International Journal of Management, Technology And Engineering ISSN NO : 2249-7455

CHARGE PUMP CIRCUIT CONCEPT FOR SINGLE-PHASE

TRANSFORMERLESS INVERTER WITH FUZZY LOGIC
CONTROLLER FOR GRID-TIED PV APPLICATIONS
ONNAMYNA HARIKRISHNA ,Bharat Institute Of Engineering And Technology College
T.SUKANTH Assistant Professor, Bharat Institute Of Engineering And Technology College
which leads to leakage current problems. For
ABSTRACT-- This project proposes a charge pump transformerless grid-connected inverters, full-bridge
circuit concept for single phase transformerless inverter (FB) inverter, neutral point clamped (NPC), active
with FLC for PV applications. The topology is derived NPC (ANPC) inverter [5], and many other topologies
from the concept of a charge pump circuit in order to such as H5, H6, and highly efficient and reliable
eliminate the leakage current. It is composed of four
inverter concept (HERIC) were proposed to reduce
power switches, two diodes, two capacitors, and an LCL
output filter. The unbiased of grid is specifically allied the leakage current with disconnecting of the grid
with negative extremity of PV board that makes a from the PV during the freewheeling modes [6].
constant basic mode vtg and zero leakage current. The However, these topologies are not totally free from
charge pump circuit generates the negative output CM current or leakage current. The leakage current
voltage of the proposed inverter during the negative still exists due to the parasitic capacitor of the switch
cycle. A analogous full control procedure is used to and stray capacitance between the PV panel and
control the injected current. The precept advantages of ground. So, some of these topologies require two or
furthered inverter are: 1) unbiased of grid is directly more filter inductors to reduce the leakage current,
allied with -ve terminal of PV board, so the spillage
which leads to a rise in the volume and cost of the
current is wiped out; 2) its minimized size; 3) ease; 4)
the exerted dc vtg of furthered inverter is same as the system [7].
FL inverter not at all like neutral point clasped (NPC), Fig. 1 illustrates a single-phase grid-tied
active NPC, and HB inverters); 5) adaptable transformerless inverter with CM current path, where
establishing setup; 6) ability of RP stream; and 7) high P and N are the positive and negative terminals of the
productivity. An entire depiction of working precept PV, respectively. The leakage current (iLeakage)
and examination of furthered in verter are exhibited. flows through a parasitic capacitor (CP ) between the
The obtained results clearly validate the performance of filters (L1 and L2), the inverter, grid, and ground
the proposed inverter and its practical application in impedance (zg). This leakage current may cause
grid-tied PV systems.
safety problems, reduce the quality of injection
INTRODUCTION
current to the grid, as well as decrease the system
Over the last two decades, the photovoltaic
efficiency [8].
(PV) power systems have become very popular
among the renewable energy sources, because they
generate electricity with no moving parts, operate
quietly with no emissions, and require little
maintenance [1], [2]. Distributed grid-connected PVs
are playing an increasingly role as an integral part of
the electrical grid. However, due to the large stray
capacitors between the PV panels and the ground, PV
Fig. 1. Block diagram of a single-phase grid-
systems suffer from a high common mode (CM)
connected transformer less inverter with a leakage
current, which reduces the system efficiency and may
current path.
cause safety issues like electric shock. In order to
In order to eliminate the leakage current, the
eliminate the leakage currents, transformers are
CM voltage (CMV) (vcm)must be kept constant
commonly used in the PV system to provide galvanic
during all operation modes according to [9]. The vcm
isolation. However, it possesses undesirable
with two filter inductors (L1, L2) is calculated as
properties including large size, high cost, and weight
follows:
with additional losses [3]. Thus, eliminating the
+ ( − )( − )
transformer is a great benefit to further improve the = + (1)
overall system efficiency, reduce the size, and weight 2 2( + )
[4]. where vAn and vBn are the voltage
One of the important issues in the differences between the midpoints A and B of the
transformerless gridconnected PV applications is the inverter to the dc bus minus terminal N, respectively.
galvanic connection of the grid and PV system, If L1 = L2 (asymmetrical inductor), vcm is calculated

Volume 8, Issue XII, DECEMBER/2018 Page No:188

International Journal of Management, Technology And Engineering ISSN NO : 2249-7455

according to (1) and the leakage current appears due modes. This topology does not allow for reactive
to a varying CMV. If L1 = L2 (symmetrical power flow [12].
inductor), vcm is simplified to Topologies based on H6 are also proposed
+ in [13] and [14] to eliminate the leakage current of
= = (2) the grid-tied PV application. These inverters consist
2
In this state, the CMV is constant and the of six power switches and some diodes for
leakage current is eliminated. In some structures such disconnecting the dc side from the grid. These
as the virtual dc-bus inverter [10] and NPC inverter, topologies are more costly than the FB inverter,
one of the filter inductors is zero and only one filter because they use extra switches and diodes. Another
inductor is used. In this state, after simplification of disadvantage of these topologies is lower efficiency
vcm, it will have a constant value according to (3) due to the current that circulates through three power
and the leakage current will be eliminated switches at the same time [15]. Several high efficient
= + = ( = 0) new H6 transformerless inverters are proposed in
[16] and [17] to achieve lightweight and also lower
cost. They have the capability of reactive power
+ − injection to the grid. The leakage current is not totally
= −
2 2 eliminated in these topologies, which is the main
= ( = 0) disadvantage of them.
(3) Another solution to eliminate the leakage
current is the direct connection of the negative PV
As shown in Fig. 2, there are various terminal to the neutral point of the grid, such as the
transformerless grid connected inverters based on the virtual dc-bus inverter in [10] and the unusual
FB inverter in the literature to overcome these topology in [18]. In these topologies, the leakage
problems. current is completely eliminated by the topology
structure. As shown in Fig. 2(c), the virtual dc-bus
inverter is composed of five insulated-gate bipolar
transistors (IGBTs), two capacitors, and one filter
inductor Lf . Only one filter inductor is used in this
topology to eliminate the leakage current, but it is
very large. The virtual dc-bus generates the negative
output voltage. The main drawback of this topology
is that there is no path to charge the capacitor C2
during the negative cycle and this will cause a high
output total harmonic distortion (THD). The topology
presented in [18], which is shown in Fig. 2(d), has a
common ground with the grid. The number of
semiconductors used in this topology is low.
However, the output voltage of this inverter is only
Fig. 2. Single-phase grid-tied transformerless PV
two levels including positive and negative voltages
inverter topologies: (a) H5 inverter, (b) HERIC
without creating the zero voltage, which requires a
inverter, (c) virtual dc-bus inverter [10], and (d) CM
large output inductor L2 and a filter. The inductor
inverter proposed in [18].
medium-type inverter [19] also called “Karschny” is
The H5 inverter that is a FB-based inverter
another topology that is derived from the buck–boost
topology, com pared to the conventional FB inverter,
topology. This inverter has a high reliability without
needs one additional switch (S5) on the dc side to
capability of giving the reactive power to the grid and
decouple the dc side from the grid as shown in Fig.
has four power switches in the current path at the
2(a). This inverter has a variable CMV with a small
same time, which will reduce the efficiency.
leakage current and it suffers from low efficiency due
This paper introduces a new transformerless
to three switches operating at the same time [11].
inverter based on charge pump circuit concept, which
As shown in Fig. 2(b), the HERIC topology
eliminates the leakage current of the grid-connected
needs two extra switches on the ac side to decouple
PV systems using a unipolar sinusoidal pulse width
the ac side from the PV module in the zero stage.
modulation (SPWM) technique. In this solution, the
HERIC combines the merits of unipolar and bipolar
neutral of the grid is directly connected to the
modulation. The main advantage of the HERIC
negative terminal of the charge pump circuit, so the
inverter is its high efficiency due to only two
voltage across the parasitic capacitor is connected to
switches operate at the same time in all operation
zero and the leakage current is eliminated. The

