additional hardware configuration, which may add burden to the cost, efficiency, and
complexity of the design for both hardware and software. The
CMV clamping method is not within the scope of this paper and
will be discussed in future work.

IV. S CALAR I MPLEMENTATION OF MDPWM

In this section, the implementation of the proposed MDPWM
is discussed. The overall block diagram for the proposed PWM
is shown in Fig. 5. It is simple and straightforward as in
conventional DPWM. It can be easily implemented using the
scalar approach as follows:
Va = Va + V0 = m sin(wt) + V0
Vb = Vb + V0 = m sin(wt 2/3) + V0
Vc = Vc + V0 = m sin(wt + 2/3) + V0

(3)
(4)
(5)

326

Fig. 5.

IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS, VOL. 62, NO. 1, JANUARY 2015

Proposed PWM modulator.

where Va , Vb , and Vc are the original sinusoidal reference

signals, V0 is the zero-sequence signal, and m is the modulation
index. V0 is calculated via the magnitude test as follows:
V0 = [sign(Vmax )] (VDC /2) Vmax

(6)

where Vmax is the original sinusoidal reference (Va , Vb , or Vc )

with the maximum magnitude.
After the injection of V0 , the resultant modulation signals
(Va , Vb , and Vc ) are then compared with a triangular carrier
wave to generate the logic signals Sa , Sb , and Sc . These logic
signals (Sa , Sb , and Sc ) are the pulse patterns of conventional
DPWM. In order to generate MDPWM with only one zero
vector (V7 ), simple logic operations are utilized to generate
the desired gating signals (S1 , S3 , S5 , and S7 ). In other words,
the implementation of the proposed MDPWM is in the same
manner as in conventional DPWM. Only additional simple
logic operations are required.
V. P ERFORMANCE A NALYSIS OF MDPWM
A. Simplicity of Design and Cost
Unlike RCMV-PWM methods, the implementation of the
proposed MDPWM is simple and straightforward. This is
because the AZPWM, NSPWM, and RSPWM methods require the simultaneous use of two different triangular carriers
(updown and downup triangular carriers) [23][25], which
needs an advance digital signal processor (DSP), such as Texas
Instruments TMS320F28335 with an enhanced PWM module,
for implementation. Hence, it is difficult to be realized on
low-cost microcontrollers with limited carrier configuration
flexibility. On the other hand, the proposed MDPWM method
requires only one carrier (updown or downup) for all three of
its phases. Furthermore, the implementation of the MDPWM is
as simple as the conventional DPWM with only additional logic
operations that can be digitally implemented into the DSP.

Fig. 6. Line-to-line voltage (top) and grid current (bottom) for (a)
SVPWM, (b) DPWM, (c) AZPWM, (d) NSPWM, and (e) H7 with the
proposed MDPWM.

transients and PWM current ripples are reduced; thus, smaller

filter inductors can be used.

B. Line-to-Line Output Voltage Patterns

C. Voltage Linearity

The elimination of the zero vectors reduces the CMV in

RCMV-PWM, but it generates undesirable bipolar line-to-line
output voltage. Zero vectors are important as it determines
the line-to-line output voltage pattern. MDPWM inherits the
unique characteristic of unipolar line-to-line output voltage
by reserving the utilization of zero vectors. In every PWM
cycle, the polarity of the voltage remains the same. Overvoltage

Each PWM modulation method has a specific linear modulation region. Outside this region, the overall performance in
terms of the THD of the output current, current ripples, voltage
magnitude, and common-mode behavior are degraded. Therefore, a wide linear modulation range is preferable. MDPWM
maintains the superior voltage linearity characteristic of conventional DPWM. By injecting the zero-sequence signal,

FREDDY et al.: MODULATION TECHNIQUES TO REDUCE LEAKAGE CURRENT IN H7 PV INVERTER

327

Fig. 8. Zoom-in waveforms of (top) CMV and (bottom) leakage current

for H7 with the proposed MDPWM, showing CMV oscillation during the
freewheeling period.

Fig. 7. (Top) CMV and (bottom) leakage current for (a) SVPWM,
(b) DPWM, (c) AZPWM, (d) NSPWM, and (e) H7 with the proposed
MDPWM.

