Escolar Documentos
Profissional Documentos
Cultura Documentos
AbstractThis paper underpins the principles for In grid-connected PV systems, an inverter is used to con-
generating reactive power in single-phase transformerless vert the direct current (dc) from the PV array into alternating
photovoltaic (PV) inverters. Two mainstream and widely current (ac) to supply the electricity to the utility grid. Gener-
adopted PV inverters are explored, i.e., H5 and HERIC.
With conventional modulation techniques, reactive power ally, there are two types of grid-connected PV inverters, that is,
cannot be realized in H5 and HERIC due to the absence with and without a 50/60 Hz transformer. Since a transformer is
of freewheeling path in negative power region. Based on bulky, expensive, and has low efficiency, transformerless invert-
the study, modulation techniques are proposed to provide ers have become a solution for high-performance application
bidirectional current path during freewheeling period. With [2][4].
proposed modulation technique, reactive power control is
achieved in H5 and HERIC inverters, without any modifica- Various topologies and modulation techniques have been in-
tion on the converter structures. The performances of the troduced and published [5][12]. H5 [7] and HERIC [8] are two
proposed modulation techniques are studied via MATLAB mainstream transformerless PV inverters due to the simple struc-
simulation and further validated with experimental results. ture and high efficiency. The leakage current has been success-
Index TermsModulation technique, reactive power, fully reduced to adhere to standard requirement. Even though
transformerless. transformerless inverters in literature are capable of suppress-
I. INTRODUCTION ing the leakage current, most of which are designed for unity
HOTOVOLTAIC (PV) energy has seen remarkable growth power factor operation only. In fact, reactive power support is
P in recent decades owing to the renewable energy policy,
feed-in-tariff and the cost-reduction of the PV installations. Ac-
required for next-generation PV inverter in order to allow high
penetration of PV system in the utility grid [13], [14]. To achieve
cording to IEA-PVPS annual report [1], the cumulative capacity this target, many international standards have been revised. Ac-
of installed PV reaches 230 GW by 2015, out of which the ma- cording to VDE-AR-N4105 [13], reactive power capability is
jority (90%) is grid-connected system. The increase in the essential for grid-connected PV inverters.
demand of the PV installation, especially grid-connected PV In light of this, conventional bipolar modulation is reported as
system, indicates that there is a need for in-depth research and a potential candidate for next-generation PV inverter [14]. Be-
development. sides leakage current elimination, bipolar modulation technique
is able to provide reactive power support. However, reactive
Manuscript received September 7, 2016; revised November 22, 2016, power capability comes at a cost of high switching loss due to
December 31, 2016, and January 20, 2017; accepted February 2, 2017.
Date of publication March 22, 2017; date of current version August 9, two-level modulation. In every switching transition, the voltage
2017. This work was supported in part by the ASEAN-ROK Academic changes across the inductor by twice of input voltage. Thus,
Exchange Programme 2016/2017, and in part by the Basic Science efficiency is low for bipolar modulation.
Research Program through the National Research Foundation of Korea
funded by the Ministry of Science, ICT and Future Planning under Grant As a result, several transformerless PV inverter topologies
2016R1A2B4010636. (Corresponding author: Tan Kheng Suan Freddy.) with reactive power capability have been proposed [14][19]
T. K. S. Freddy is with the School of Engineering, Faculty of Comput- via three-level modulation (unipolar modulation). In order to
ing, Engineering and Technology, Asia Pacific University of Technology
and Innovation, Kuala Lumpur 57000, Malaysia (e-mail: freddytan.86@ provide reactive power control in conventional H5 topology,
gmail.com). combined unipolar and bipolar pulse width modulation (PWM)
J.-H. Lee, H.-C. Moon, and K.-B Lee are with the Department of was proposed in [15]. The PWM is switched from unipolar
Electrical and Computer Engineering, Ajou University, Suwon 16499,
South Korea (e-mail: ljh20609@ajou.ac.kr; mhc0201@ajou.ac.kr; to bipolar during the negative power region. Although reactive
kyl@ajou.ac.kr). power control is attained, the implementation is complicated.
N. A. Rahim is with the UM Power Energy Dedicated Advanced Centre The current ripple and switching loss are high due to the adop-
(UMPEDAC), University of Malaya, Kuala Lumpur 59990, Malaysia, and
also with King Abdulaziz University, Jeddah 23218, Saudi Arabia (e-mail: tion of bipolar PWM.
nasrudin@um.edu.my). High-efficiency transformerless inverters with reactive power
Color versions of one or more of the figures in this paper are availablecapability were recently proposed in [16] and [17] by utilizing
online at http://ieeexplore.ieee.org.
Digital Object Identifier 10.1109/TIE.2017.2686366 MOSFETS and SiC diodes. Even though high efficiency is realized
0278-0046 2017 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.
6990 IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS, VOL. 64, NO. 9, SEPTEMBER 2017
Fig. 3. Operation modes of H5 topology with conventional PWM. (a) Mode 1conduction mode and (b) Mode 2freewheeling mode during
positive half cycle. (c) Mode 3conduction mode and (d) Mode 4freewheeling mode during negative half cycle.