Volume 8, Issue XII, DECEMBER/2018 Page No:189

International Journal of Management, Technology And Engineering ISSN NO : 2249-7455

charge pump circuit is implemented to generate

negative output voltage the modulation strategy of
the proposed inverter because the leakage current is
eliminated by the circuit topology. The pro posed
topology consists of only four power switches, so the
cost of the semiconductors is reduced and the power
quality is improved by three-level output voltage in
order to reduce the output current ripple. During
operation of the proposed inverter, the current flows
through two switches; thus, the conduction loss is
also lower. The used dc voltage of the proposed
inverter is the same as the FB inverter (unlike NPC,
ANPC, and half-bridge (HB) inverters). The
proposed inverter is capable of delivering reactive
power into grid too. Thus, it can satisfy the require
ment of the standard VDE-AR-N 4105. This paper is
organized as follows. The charge pump circuit
concept is explained in Section II. Based on it, a
novel inverter topology is derived, and the Fig. 3. Schematic diagram of the proposed inverter
modulation strategy and operation principles are including the charge pump circuit.
described in detail. The current stress analysis of the =− + + (4)
switches, calculation of the power losses, and control where Vdc is the input voltage, Vcut−in−D
scheme along with a compari son have been done in 1 and Vcut−in−D 2 are the cut-in voltages of the
Section III. Experimental validations of the proposed diodes D1 and D2, respectively. For high power
grid-tied inverter is presented in Section IV, and applications, these values can be negligible.
finally, conclusions are given in Section V. The above principle is integrated into the
PROPOSED TOPOLOGY AND MODULATION proposed inverter by using additional switching
STRATEGY devices. The voltage difference between point A of
A.Charge Pump Circuit Concept the inverter to the point n is + Vdc or zero, according
The concept of a simple charge pump circuit to the switching state of the S1 and S2, respectively.
to be used in the proposed topology to generate the The voltage of C1 and C2 must be kept constant
inverter negative output voltage is shown in Fig. 3. during all operation modes by selecting proper
The circuit consists of two diodes (D1, D2) and two switching states. Point B must be connected to the
capacitors (C1, C2). This simple charge pump circuit points A or C with extra switches. This creates three
gives a negative dc output voltage at the point of C different voltages, namely + Vdc, zero, and −Vdc for
equal to the voltage of point A [20]. The capacitor C1 the inverter operation. In summary, the charge pump
is used to couple the voltage point of A to the node circuit in the transformerless inverter has the
D. Two Schottky diodes D1 and D2 are used to pump following characteristics for grid-tied applications.
the output volt age. When the diode D2 is forward 1) This circuit has a common line with the
biased, the capacitor C1 is charged by diode D2. The negative terminal of the input dc voltage and the
diode D1 is reversed in this state. When the diode D1 neutral point of the grid that causes the leakage
conducts, capacitor C2 is charged through the current to be eliminated.
capacitor C1 by using node n and switch S2. In 2) The charge pump circuit has no active
steady state, the output voltage of the negative charge device and it has a lower cost for grid-tied
pump circuit (vCn) can be derived by applications.
3) The capacitor of the proposed inverter
charges every switching cycle, which reduces the size
of the required capacitor with the switching
frequency.
4) The capacitor of the charge pump circuit
charges with a switching cycle that eliminates the
pulse duration sensitivity to generate the negative
voltage. B. Proposed Topology As shown in Fig. 4,
the proposed topology consists of four power
switches (S1 − S4), two diodes (D1, D2), two
capacitors (C1, C2) based on the charge pump circuit

Volume 8, Issue XII, DECEMBER/2018 Page No:190

International Journal of Management, Technology And Engineering ISSN NO : 2249-7455