MDPWM is able to extend the utilization of the dc-link voltage

by 15%, which enable it to operate linearly for 0 < m < 0.907.
VI. S IMULATION R ESULTS
In order to verify the operation and the overall performance
of the PV system, simulations are carried out, as shown in
Fig. 1, using Matlab/Simulink. All the simulations are done
based on the same parameters. The PV array is simulated with
a dc voltage source of 600 V. The stray capacitance (CPV )

is modeled with two capacitors of 220 nF, each connected

between the PV terminal and the ground. The dc-link capacitor
(CDC ) is 2mF. The ground resistance (RG ) is 11 . The
filter is made up of three inductors LF , each having a value
of 5 mH. The load resistance is 100 , and the modulation
index m = 0.9. For fair comparison, the switching frequency
is selected such that all the modulation techniques have equal
average switch commutation, i.e., SVPWM, AZPWM, and the
proposed MDPWM are modulated at 10 kHz, whereas DPWM
and NSPWM at 15 kHz. RSPWM is not taken into account due
to its impracticality in grid-connected applications.
Fig. 6 shows the line-to-line output voltage waveforms and
the output currents for various modulation techniques. As
shown in Fig. 6(a) and (b), conventional SVPWM and DPWM
generate unipolar line-to-line output voltage. Both AZPWM
and NSPWM share the common bipolar output voltage waveform. As observed in Fig. 6(c) and (d), the output voltage varies
from +VDC to VDC . This doubles the voltage stress across
the inductors by twice of the input voltage. As a result, losses
and current ripples double, which requires the use of larger
filter inductors. The proposed MDPWM shares the similar
characteristic of conventional DPWM. It generates unipolar
output voltage, as shown in Fig. 6(e).
Fig. 7 indicates the common-mode behavior for various
topologies. Due to the utilization of zero vectors, both SVPWM
and DPWM have the poorest common-mode behavior. As
shown in Fig. 7(a) and (b), the CMVs vary from 0 to VDC ; thus,
high leakage currents are generated. This also explains why
the output current ripples are higher than the other topologies,
as observed in Fig. 6(a) and (b). Without any zero vectors,
AZPWM and NSPWM are able to reduce the leakage current by
reducing the CMV, varying from VDC /3 to 2VDC /3, as shown
in Fig. 7(c) and (d), respectively. Thus, the leakage currents are
reduced. Similarly, the proposed MDPWM is able to reduce the
leakage current due to the reduced CMV and galvanic isolation.
The CMV changes depending on S7 . When S7 is turned on, the
CMV varies from VDC /3 to 2VDC /3 and from VDC /3 to VDC ,
otherwise.
Fig. 8 shows the microscopic view of the CMV waveform
for the proposed MDPWM. Instead of VDC /3 and 2VDC /3,
small oscillation with magnitude up to 100 V is observed in the
CMV waveform. This oscillation occurs when S7 is turned off
to provide galvanic isolation during the freewheeling period.
Practically, VAN , VBN , and VCN cannot reach the common
voltage of VDC /3 due to the switches junction capacitance and
resonant circuit effects, as explained earlier.

328

IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS, VOL. 62, NO. 1, JANUARY 2015

Fig. 9. Line-to-line voltage (500 V/div), output current (2 A/div), and leakage current (500 mA/div) for (a) SVPWM, (b) DPWM, (c) AZPWM,
(d) NSPWM, and (e) H7 topology with the proposed MDPWM.

However, it is worth mentioning that the H7 topology with

the proposed MDPWM still has the best overall performance,
which combines the advantages of the unipolar output voltage
pattern, reduced switching losses of DPWM, and the low CMV
and low leakage currents of the RCMV-PWM methods.
VII. E XPERIMENTAL R ESULTS
The experimental setup has the same parameters as the simulations. International Rectifier IRGP30B120KD-EP 1200-V
60-A fast IGBTs are used for switches. All the control algorithms are implemented in DSP TMS320F28335, with dead
time = 2.5 s.
Fig. 9 shows the experimental results of line-to-line output voltage, output current, and leakage current for various
topologies. Both conventional SVPWM and DPWM share
similar unipolar output voltage and sinusoidal output current
waveforms. Nevertheless, high current ripples are observed, as
shown in Fig. 9(a) and (b), respectively, due to high leakage
currents. The magnitude of the leakage current for SVPWM
and DPWM are 610 and 414 mA (RMS), respectively, which
have exceeded the requirement of VDE 0126-1-1 standard.
Moreover, common-mode noise is observed. Therefore, they
are not suitable for transformerless PV applications.
Without any zero vectors, the leakage current are reduced
significantly to 240 and 229 mA (RMS) for AZPWM [see
Fig. 9(c)] and NSPWM [see Fig. 9(d)], respectively, due to
reduced CMV. This explains why the output current ripples
are smaller compared with the conventional modulation techniques. However, the replacement of the zero vectors causes the
bipolar output line-to-line voltage issues. As shown in Fig. 9(c)
and (d), in certain PWM cycles, the output voltage varies from