Fig. 4. Operation modes of HERIC topology with conventional PWM. (a) Mode 1conduction mode and (b) Mode 2freewheeling mode during
positive half cycle. (c) Mode 3conduction mode and (d) Mode 4freewheeling mode during negative half cycle.
VDC Grid voltage (Vg ) and grid current (Ig ) have opposite polarity
VCM = . (6) during the negative power region. In order to generate unipolar
2
PWM in negative power region, VAB should be +VDC and 0,
In mode 3 Fig. 3(c), S5 , S2 , and S3 are ON while S1 and or VDC and 0, depending on the polarities of both Vg and
S4 are OFF. Current increases and flows through S5 , S2 , and S3 . Ig . When Vg is negative and Ig is positive, current freewheels
VAN = 0 and VBN = VDC , so the VCM and VDM are expressed through antiparallel diode of S2 , of S3 and of S5 , as depicted in
6992 IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS, VOL. 64, NO. 9, SEPTEMBER 2017
Fig. 5. Operation modes of conventional H5 topology in negative power Fig. 6. Operation modes of conventional HERIC topology in negative
region. (a) V g is negative and Ig is positive. (b) V g is positive and Ig is power region. (a) V g is negative and Ig is positive. (b) V g is positive and
negative. Ig is negative.
Fig. 5(a). VAN = 0 and VBN = VDC , so III. PROPOSED MODULATION TECHNIQUES WITH REACTIVE
VDM = VAB = VDC (11) POWER CAPABILITY
According to Section II, H5 and HERIC with conventional
VDC
VCM = . (12) PWMs are not suitable for reactive power application because
2
they fail to generate zero voltage during the negative power re-
gion. In light of this, simple modulation techniques are proposed
When Vg is positive and Ig is negative, current freewheels here in order to create a new current path to achieve zero-voltage
through antiparallel diode of S4 , of S1 and of S5 , as depicted in state during the negative power region.
Fig. 5(b). VAN = VDC and VBN = 0. Hence,
VDM = VAB = VDC (13) A. Operation Principle of Proposed PWM
VDC Fig. 7 shows the proposed modulation techniques for both
VCM = . (14) topologies. As shown in Fig. 7, region I and region II belong
2
to positive power region (Vg and Ig are in same polarity) while
However, there is no path for the current to flow in order
region III and region IV are negative power region (Vg and Ig are
to generate zero voltage. In fact, the same occasion follows in
in opposite polarity). The detailed principle operation of region I
HERIC inverter. With conventional switching pattern, current
and region II have been discussed earlier in Section II.
freewheels through the antiparallel diodes of the corresponding
The proposed PWM for H5 inverter is shown in Fig. 7(a). In
switches to generate VDC or +VDC as shown in Fig. 6(a) and
region III, Vg is negative while Ig remains positive. During the
(b), respectively. There is no current path to achieve zero-voltage
freewheeling period, S1 is OFF while S3 and S5 are ON, current
state. This explains why H5 and HERIC with the conventional
freewheels through the antiparallel diode of S2 , of S3 and of S5 ,
modulation techniques are not suitable for reactive power appli-
as shown in Fig. 5(a). VAN = 0 and VBN = VDC , and VCM
cation. In fact, the whole principle operation is similar to most
and VDM can be calculated as
of the topologies in the literature. They operate optimally at
unity power factor. Nonetheless, during negative power region, VDM = VAB = VDC (15)
zero-voltage state could not be realized because there is no path
VDC
for the current to flow. VCM = . (16)
Based on the analysis, principle of generating reactive power 2
in single-phase transformerless PV inverter is underlined cur- In order to achieve zero voltage in region III, instead of turning
rent path must be provided so that zero-voltage state is attained OFF S1 in the negative half cycle, S1 commutates complemen-
during negative power region. tarily to S2 to create a new current path. S1 and S3 are ON
FREDDY et al.: MODULATION TECHNIQUE FOR SINGLE-PHASE TRANSFORMERLESS PHOTOVOLTAIC INVERTERS WITH REACTIVE POWER 6993
Instead of turning OFF S3 in the positive half cycle, S3 com- common point, VDC /2. VCM and VDM become
mutates complementarily to S4 to achieve zero-voltage state. S1 VDM = VAB = 0 (21)
and S3 are ON while S2 , S4 and S5 are OFF, current flows through
S3 and the antiparallel diode of S1 , as shown in Fig. 8(b). VDC
VCM = . (22)
VAN increases and VBN decreases until their values reach the 2
6994 IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS, VOL. 64, NO. 9, SEPTEMBER 2017
Fig. 12. V g and Ig (20) for H5 and HERIC inverters under unity power
factor operation. (a) Conventional PWM. (b) Proposed PWM. Fig. 14. V g and Ig (20) for H5 and HERIC inverters under lagging
power factor operation. (a) Conventional PWM. (b) Proposed PWM.
Fig. 15. Transient from unity power factor to lagging power factor for H5
inverter with proposed PWM. (a) V g and Ig (20). (b) Leakage current,
Fig. 13. V g and Ig (20) for H5 and HERIC inverters under leading IL e a ka g e .
power factor operation. (a) Conventional PWM. (b) Proposed PWM.