as described in Section II-A. An LCL-filter is used to When the switches S1 and S2 are ON, the
eliminate harmonics of the output current. output voltage of the inverter (vAn) will be + Vdc
(positive state) as shown in Fig. 6(a) and (g). During
this time interval, diode D1 is reverse biased and D2
is ON, so the capacitor C1 is charged through diode
D2 and the voltage across the capacitor C2 maintains
to be constant. In this state, when the switches S2 and
S3 are ON, vAn will be 0 (zero state) as shown in
Fig. 6(b) and (h).
In the zero state at positive cycle, the
capacitors C1 and C2 are connected in parallel
Fig. 4. Proposed single-phase transformerless grid- through D1. The capacitor C2 is charged by the
connected inverter capacitor C1 with negative polarities by the charge
This new topology is modulated using pump circuit to provide the negative voltage level.
simple SPWM. Fig. 5 shows the gate drive signals for
the proposed inverter under the current lagging In the regions II and III, the negative and
condition. According to the direction of the inverter zero voltage levels are produced. Fig. 6(c) and (e)
output voltage and output current, theoperation of the shows the equivalent circuit that S4 and S1 are ON.
proposed inverter is divided in four regions as shown The negative voltage is generated, when switch S4 is
in Fig. 6. These four different regions can be defined turned ON and the voltage across the capacitor C2
as follows. appears at the inverter output voltage (vAn = −Vdc)
(negative state). The negative output voltage of the
inverter is produced by the capacitor C2. In this state,
S1 is switching simultaneously with S4. In addition
to this, C1 is charged by the capacitor CB through S1
in order to maintain a constant voltage for the
capacitor C1.
The voltage across the capacitor C1 can be
kept constant in this state by the modulation strategy.
In this period, the circuit operation of the zero state is
similar to the zero state of positive half-period of the
grid as shown in Fig. 6(b) and (h). Based on the
analysis given above, three sequences exist for the
output voltage and current of the proposed inverter.
1) If the sequence is I → III, then the inverter is in
Fig. 5. Switching pattern of the proposed topology the unity power factor condition (PF = 1). 2) If the
with reactive power flow. sequence is IV → I → II → III, then the inverter is in
the current lagging condition. 3) If the sequence is I
1) Region I: the inverter output voltage and the → IV → III → II, then the inverter is in the current
output cur rent are positive; energy is transferred leading condition. The aluminum electrolytic
from dc side to grid side as shown in Fig. 6(a). capacitors are used for the capacitors C1 and C2 of
2) Region II: the inverter output voltage is negative the proposed inverter in the experimental setup.
and the output current is positive; energy is These capacitors have a limited life span
transferred from grid side to dc link as shown in Fig. because the electrolyte finally dissipates in to the
6(c). element. Inrush current and voltage stress on the
3) Region III: the inverter output voltage and the capacitors decrease the lifetime of them.
output current are negative; energy is transferred Consequently, in order to reduce the equivalent series
from dc link to grid side as shown in Fig. 6(e). resistance (ESR) losses and to ensure the capacitor’s
4) Region IV: the inverter output voltage is positive lifetime, the capacitance of the aluminum electrolytic
and the output current is negative; energy is capacitors is chosen a little higher than the design
transferred from grid side to dc side as shown in Fig. values. One of the important elements in the
6(g). In the regions I and IV, the switches S1 and S3 capacitor design is the inrush current of the
will be ON and OFF with the switching frequency fs capacitors. In this case, the charging time constant of
to produce positive and zero voltage while S2 capacitor C2 (τC 2) can be expressed as follows:
remains ON for the whole positive half cycle.

Volume 8, Issue XII, DECEMBER/2018 Page No:191

International Journal of Management, Technology And Engineering ISSN NO : 2249-7455

switch S1 and the capacitor C2 is charged through

switch S3. Using the PSCAD software, simulation
results of the current in the switches S1 and S3 for an
output power of 500 W are shown in Fig. 7.

Fig. 7. Simulations of the current of switches: (a) grid

current (ig ) [1 A/div], (b) (iS 1 ) [2 A/div], and (c)
(iS 3 ) [2 A/div].
The selected parameters of the simulations
are the same as the experimental 500 W prototype.
These parameters are listed in Table IV. The
maximum value of the current in these switches
occurs at the negative state in simulation results as
shown in Fig. 7. In the negative half-period of the
grid, the circuit shifts between the negative and zero
states. During the negative state, the capacitor C1 is
Fig. 6. Operational stages of the proposed inverter charged by the capacitor CB , while the capacitor C2
during (a), (b) re gion I, (c), (d) region II, (e), (f) is discharged by the grid. As a result, the voltage
region III, and (g), (h) region IV. (a) vAn = +Vdc , ig values of the vC 2 − vC 1 and vCB − vC 1 are
> 0. (b) vAn = 0, ig > 0. (c) vAn = −Vdc , ig > 0. (d) increased. When the circuit shifts to zero state, the
vAn = 0, ig > 0. (e) vAn = −Vdc , ig < 0. (f) vAn = 0, capacitor CB is charged by the PV panel, whereas the
ig < 0. (g) vAn = +Vdc , ig < 0. (h) vAn = 0, ig < 0. capacitor C1 is discharged by the capacitor C2 in the
= (5) paralleling mode.
where Re1 and Ce1 will be as follows: Therefore, the voltage difference between
the capacitors C1 and CB as well as the capacitors C1
= + + + , = and C2 is decreased. The equivalent circuits of the
+
(6) zero and negative states of the proposed topology for
in which RC 1, RC 2 are the ESR of the the current stress analysis are shown in Fig. 8(a) and
capacitors, RS3 de notes the resistor of switch S3 in (b), respectively. The power switches are modeled
the conducting state, and RD1 represents the resistor with a resistor (Rs) and a constant voltage source
of diode D1. The current through capacitors (von). The capacitor is considered as an ideal
(iCapacitors) is calculated by capacitor with a series resistor (RC ), the PV panel is
− regarded as a current source (IPV ), and the diode of
= (7) the topology is modeled with a resistor (Rd). The
forward voltage drop of the diodes is neglected.
According to (5), the charging time constant According to the electric circuit theory, the
of C2 is larger than its natural discharging time grid voltage in series with a grid-side inductor (Lg )
constant and VC 1 − VC 2 has a very small value in can be equivalent with the current source (ig ) as
steady state. Due to these analyses and according to shown in Fig. 8. The value of the capacitors C1 and
(7), the capacitor current value is small in this period C2 determines the max imum current of the power
and leads to an improvement in the lifespan. switches. The additional current on the switches S1
ANALYSIS OF THE PROPOSED TOPOLOGY and S3 is created by the voltage differ ence between
A. Current Stress Analysis and Capacitors Design the capacitors C1, C2, and CB . So, by defin ing two
A current stress analysis of the proposed variables vdiff ,1 = vC 1 − vC 2 − von and vdiff ,2 =
topology is presented in this section. The maximum vCB − vC 1 − von, the state equation for zero and
value of the current stress occur on the switches S1 negative states can be taken separately.
and S3, because the capacitor C1 is charged through

Volume 8, Issue XII, DECEMBER/2018 Page No:192

International Journal of Management, Technology And Engineering ISSN NO : 2249-7455