Fig. 10. Experiment result of dc-link voltage ripples for various modulation techniques.

+VDC to VDC . This doubles the voltage stress across the filter
inductors and hence reduces the overall efficiency. Issues such
as overvoltage transients may arise as well.
DPWM and MDPWM have the same modulation signal.
Hence, MDPWM demonstrates the same attributes as DPWM.
With the use of zero vectors, it generates the favorable unipolar
output voltage waveform, as shown in Fig. 9(e). On the other
hand, leakage current is reduced to 212 mA (RMS) due to
the reduced CMV and galvanic isolation of H7 topology. Although both AZPWM and NSPWM meet the requirement of
VDE 0126-1-1 standard, the proposed method gives the lowest
leakage current.

FREDDY et al.: MODULATION TECHNIQUES TO REDUCE LEAKAGE CURRENT IN H7 PV INVERTER

329

TABLE III
P ERFORMANCE C OMPARISONS FOR VARIOUS PWM T ECHNIQUES

The voltage ripple performance is another critical factor as

it determines the dc-link capacitor size, lifetime, and, hence,
cost. Therefore, lower ripples are preferred. The dc-link voltage ripple waveforms for various PWM methods are captured as shown in Fig. 10. The peak-to-peak voltage ripples
for SVPWM, DPWM, AZPWM, NSPWM, and H7 with the
proposed MDPWM are 3.88, 1.66, 1.68, 1.62, and 1.60 V,
respectively. Although conventional SVPWM and DPWM are
more superior in terms of dc-link current ripples, as reported in
[22], they are not valid for transformerless PV applications due
to the very high leakage current. As expected, SVPWM has the
highest dc-link voltage ripples. The proposed H7 topology with
MDPWM inherits the superior characteristics of DPWM. This
explains why the proposed topology yields the lowest dc-link
voltage ripples among all methods.
The THD of the output currents for different topologies
are calculated via the fast Fourier transform analysis of the
Matlab/Simulink and are summarized in Table III. Although
conventional SVPWM and DPWM are expected to have the
best THD performance, this is not valid in transformerless
PV applications. The high leakage currents of both SVPWM
and DPWM increase the current ripples significantly, which
degrade the THDs that read 19.87% and 10.87%, respectively.
Among the recently proposed topologies, AZPWM presents
relatively higher THD because of bipolar output voltage that
changes between VDC and VDC in certain PWM cycle. Although NSPWM has bipolar output voltage, its reduced switch
count pulse pattern (similar to DPWM), which allows NSPWM
to switch at higher frequency, has improved the THD performance. As a result, NSPWM has the best THD performance
(5.27%) among the discussed topologies. The proposed H7
topology with MDPWM yields slightly higher THD than that
with NSPWM due to the simultaneous switching of the proposed MDPWM. However, the THD performance is still the
best after NSPWM.
The experimental performance comparisons for all the PWM
techniques are summarized in Table III. Although DPWM is
well known for its superior performance in terms of switching
losses, current ripples, and voltage linearity, such optimum
performance is distorted in transformerless PV applications
due to high leakage current. It is experimentally proven that
the proposed H7 topology combines the superior performance
of both DPWM (unipolar line-to-line output voltage, voltage
linearity, low output current ripples, low dc-link current ripples,
and low THD) and RCMV-PWM (reduced CMV with low
leakage current). This indicates that the proposed topology has

the potential to be applied in three-phase transformerless gridconnected PV applications.