IV. SIMULATION
Simulations are carried out using MATLAB/Simulink to ana-
lyze the theoretical analysis. The dc input and the output voltages
are 400 V and 230 V/50 Hz, respectively. The stray capacitance
(CPV ) is modeled with two capacitors of 220 nF. The filters are
made up of two inductors (Lf ), each has a value of 3 mH. The
switching frequency is 10 kHz.
Fig. 16. Transient from unity power factor to lagging power factor for
Fig. 12 shows the simulated waveform of H5 and HERIC HERIC inverter with proposed PWM. (a) V g and Ig (20). (b) Leakage
inverters with conventional and proposed modulation methods current, IL e a ka g e .
under unity power operation. As expected, both methods are
working properly under unity power factor. However, the con-
ventional modulation methods are not operating correctly for leading and lagging power factor. In other words, the proposed
nonunity power factor (lagging and leading power factor). modulation methods provide H5 and HERIC inverters with re-
Current distortions are observed during negative power re- active power capability.
gion, as shown in Figs. 13(a) and 14(a). As addressed in The transient performance of the proposed PWM is studied
Section II, there is no current path for the current to flow during and the result is shown in Figs. 1518. It is clearly proven
negative power region. On the other hand, the proposed PWMs that the proposed modulation techniques are working properly
allow both H5 and HERIC inverters to operate beyond unity during the transition period for both inverters, either from unity
power factor. As shown in Figs. 13(b) and 14(b), the discussed power factor to lagging power factor as shown in Figs. 15 (H5)
inverters with proposed PWMs are operating optimally for both and 16 (HERIC), or from unity power factor to leading power
6996 IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS, VOL. 64, NO. 9, SEPTEMBER 2017
Fig. 17. Transient from unity power factor to leading power factor for H5
inverter with proposed PWM. (a) V g and Ig (20). (b) Leakage current,
IL e a ka g e .
Fig. 18. Transient from unity power factor to leading power factor for
HERIC inverter with proposed PWM. (a) V g and Ig (20). (b) Leakage
current, IL e a ka g e .
TABLE I
PARAMETERS OF INVERTER
Parameter Value
Fig. 23. Transient from unity power factor to leading power factor for
HERIC inverter with proposed PWM: V g (200 V/div) and Ig (5/div).
Fig. 24. Transient from unity power factor to lagging power factor for
H5 inverter with proposed PWM: V g (200 V/div) and Ig (5/div).
Fig. 22. Transient from unity power factor to leading power factor for
H5 inverter with proposed PWM: V g (200V/div) and Ig (5/div).
TABLE II
MEASURED EFFICIENCY OF THE DISCUSSED TOPOLOGIES
H5 with conventional PWM 90.9% 93.1% 95.1% 95.9% 96.4% 96.4% 95.75%
H5 with proposed PWM 90.2% 92.7% 94.6% 95.7% 96.1% 96.1% 95.43%
HERIC with conventional PWM 92.2% 95.2% 96.1% 96.75% 97% 97% 96.56%
HERIC with proposed PWM 91.2% 93.6% 95.2% 96.1% 96.7% 96.7% 96.02%
[21] B. Yang, W. Li, Y. Gu, W. Cui, and X. He, Improved transformerless Kyo-Beum Lee (S02M04SM10) received
inverter with common-mode leakage current elimination for a photovoltaic the B.S. and M.S. degrees in electrical and
grid-connected power system, IEEE Trans. Power Electron., vol. 27, electronic engineering from Ajou University,
no. 2, pp. 752762, Feb. 2012. Suwon, South Korea, in 1997 and 1999, re-
spectively, and the Ph.D. degree in electri-
cal engineering from Korea University, Seoul,
South Korea, in 2003.
Tan Kheng Suan Freddy received the B.Eng. From 2003 to 2006, he was with the Insti-
degree in electrical engineering from Multimedia tute of Energy Technology, Aalborg University,
University, Cyberjaya, Malaysia, in 2010, and the Aalborg, Denmark. From 2006 to 2007, he was
Ph.D. degree in electrical engineering from the with the Division of Electronics and Information
University of Malaya, Kuala Lumpur, Malaysia, Engineering, Chonbuk National University, Jeonju, South Korea. In 2007,
in 2015. he joined the School of Electrical and Computer Engineering, Ajou
He is currently a Lecturer at the Asia Pacific University. He is an Associate Editor of the IEEE TRANSACTIONS ON
University of Technology and Innovation (APU), POWER ELECTRONICS, Journal of Power Electronics, and Journal of Elec-
Kuala Lumpur, Malaysia. Prior to APU, he was a trical Engineering and Technology. His research interests include elec-
Postdoctoral Research Fellow in the UM Power tric machine drives, renewable power generations, and electric vehicle
Energy Dedicated Advanced Centre, University applications.
of Malaya. He was a Visiting Research Scholar in the Power Electronics
Laboratory, Ajou University, Suwon, South Korea. His research interests
include power electronics converters, renewable energy, and smart grids.
Dr. Freddy received the ASEAN-Korea Exchange Fellowship Award.