( )
1, ℎ ℎ
= 0, ℎ ℎ (15)
−1, ℎ ℎ
Rearranging (13) and (14) and using the averaging
method in switching cycle Ts, we have
(a) 〉 (1 + ( )) 〈 .
〈 . 〉 −
( + )
〈 . 〉 ( )
+
( + )
〈 〉 (1 + ( )) ( ) ( )
=− −
( + ) ( + )
(b) 〈 〉 ( )
Fig. 8. Equivalent circuit of the proposed converter − (16)
during (a) zero state and (b) negative state. 〈 〉 ( )
.
According to Fig. 8(a), at the zero state 〈 . 〉 +
. . −
= (8) 〉 (1 + ( ))〈 .
( + ) +
. . −
( + )
= − (9) 〈 〉 (1 + ( ))
( + ) = −
According to Fig. 8(b), in -ve state ( + )
. + .
( ) ( )
= + (10) − (17)
( + )
. − .
= + (11) where d(t) and ig (t) are
( )= sin (18)
where in (8)–(11), Re2 and Ce2 will be as follows:
( )= sin (19)
= + + + , where ω signifies the rakish frequency of O/P
current, M is proportion of dc vtg and grid vtg, and
= (12) Im speaks to pinnacle O/P current of inverter.
+
By utilizing the averaging strategy at the The normal current of iS 1 and iS 3 amid Ts feasible
switching cycle Ts, and linearizing (8)– (11), the found as takes after:
normal estimation of iS 1 and iS 3 at the -ve and zero 〈 〉 + 〈 〉
.
states is equivalent to (13) and (14), separately: 〈 〉 = (20)
+
. . −
= 1+ ( ) 〈 . 〉 +
( + ) 〈 〉 = (21)
+ .
− ( ) The simulation results of present coursing via the
( + ) switches S1 and S3 for the O/P power 500 W are
+ (13) appeared in Fig. 7. The estimation of present that
goes via S1 and S3 to reach to its greatest at the -ve
. .− state is appeared in this figure as well. At the -ve
= 1+ ( ) − state,
( + )
1
− ( ) 〈 , 〉 = +1
2 + 1−
− 1−
+
.
(14) + (22)
2
1
〈 , 〉 = +1
where s(t) denotes the switching state function given 2 + 1−
as follows: 1−
+ (23)
2

Volume 8, Issue XII, DECEMBER/2018 Page No:193

International Journal of Management, Technology And Engineering ISSN NO : 2249-7455

Equations (22) and (23) indicate that the : ( )= ( ) (26)

values of C1/(C1 + C2) and C1/(C1 + CB ) should be : ( )= + ( ) (27)
calculated small enough, and the values of the
Re1Ce1 and Re2Ce2 should be smaller than the where vDS is the drain source voltage drop
switching period in order to minimize the current of the MOSFET, RDS is the drain source resistance
stress on the switches. Therefore, there is a tradeoff of the MOSFET during on the state operation, vAK is
between the current stress on the switches and the the anode cathode voltage drop of the diode, VF is
size of the switched capacitors. To achieve these the equivalent voltage drop under zero current
objectives, the capacitors C1, C2, and CB must be set condition of the diode, RAK is the anode cathode
to C1 < C2 and C1 < CB to reduce C1/(C1 + C2) and resistance of the diode during the on state, and i(t) is
C1/(C1+CB), respectively. the grid current. The duty ratio of the conducting
The value of C1/(C1 + C2) and C1/(C1 + devices is given in Table I us ing the unipolar SPWM
CB ) is almost set to 1/4 by considering the strategy in the proposed grid-connected inverter. The
capacitance of the capacitors, and hence, iS 1,max = average value of the conduction losses of the MOS
4Im and iS 3,max = 4Im . These values can be FET switch (PMOSFET Cond) during half of the
limited by a small resistor or a small inductor fundamental period is calculated by
between the capacitors if needed. According to this TABLE I
analysis, the current flowing through the capacitors DUTY RATIO OF EACH CONDUCTING DEVICE
of C1 and CB has the same value of the current
flowing through switch S3 and S1, respectively. The
relationship between the voltage and current passing
through the capacitors is calculated by
∆
= (24)
∆
The required capacitance of C1 and CB for
the proposed in verter can be derived by equaling the . =
capacitor power magnitude to the grid power ripple ∫ ( )( ) ( ) (28)
magnitude. The capacitance C1 and CB can be
calculated as follows:
( ) The average value of the conduction loss in
= (25)
the diode (PDiode Cond) during the on state mode is
where IC 1max and IC B max are the calculated by
maximum current that passes through the capacitors
C1 and CB , respectively. ∧V r, Vn, and f are the .
capacitor voltage ripple magnitude, nominal voltage 1
on the capacitor, and the frequency during maximum = ( ) ( ) (29)
current, respectively. By applying suitable values for
1
these parameters in (25), the capacitance of C1 and ( + ( ) ) ( ) ( )
CB is calculated.
The aggregate conduction losses of MOSFET
B.CONDUCTION AND SWITCHING LOSSES switches and diodes (PFurthered Cond) for the
OF POWER DEVICES furthered inverter amid a fundamental period feasible
During the positive power cycle, the grid computed from
current flows through switches S1 and S2 and the . = +
capacitor C1 is charged through diode D2 at the 4
positive state as shown in Fig. 6(a) and (g). At zero =4 +
state, the grid current flows through S2 and S3 and 3 2
the capacitor C2 is charged through diode D1. In this 4
cycle, there are two switches on the grid current path + −
3
and one diode is conducting in each state. In the
4
negative cycle of the grid, the grid current flows +2 +
through S4. The capacitor C1 is charged through 2 3
diode D2 and switch S1 at the negative state as = 3.1 (30)
shown in Fig. 6(c) and (e). The voltage drop of the
power devices can be derived by

Volume 8, Issue XII, DECEMBER/2018 Page No:194

International Journal of Management, Technology And Engineering ISSN NO : 2249-7455

The device manufacturer and circuit generate a sinusoidal current and the outer control
parameters for efficiency evaluation of proposed loop is implemented for the current reference
inverter are listed in Table II. In silicon carbide generation, where the power is controlled. A
power MOSFET switches, the recovery cur rent of proportional resonant (PR) controller provide an
the diodes is eliminated and therefore the switching infinite gain at the resonant frequency (fres) and can
losses for the diodes are negligible. The switching eliminate the steady state error when tracking a
losses of the MOSFET switch can be found as sinusoidal signal, which is an index of power quality.
follows: Due to these features, the PR controller is selected
TABLE II instead of the FLC controller in the current control
SPECIFICATIONS AND POWER DEVICES FOR loop in this topology [22]. The transfer function of
EFFICIENCY EVALUATION this controller can be found as follows:

Fig. 9. Control block diagram of the proposed single-

phase grid-tie inverter based on single-phase PQ
. = (31) theory.
where Eoss is put away energy that feasible 2
accomplished from datasheet that is equivalent to 45 ( )= + (35)
μJ. The aggregate switching losses of switches in +
furthered inverter feasible inferred as takes after:
where k p is the proportional gain, kr is the
=4 = 3.46 (32) fundamental resonant gain, and ω is the resonant
C.Conduction Losses in the Capacitors frequency. The power control loop requires
The ESR of the capacitors of the proposed
orthogonal signal genera tion systems to create
inverter is achieved from aluminum electrolyte
quadrature components (vgα, vgβ and i gα, igβ )
capacitor datasheets and it is divided into two parts.
corresponding to the grid voltage vg and grid current
The first part of the conduction losses in the ca
i g, and then it generates a current reference, which is
pacitors is similar to the conventional transformerless
used in the inner current control loop. According to
inverters including FB, H5, HERIC, etc. The second
the single-phase PQ theory [23], [24], the current
part of the conduc tion losses is related to the inrush
reference can be produced by reg ulating the active
current during the charging of the capacitors. These
and reactive powers. The active power (P) and
losses can be defined as follows:
reactive power (Q) for the proposed topology can be
. . calculated by [24]
2( + ) +
= ( ) ( ) (33) =
2
. . −
2( + ) = (36)
( ) ( ) 2
= (34)
where v gα, vgβ , igα, and igβ are the α and
where dc(t) is the duty ratio of the capacitor. β components of grid voltage and current,
The aluminum electrolytic capacitors are used in the respectively. The active power and reactive power
experimental hardware. The simulation result of this references (P∗ and Q∗) can be tuned by the operators
loss with PSCAD software is totally around 3.1 W at {R– 3} or in the control unit, when the MPPT control
500 W prototype. is activated. The current reference can be computed
D.Control Scheme in the αβ-reference frame, which simplifies the
The control strategy of the proposed grid- overall control. If PI controllers are used for power
tied single-phase inverter is shown in Fig. 9. It regulations, the grid current reference (ig-ref ) can be
contains two cascaded loops [21]: the first loop is an derived as follows [24]:
inner control loop, which has the responsibilities to

Volume 8, Issue XII, DECEMBER/2018 Page No:195

International Journal of Management, Technology And Engineering ISSN NO : 2249-7455

∗ +
=
1
[ ]
( ) ( −
∗
)
(37) =1 2 (45)
+ ( ) ( − )

where G p(s) and Gq(s) are the PI The second consideration is the current
controllers for active power and reactive power, control strategy effec on the output filter design. In
respectively. the closed-loop control system, the combination of L
The LCL filter is adopted as the grid and C can cause a resonance problem, which may
interfaced filter in this proposed topology. High lead to instability of the controller. The active
output current quality in the proposed inverter can be damping is used to smooth the resonance peak of the
obtained if the output filter is configured cor rectly. LCL filter as shown in Fig. 9 [29], [30]. A block
The first design consideration is the calculation of the diagram of the control system is shown in Fig. 10.
filter parameters, which can be determined by current The Bode diagram of the transfer function from vcref
ripple and filter values [25]. The inverter-side to ig is defined by G(s). This diagram is shown in
inductor (Lf ) value is calculated by considering 10– Fig. 11 in order to demonstrate the effect of active
20% of the ripple on the output current, which is damping by the filter capacitor current (Hcic)
given by
( − )( sin )
= (38)
∆
where fsw is the switching frequency and
ΔiL represents the peak-to-peak ripple current on the
Lf . The inverter output volt age (vAn) can be
calculated as follows: Fig. 10. Control diagram of the injected current with
capacitor current feedback active damping.
= sin (39)

By replacing (39) with (38) and simplifying it, we

have
( )( )
= (40)
∆

where RF is the ripple current and can be calculated

from
Fig. 11. Bode plot of the system in the case of
= sin − sin (41)
different values for the active damping gain HC
The maximum achievable value of
( )
modulation index (M) is RFmax = 0.25 [26]. The ( )= (46)
maximum value of the filter capacitor is calculated 1+ ( )
by (42), limiting to be less than 5% of the nominal where
value [27] 1
0.05 ( )= (47)
= (42) + (1 +
,
2
and K pwm = 1 and HC is the active damping gain.
where Pn is the nominal power, Vrms The effec of HC is shown in Fig. 11, where a high
denotes the root mean square (RMS) grid voltage, value of HC shows better resonance peak damping
and f presents the grid frequency. There is a relation capabilities.
between the inverter-side inductor (Lf ) and the grid E.Comparison With Other Known Topologies
side (Lg). This value is determined with the ratio be A comparison of the transformerless inverter
tween the ripple attenuation (r) as described in [28] structures with respect to number of semiconductor
= (43) devices, passive elements, THD of the current, and
The grid-side inductor (Lg) value feasible number of switches in the current path is listed in
determined by Table III.
10 ≤ ≤ 0.5 (44)
The resonant frequency for the LCL filter is TABLE III
given by

Volume 8, Issue XII, DECEMBER/2018 Page No:196

International Journal of Management, Technology And Engineering ISSN NO : 2249-7455