VIII. C ONCLUSION
In this paper, a three-phase transformerless PV inverter,
together with its PWM technique, have been investigated for
leakage current reduction. A MDPWM based on conventional
DPWM has also been proposed.
Many research works have shown that DPWM has superior
performance in terms of dc-link voltage ripples, output current ripples, and THD. However, these merits are invalid in
transformerless PV applications due to high leakage current.
In addition to safety issues, the leakage current has deteriorated
the overall performances of the PV systems. The present solution that utilizes RCMV-PWM solves the leakage current issues
by reducing the CMV through the elimination of zero voltage vector switching. However, these RCMV-PWM methods
only focus on leakage current elimination while neglecting the
overall performance of the PV systems. Issues such as voltage
linearity, output current ripples, dc-link current ripples, and
harmonic distortion have been compromised.
The H7 topology with the proposed MDPWM combines
the advantages of both DPWM and RCMV-PWM methods.
By reducing the CMV and by breaking the leakage current
conduction path during the freewheeling period, the leakage
current can be reduced to meet the requirements of the standards without compromising the overall performances of the
PV systems. The performance of H7 with the MDPWM is
compared with other available RCMV-PWM methods in terms
of CMV, leakage current, output current ripples, dc-link current
ripples, and THD. It is experimentally proven that the H7
topology gives the best overall performance and is suitable for
transformerless PV applications.
R EFERENCES
[1] F. T. K. Suan, N. A. Rahim, and H. W. Ping, Modeling, analysis and
control of various types of transformerless grid connected PV inverters,
in Proc. IEEE CET, Jun. 2011, pp. 5156.
[2] E. S. Sreeraj, K. Chatterjee, and S. Bandyopadhyay, One-cyclecontrolled single-stage single-phase voltage-sensorless grid-connected pv
system, IEEE Trans. Ind. Electron., vol. 60, no. 3, pp. 12161224,
Mar. 2013.
[3] N. A. Rahim, K. Chaniago, and J. Selvaraj, Single-phase sevenlevel grid-connected inverter for photovoltaic system, IEEE Trans. Ind.
Electron., vol. 58, no. 6, pp. 24352443, Jun. 2011.
[4] M. Calais, J. Myrzik, T. Spooner, and V. G. Agelidis, Inverters for singlephase grid connected photovoltaic systemAn overview, in Proc. IEEE
PESC, 2000, pp. 19952000.

330

IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS, VOL. 62, NO. 1, JANUARY 2015