comparison of the proposed topology with the The switching losses are calculated based
conventional transformerless topologies on the datasheet of the devices. The calculation
process is studied in detail in the literature [34]. The
total semiconductor losses consist of the switching
losses, conduction losses, switching losses of diode,
and freewheeling losses. It is important to know that
the con duction loss internal SiC diode of MOSFET
This comparison includes conventional H5, implemented in the proposed inverter is less than the
HERIC inverter, and several common ground conduction loss of the IGBT body diode used in the
transformer less inverters like [31]–[33] with the H5, [31]–[33]. As shown in Fig. 12, the H5 and
proposed inverter. It can be seen from Table III that HERIC topologies have almost the same switching
the proposed topology utilizes the least amount of losses, because the number of high frequency
active and passive components compared to the other switches is the same. The proposed topology has the
topologies, which reduces the number of driver cir lowest switching losses according to the analysis
cuits, complexity of control, and power losses of the presented in Section III.
inverter. In this comparison, it is shown that the H5 According to Table III, the conduction
and HERIC inverters are using extra switches to losses of [31] are high because many switches are
disconnect the grid side from the dc side, and this used in the current path during the inverter oper
disconnection is incomplete due to the para sitic ation. Freewheeling losses of all inverter topologies
capacitance of switches. Therefore, the high- are almost the same. As a result, Shen et al. [31] have
frequency CM current flows through parasitic the highest total losses and the proposed topology has
capacitors. Thus, these topolo gies need extra filters the lowest losses because of low switching and
to absorb the CM current according to Table III and conduction losses at 500 W output power, which
the analysis given in Section I. Due to the con validate the theoretical analysis.
figuration of the charge pump circuit in the proposed SIMULATION RESULTS
topology, the CM current is completely eliminated In order to verify the feasibility of the
without extra filters. In addition, the dc voltage used proposed topology, a 500 W, single-phase
in the proposed inverter is the same as the transformerless grid-tied inverter has been built in the
conventional FB inverter, and only half of the dc laboratory and experimentally tested. The
voltage used in conventional HB, NPC, and ANPC configuration and parameters used for the ex
inverters, while the performance in eliminating the perimental tests are listed in Table IV. The
CM current is better than the FB-based inverters. The experimental results of the proposed grid-connected
THD of the grid currents for these topologies is listed inverter with unity power factor (PF = 1) operation
in Table III. It can be seen that all the single-phase are presented in Fig13. The grid voltage has the same
transformerless topologies have similar THD of the phase as the injected grid current as shown in Fig.
grid current. 13(a) and (b). Fig. 13(c) shows that the inverter
The number of switches in the current path output voltage vAn has three levels as + Vdc, 0, and
is related to the conduction losses. It can be seen −Vdc without the LCL-filter circuit. This determines
from Table III that the num ber of switches in the that the proposed topology is
current path and semiconductor devices of the CM TABLE IV
inverter [31] are larger than the other topologies. Fig. PARAMETERS FOR THE 500 W
12 shows the power semiconductor losses PROTOTYPE
distribution ac cording to the switching frequency at
the rated power of all inverter topologies listed Table
III with the same circuit param eters.

Fig. 12. Total semiconductor power losses

distribution for the H5, HERIC, [31]–[33] and the
proposed topologies at rated power.

Volume 8, Issue XII, DECEMBER/2018 Page No:197

International Journal of Management, Technology And Engineering ISSN NO : 2249-7455

The partial enlargement of the grid current

ig, grid voltage vg, and three level output voltage
vAn is provided in Fig. 16. It is clear that the pulse
duration of the output voltage (vAn) is in agreement
with the switching frequency. The experimental
results of the voltage across the switches S1, S2, S3,
and S4 under 400 V input voltage conditions are
shown in Fig. 18. The voltage stress of the switches
Fig. 13 Simulation results of the proposed topology
S1, S2, and S3 is the same as the input dc voltage and
with unity power factor (PF = 1) operation. (a) vg
on the switch S4 is double of the input voltage. The
[500 V/div], (b) ig [5 A/div], (c) vA n [500 V/div],
voltage stress of the capacitors and diodes is shown
and (d) fast Fourier transform analysis of ig [50 V/div
in Fig. 18. From the experimental waveforms of the
and 125 Hz/div].
output voltage, it can be seen that there are no voltage
modulated with unipolar SPWM. From Fig.
overshoots across the MOSFET switches.
13, it is clear that the output current and voltage of
the proposed inverter are highly sinusoidal with low
harmonic distortion due to the three-level inherency
of the output voltage. The current harmonic
distribution is demonstrated in Fig. 13(d). It can be
seen that the 5th-, 7th-, and 11th-order harmonics are
very low and the output voltage and current of the
topology are very close to be sinusoidal.
The experimental results for the THD
measurement of the grid current are equal to 2.1% in Fig. 16. Experimental enlarged results of the
unity power factor operation and it is lower than the proposed topology. (a) vAn [500 V/div], (b) vg [250
recommended value by the IEEE STD 519- 1992 V/div], and (c) ig [5 A/div], time [400 μs/div]
[35]. The RMS value of the current injected to the
grid is equal to 2.3 A, and the output power is 500 W.
Figs. 14 and 15 demonstrate the experimental results
for the inverter operating under current lagging
condition (PF = + 0.8) and leading condition (PF = –
0.8), respectively.

Fig. 17. Experimental results for drain source voltage

of switches. (a) vS 1 [250 V/div], (b) vS 2 [500
V/div], (c) vS 3 [250 V/div], and (d) vS 4 [500
V/div].

Fig. 14. Experimental results of the proposed

topology with lagging power factor operation. (a) vg
[500 V/div] and (b) ig [2 A/div].

Fig. 19. Experimental results of the capacitor and

diode voltages. (a) vC 1 [250 V/div], (b) vC 2 [500
V/div], (c) vD 1 [250 V/div], and (d) vD 2 [250
V/div].
Fig. 19 (a)–(c) shows the experimental waveforms of
Fig. 15. Experimental results of the proposed the grid voltage vg, the grid current ig, and the CM
topology with leading power factor operation. (a) vg voltage vCM in the proposed topology. It can be seen
[500 V/div] and (b) ig [2 A/div]. that by applying unipolar mod ulation SPWM

Volume 8, Issue XII, DECEMBER/2018 Page No:198

International Journal of Management, Technology And Engineering ISSN NO : 2249-7455

strategy, the CMV of the proposed inverter has been may be noted that the efficiency diagram covers the
kept constant. As a result, the CM current of the total power device losses and the filter inductor
proposed inverter is eliminated due to the losses.
configuration of the charge pump circuit. Therefore,
this solution with a charge pump cir cuit provides a
new idea for using the proposed inverter for PV
applications.

Fig. 21. Efficiency comparison for transformerless

topologies under different output power and unit
power factor.

The curves show that the HERIC has the

highest efficiency for the low power applications but
Fig. 19. Experimental results of the proposed the leakage current characteristic is worse in
topology. (a) vg [250 V/div], (b) ig [5 A/div], and (c) comparison to the proposed topology. The efficiency
CMV [200 V/div]. of the proposed inverter is lower than the H5 and a
The performance of the control strategy is little less than the HERIC topology at low power,
confirmed by ap plying a step change to the proposed while it is better at high power. The measured
inverter. A PR current controller is adopted with ki = maximum efficiency of the proposed inverter is
2000 and kp = 20 as shown in (35). Fig. 20 shows the around 97.4% at 500 W. The experimental efficiency
performance of the proposed inverter under the load of the proposed inverter confirms the power loss
step change. It has been demonstrated that the calculation in Section III-B . As shown in (48), the
proposed inverter can track step change when the California Energy Commission (CEC) efficiency is
output power is decreased from 200 W to zero. As it calculated combining different weighted factors at
can be seen, this test demonstrates the effectiveness different output power levels [36]
of the power control strategy used in this paper in
terms of fast response