[5] T. Kerekes, R. Teodorescu, M. Liserre, C. Klumpner, and M. Sumner,

Evaluation of three-phase transformerless photovoltaic inverter topologies, IEEE Trans. Power Electron., vol. 24, no. 9, pp. 22022211,
Sep. 2009.
[6] E. C. dos Santos, Jr., A. M. Farias, M. C. Cavalcanti, and F. Bradaschia,
Integrated three-phase transformerless PV inverter, in Proc. IEEE ISIE,
May 2012, pp. 17801784.
[7] T. Kerekes, R. Teoderescu, and U. Borup, Transformerless photovoltaic
inverters connected to the grid, in Proc. IEEE APEC, Feb./Mar. 2007,
pp. 17331737.
[8] F. Blaabjerg, R. Teodorescu, M. Liserre, and A. V. Timbus, Overview
of control and grid synchronization for distributed power generation
systems, IEEE Trans. Ind. Electron., vol. 53, no. 5, pp. 13981409,
Oct. 2006.
[9] E. Guba, P. Sanchis, A. Ursa, J. Lpez, and L. Marroyo, Ground currents in single-phase transformerless photovoltaic systems, Prog. Photovolt., Res. Appl., vol. 15, no. 7, pp. 629650, Nov. 2007.
[10] H. Xiao and S. Xie, An optimized transformerless photovoltaic gridconnected inverter, IEEE Trans. Ind. Electron., vol. 58, no. 5, pp. 1887
1895, May 2011.
[11] D. Meneses, F. Blaabjerg, O. Garca, and J. A. Cobos, Review and
comparison of step-up transformerless topologies for PV ac-module application, IEEE Trans. Power Electron., vol. 28, no. 6, pp. 26492663,
Jun. 2013.
[12] Y. Bae and R. Y. Kim, Suppression of common-mode voltage using
a multicentral photovoltaic inverter topology with synchronized PWM,
IEEE Trans. Ind. Electron., vol. 61, no. 9, pp. 47224733, Sep. 2014.
[13] M. C. Cavalcanti, A. M. Farias, K. C. de Oliveira, F. A. S. Neves, and
J. L. Afonso, Eliminating leakage currents in neutral point clamped
inverters for photovoltaic systems, IEEE Trans. Ind. Electron., vol. 59,
no. 1, pp. 435443, Jan. 2012.
[14] O. Lopez et al., Eliminating ground current in a transformerless photovoltaic application, IEEE Trans. Energy Convers., vol. 25, no. 1, pp. 140
147, Mar. 2010.
[15] Y. Wang and R. Li, Novel high-efficiency three-level stacked-neutralpoint-clamped grid-tied inverter, IEEE Trans. Ind. Electron., vol. 60,
no. 9, pp. 37663774, Sep. 2013.
[16] R. Gonzalez, E. Gubia, J. Lopez, and L. Marroyo, Transformerless
single-phase multilevel photovoltaic inverter, IEEE Trans. Ind. Electron.,
vol. 55, no. 7, pp. 26942702, Jul. 2008.
[17] Automatic Disconnection Device between a Generator and the Public
Low-Voltage Grid, DIN Electrotech. Std. DIN VDE 0126-1-1, 2005.
[18] M. C. Cavalcanti et al., Modulation techniques to eliminate leakage currents in transformerless three phase photovoltaic systems, IEEE Trans.
Ind. Electron., vol. 57, no. 4, pp. 13601368, Apr. 2010.
[19] G. Vazquez et al., A photovoltaic three-phase topology to reduce
common-mode voltage, in Proc. IEEE ISIE, Jul. 2010, pp. 28852890.
[20] G. Oriti, A. L. Julian, and T. A. Lipo, A new space vector modulation
strategy for common mode reduction, in Proc. IEEE PESC, Jun. 1997,
pp. 15411546.
[21] E. n and A. M. Hava, A near-state PWM method with reduced switching losses and reduced common-mode voltage for three-phase voltage
source inverters, IEEE Trans. Ind. Appl., vol. 45, no. 2, pp. 782793,
Mar./Apr. 2009.
[22] M. Cacciato, A. Consoli, G. Scarccella, and A. Testa, Reduction of
common-mode currents in PWM inverter motor drives, IEEE Trans. Ind.
Appl., vol. 35, no. 2, pp. 469476, Mar./Apr. 1999.
[23] A. M. Hava and E. n, Performance analysis of reduced common-mode
voltage PWM methods and comparison with standard PWM methods
for three-phase voltage-source inverters, IEEE Trans. Power Electron.,
vol. 24, no. 1, pp. 241252, Jan. 2009.
[24] A. M. Hava and E. n, A high-performance PWM algorithm for
common-mode voltage reduction in three-phase voltage source inverters, IEEE Trans. Power Electron., vol. 26, no. 7, pp. 19982008,
Jul. 2011.
[25] C. C. Hou, C. C. Shih, P. T. Cheng, and A. M. Hava, Common-mode
voltage reduction pulse-width modulation techniques for three-phase grid
connected converters, IEEE Trans. Power Electron., vol. 28, no. 4,
pp. 19711979, Apr. 2013.
[26] F. Bradaschia et al., Modulation for three-phase transformerless z-source
inverter to reduce leakage currents in photovoltaic systems, IEEE Trans.
Ind. Electron., vol. 58, no. 12, pp. 53855395, Dec. 2011.
[27] J. Selvaraj and N. A. Rahim, Multilevel inverter for grid-connected
PV system employing digital PI controller, IEEE Trans. Ind. Electron.,
vol. 56, no. 1, pp. 149158, Jan. 2009.
[28] M. Victor, F. Greizer, S. Bremicker, and U. Hbler, Method of converting
a direct current voltage from a source of direct current voltage, more

[29]
[30]
[31]
[32]
[33]
[34]

[35]

[36]
[37]