The calculated CEC efficiency of the

proposed transformer less inverter is 97.2%. As a
result, the conduction loss in the semiconductors of
the proposed inverter is similar to the H5 and HERIC
inverters and the total device loss for the HERIC
topol ogy at low power is higher than the proposed
inverter. Hence, the proposed inverter reduces the
output current ripple in addition to maintain a high
Fig. 20. Experimental results of the proposed efficiency without leakage current.
topology in dynamic state. (a) vg [250 V/div], (b) ig CONCLUSION
[10 A/div], and (c) active power (P) [400 W/div].. This paper proposed a new single-phase
transformerless in verter for a grid-tied PV system
As described in the previous section, the using a charge pump circuit concept. The concept is
efficiency of the pro posed inverter mainly relies on proposed to generate the negative output voltage in
the losses of the power devices. Fig. 21 shows the the proposed inverter. This new topology generates a
efficiency of the H5, HERIC, [31]–[33] and the three level output voltage by employing unipolar
proposed inverter {R–2} having the same output SPWM. The negative terminal of the proposed
power and unity PF. The efficiency comparison topology is the same as the neutral line in the grid;
curves illustrate the exper imental results under the thus, the leakage current is well sup pressed and the
same dc-link voltages, output power, switching transformer is eliminated. The proposed topol ogy
frequency, and 220 Vrms ac output voltage has also the ability to deliver reactive power into the
conditions. The only difference is the components. grid. In addition, the proposed topology can be
IGBTs are used in the H5, HERIC, [31]–[33], while realized with a mini mum number of components;
SiC MOSFETs are used in the pro posed inverter. A hence, a higher power density can be achieved with
VOLTECH PM300 power analyzer has been used to lower design cost. Compared to other existing
measure the efficiency of the proposed inverter. It

Volume 8, Issue XII, DECEMBER/2018 Page No:199

International Journal of Management, Technology And Engineering ISSN NO : 2249-7455

transformerless topologies, the performance depicted photovoltaic systems,” IEEE Trans. Ind. Electron.,
by the pro posed inverter is good. A theoretical vol. 58, no. 12, pp. 5385–5395, Dec. 2011.
analysis performed and it is validated by [10] Y. Gu, W. Li, Y. Zhao, B. Yang, C. Li, and X.
experimental results for a grid-connected inverter He, “Transformerless inverter with virtual dc bus
prototype. The proposed topology is verified with a concept for cost-effective grid-connected PV power
500 W pro totype. The maximum efficiency of the systems,” IEEE Trans. Power Electron., vol. 28, no.
proposed inverter is mea sured to be 97.4%. 2, pp. 793–805, Feb. 2013.
Experimental results and loss calculations [11] W. Li, Y. Gu, H. Luo, W. Cui, X. He, and C.
demonstrate the validity of the proposed inverter with Xia, “Topology review and derivation methodology
lower THD for the grid-connected inverter. It can be of single phase transformerless photovoltaic inverters
concluded that the proposed topology is suitable for a for leakage current suppression,” IEEE Trans. Ind.
grid-tied transformerless inverter. Electron., vol. 62, no. 7, pp. 4537–4551, Jul. 2015.
REFERENCES [12] H. Schmidt, C. Siedlke, and J. Ketterer,
[1] S. B. Kjaer, J. K. Pedersen, and F. Blaabjerg, “A “DC/AC converter to convert direct electric voltage
review of single-phase grid connected inverters for to alternating voltage or current,” U.S. Patent 7 046
photovoltaic modules,” IEEE Trans. Ind. Electron., 534 B2, May 16, 2006.
vol. 41, no. 5, pp. 1292–1306, Sep./Oct. 2005. [13] M. Islam and S. Mekhilef, “H6-type
[2] X. Guo, R. He, J. Jian, Z. Lu, X. Sun, and J. M. transformerless single-phase in verter for grid-tied
Guerrero, “Leakage current elimination of four-leg photovoltaic system,” IET Power Electron., vol. 8,
inverter for transformerless three-phase PV systems,” pp. 636–644, 2015. [14] W. Yu, J. Lai, H. Qian, and
IEEE Trans. Power Electron., vol. 31, no. 3, pp. C. Hutchens, “High-efficiency MOSFET in verter
1841–1846, Mar. 2016. with H6-type configuration for photovoltaic non
[3] W.-J. Cha, K.-T. Kim, Y.-W. Cho, S.-H. Lee, and isolated ac-module applications,” IEEE Trans. Power
B.-H. Kwon, “Evalu ation and analysis of Electron., vol. 26, no. 4, pp. 1253–1260, Apr. 2011.
transformerless photovoltaic inverter topology for [15] W. Cui, B. Yang, Y. Zhao, W. Li, and X. He,
efficiency improvement and reduction of leakage “A novel single-phase transformerless grid-connected
current,” IET Power Electron., vol. 8, no. 2, pp. 255– inverter,” in Proc. 37th Annu. Conf. IEEE Ind.
267, 2015. Electron. Soc., 2011, pp. 1067–1071.
[4] H. Xiao and S. Xie, “Leakage current analytical [16] M. Islam, N. Afrin, and S. Mekhilef, “Efficient
model and application in single-phase transformerless single phase transformer less inverter for grid-tied
photovoltaic grid-connected inverter,” IEEE Trans. PVG system with reactive power control,” IEEE
Electromagn. Compat., vol. 52, no. 4, pp. 902–913, Trans. Sustain. Energy, vol. 7, no. 3, pp. 1205–1215,
Nov. 2010. Jul. 2016.
[5] T. Bruckner, S. Bernet, and H. G ¨ uldner, “The [17] M. Islam and S. Mekhilef, “Efficient
active NPC converter and ¨ its loss-balancing transformerless MOSFET inverter for a grid-tied
control,” IEEE Trans. Ind. Electron., vol. 52, no. 3, photovoltaic system,” IEEE Trans. Power Electron.,
pp. 855–868, Jun. 2005. vol. 31, no. 9, pp. 6305–6316, Sep. 2016.
[6] D. Barater, E. Lorenzani, C. Concari, G. [18] N. Vazquez, M. Rosas, C. Hern ´ andez, E. V ´
Franceschini, and G. Buticchi, “Recent advances in azquez, and F. Perez, “A new ´ common-mode
single-phase transformerless photovoltaic inverters,” transformerless photovoltaic inverter,” IEEE Trans.
IET Renewable Power Gener., vol. 10, no. 2, pp. Ind. Electron., vol. 62, no. 10, pp. 6381–6391, Oct.
260–273, 2016. 2015.
[7] W. Yu et al., “High-efficiency inverter with H6- [19] Wechselrichter, German Patent DE 19642522
type configuration for photovoltaic non-isolated ac C1, Apr. 23, 1998.
module applications,” in Proc. Annu. IEEE Appl. [20] G. F. W. Khoo, R. H. Douglas, and R. A.
Power Electron. Conf. Expo., 2010, pp. 1056–1061. McMahon, “Analysis of a charge pump power supply
[8] B. Yang, W. Li, Y. Gu, W. Cui, and X. He, with a floating voltage reference,” IEEE Trans.
“Improved transformerless inverter with common- Circuits Syst. I, Fundam. Theory Appl., vol. 47, no.
mode leakage current elimination for a photovoltaic 10, pp. 1494–1501, Oct. 2000.
grid connected power system,” IEEE Trans. Power [21] F. Blaabjerg, R. Teodorescu, M. Liserre, and A.
Electron., vol. 27, no. 2, pp. 752–762, Feb. 2012. V. Timbus, “Overview of control and grid
[9] F. Bradaschia, M. C. Cavalcanti, P. E. P. Ferraz, synchronization for distributed power generation sys
F. A. S. Neves, E. C. Santos, and J. H. G. M. da tems,” IEEE Trans. Ind. Electron., vol. 53, no. 5, pp.
Silva, “Modulation for three-phase transformerless z- 1398–1409, Oct. 2006.
source inverter to reduce leakage currents in [22] W. L. Chen and J. S. Lin, “One-dimensional
optimization for proportional resonant controller

Volume 8, Issue XII, DECEMBER/2018 Page No:200

International Journal of Management, Technology And Engineering ISSN NO : 2249-7455

design against the change in source impedance and [34] F. Hong, R. Z. Shan, H. Z. Wang, and Y.
solar irradiation in PV systems,” IEEE Trans. Ind. Yangon, “Analysis and calculation of inverter power
Electron., vol. 61, no. 4, pp. 1845–1854, Apr. 2014. loss,” Proc. CSEE, vol. 28, no. 15, pp. 72–78, May
[23] M. Saitou and T. Shimizu, “Generalized theory 2008.
of instantaneous active and reactive powers in single- [35] IEEE Recommended Practices and
phase circuits based on Hilbert transform,” in Proc. Requirements for Harmonic Control in Electrical
33rd Annu. IEEE Power Electron. Spec. Conf., 2002, Power Systems, IEEE Standard 519-1992, 1993.
vol. 3, pp. 1419–1424. [36] C. Liu, Y. Wang, J. Cui, Y. Zhi, M. Liu, and G.
[24] Y. Yang, F. Blaabjerg, and H. Wang, “Low Cai, “Transformerless photovoltaic inverter based on
voltage ride-through of single phase transformerless interleaving high-frequency legs having bidirectional
photovoltaic inverters,” IEEE Trans. Ind. Appl., vol. capability,” IEEE Trans. Power Electron., vol. 31, no.
50, no. 3, pp. 1942–1952, May/Jun. 2014. 2, pp. 1131–1142, Feb. 2016.
[25] A. Timbus, M. Liserre, R. Teodorescu, P.
Rodriguez, and F. Blaab jerg, “Evaluation of current
controllers for distributed power generation systems,” Author’s profile:
IEEE Trans. Power Electron., vol. 24, no. 3, pp. 654– T.Sukanth,M.tech(PhD).
664, Mar. 2009. Completed B.Tech from
[26] S. V. Araujo, P. Zacharias, and R. Mallwitz, Sree Vishveswaraya
“Highly efficient single-phase transformerless
Institute Of Technology &
inverters for grid-connected photovoltaic systems,”
IEEE Trans. Ind. Electron., vol. 57, no. 9, pp. 3118– Science , Mahabubnagar,and
3128, Sep. 2010. has done M.Tech from
[27] P. Channegowda and V. John, “Filter Sreenidhi Institute Of
optimization for grid interactive voltage source Science And Technology
inverters,” IEEE Trans. Ind. Electron., vol. 57, no. Hyderabad.Currently
12, pp. 4106–4114, Dec. 2010. working as assistant
[28] A. G. Matias, F. Alejandro, L. Piotr, C. Tomasz,
professor in Bharat Institute Of Engineering And
D. Valcan, and P. Ale jandro, “Control of grid
converter for wind turbines,” Partial fulfillment of the TechnologyMangalpally,Ibrahimpatnam.
requirements for M.S. degree, Aalborg Univ., Mail Id: tsukanth@biet.ac.in
Aalborg, Denmark, Groups PED-842, MCE2-830,
2009. Onnamyna Harikrishna.
[29] R. P. Alzola, M. Liserre, F. Blaabjerg, R. Completed B.Tech from
Sebastian, J. Dannehl, and F. W. Fuchs, “Analysis of Vathsalya Institute Of Science
the passive damping losses in LCL-filter-based grid And Technology Bhongir,
converters,” IEEE Trans. Power Electron., vol. 28, Yadhadhri buvanagiri,
no. 6, pp. 2642– 2646, Jun. 2013.
[30] W. Wu, Y. He, T. Tang, and F. Blaabjerg, “A Telangana .Pursuing M.Tech
new design method for the passive damped LCL and from Bharat Institute Of
LLCL filter-based single-phase grid-tied inverter,” Engineering And
IEEE Trans. Ind. Electron., vol. 60, no. 10, pp. 4339– TechnologyMangalpally, Ibrahimpatnam.
4350, Oct. 2013. Mail Id: harikrishna830@gmail.com
[31] J. M. Shen, H. L. Jou, and J. C. Wu, “Novel
transformerless grid-connected power converter with
negative grounding for photovoltaic generation
system,” IEEE Trans. Power Electron., vol. 27, no. 4,
pp. 1818–1829, Apr. 2012.
[32] L. C. Breazeale and R. Ayyanar, “A
photovoltaic array transformer-less inverter with film
capacitors and silicon carbide transistors,” IEEE
Trans. Power Electron., vol. 30, no. 3, pp. 1297–
1305, Mar. 2015. [33] Y. Wang and R. Li, “Novel
high-efficiency three-level stacked-neutral point-
clamped grid-tied inverter,” IEEE Trans. Ind.
Electron., vol. 60, no. 9, pp. 3766–3774, Sep. 2013.

Volume 8, Issue XII, DECEMBER/2018 Page No:201