specifically from a photovoltaic source of direct current voltage, into a

alternating current voltage, U.S. Patent 7 411 802, Aug. 12, 2008.
T. Kerekes, R. Teodorescu, P. Rodrguez, G. Vzquez, and E. Aldabas, A
new high-efficiency single-phase transformerless PV inverter topology,
IEEE Trans. Ind. Electron., vol. 58, no. 1, pp. 184191, Jan. 2011.
T. K. S. Freddy, N. A. Rahim, W. P. Hew, and H. S. Che, Comparison and
analysis of single-phase transformerless grid-connected PV inverters,
IEEE Trans. Power Electron., vol. 29, no. 10, pp. 53585369, Oct. 2014.
H. F. Xiao, X. P. Liu, and K. Lan, Zero-voltage-transition full bridge
topologies for transformerless photovoltaic grid-connected inverter,
IEEE Trans. Ind. Electron., vol. 61, no. 10, pp. 53935401, Oct. 2014.
L. Zhang, K. Sun, Y. Xing, and M. Xu, H6 transformerless full-bridge PV
grid-tied inverters, IEEE Trans. Power Electron., vol. 29, no. 3, pp. 1229
1238, Mar. 2014.
B. Ji, J. Wang, and J. Zhao, High-efficiency single-phase transformerless PV H6 inverter with hybrid modulation method, IEEE Trans. Ind.
Electron., vol. 60, no. 5, pp. 21042115, May 2013.
B. Yang, W. Li, Y. Gu, W. Cui, and X. He, Improved transformerless inverter with common-mode leakage current elimination for a photovoltaic
grid-connected power system, IEEE Trans. Power Electron., vol. 27,
no. 2, pp. 752762, Feb. 2012.
W. Yu, J.-S. Lai, H. Quian, and C. Hutchens, High-efficiency
MOSFET inverter with H6-type configuration for photovoltaic noniolated
AC-module applications, IEEE Trans. Power Electron., vol. 26, no. 4,
pp. 12531260, Apr. 2011.
S. Heribert, S. Christoph, and K. Jurgen, Inverter for transforming a DC
voltage into an AC current or an AC voltage, European Patent 1 369 985
(A2), May 13, 2003.
A. M. Hava and N. O. Cetin, A generalized scalar PWM approach with
easy implementation features for three-phase, three-wire voltage-source
inverters, IEEE Trans. Power Electron., vol. 26, no. 5, pp. 13851395,
May 2011.

Tan Kheng Suan Freddy received the B.Eng.

degree in electrical engineering from Multimedia University, Melaka, Malaysia, in 2010.
He is currently working toward the Ph.D. degree in the Power Energy Dedicated Advanced
Center (UMPEDAC), University of Malaya,
Kuala Lumpur, Malaysia.
He is also currently with the Faculty of Engineering, University of Malaya. His research
interests include transformerless photovoltaic
inverters, power electronics, and renewable
energy.

Nasrudin A. Rahim (M89SM08) received

the B.Sc. (Hons.) and M.Sc. degrees from the
University of Strathclyde, Glasgow, U.K., and
the Ph.D. degree from Heriot-Watt University,
Edinburgh, U.K.
He is currently a Professor and the Director
of the Power Energy Dedicated Advanced Center (UMPEDAC), University of Malaya, Kuala
Lumpur, Malaysia. He is also an Adjunct Professor with King Abdulaziz University, Jeddah,
Saudi Arabia. His research interests include
power electronics, real-time control systems, and electrical drives.
Dr. N. A. Rahim is a Fellow of the Institution of Engineering and
Technology, U.K., and the Academy of Sciences Malaysia.

FREDDY et al.: MODULATION TECHNIQUES TO REDUCE LEAKAGE CURRENT IN H7 PV INVERTER

Wooi-Ping Hew (M06) received the B.Eng.

and M.S. degrees in electrical engineering from
the University of Technology Malaysia, Johor,
Malaysia, and the Ph.D. degree from the University of Malaya, Kuala Lumpur, Malaysia.
He is currently a Professor with the Faculty of Engineering, University of Malaya, Kuala
Lumpur. He is also with the Power Energy Dedicated Advanced Center (UMPEDAC), University
of Malaya. His research interests include electrical drives and electrical machine design.
Dr. Hew is a member of the Institution of Engineering and Technology,
U.K., and a Chartered Engineer in the U.K.

331

Hang Seng Che (M14) received the B.Eng. degree in electrical engineering from the University
of Malaya, Kuala Lumpur, Malaysia, in 2009 and
the Ph.D. degree jointly from the University of
Malaya and Liverpool John Moores University,
Liverpool, U.K, in 2013.
He is currently a Senior Lecturer with the
Power Energy Dedicated Advanced Center
(UMPEDAC), University of Malaya. His research
interests include multiphase machines drives,
fault-tolerant control, and renewable energy.
Dr. Che was the recipient of the 2009 Kuok Foundation Postgraduate
Scholarship Award for his Ph.D. studies.