IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS, VOL. 65, NO.

7, JULY 2018 5561

A Buck and Boost Based Grid Connected PV

Inverter Maximizing Power Yield From Two PV
Arrays in Mismatched Environmental Conditions
Subhendu Dutta and Kishore Chatterjee, Member, IEEE

Abstract—A single phase grid connected transformerless in a PV array is large. In order to achieve desired magnitude
photovoltaic (PV) inverter, which can operate either in buck for the input dc-link voltage of the inverter of a grid connected
or in boost mode, and can extract maximum power simulta- transformerless (GCT) PV system, the requirement of series
neously from two serially connected subarrays while each of
the subarray is facing different environmental conditions, is connected modules becomes high. Therefore, the power output
presented in this paper. As the inverter can operate in buck from a GCT PV system such as single phase GCT (SPGCT)
as well as in boost mode, depending on the requirement, the inverter based systems derived from H-bridge [2], [3] and
constraint on the minimum number of serially connected so- neutral point clamp (NPC) inverter based systems [4], [5] get
lar PV modules that is required to form a subarray is greatly
affected significantly during MEC.
reduced. As a result, power yield from each of the subarray
increases when they are exposed to different environmental In order to address the problem arising out of MEC in a PV
conditions. The topological configuration of the inverter and system, various solutions are reported in the literature. An ex-
its control strategy are designed so that the high-frequency haustive investigation of such techniques has been presented in
components are not present in the common mode voltage, [6]. Power extraction during MEC can be increased by choos-
thereby restricting the magnitude of the leakage current as-
ing proper interconnection between PV modules [6], [7] or
sociated with the PV arrays within the specified limit. Fur-
ther, high operating efficiency is achieved throughout its by tracking global maximum power point (MPP) of PV ar-
operating range. A detailed analysis of the system leading ray by employing complex MPP tracking (MPPT) algorithm
to the development of its mathematical model is carried out. [6], [8]. However, these techniques are not effective for low
The viability of the scheme is confirmed by performing de- power SPGCT PV system. Similarly, reconfiguration of the PV
tailed simulation studies. A 1.5 kW laboratory prototype is
modules in a PV array by changing the electrical connection
developed, and detailed experimental studies are carried
out to corroborate the validity of the scheme. of PV modules [9], [10] is not effective for SPGCT PV system
due to the considerable increment in component count and es-
Index Terms—Buck and Boost based photovoltaic (PV) in- calation in operating complexity. In order to extract maximum
verter, grid connection, maximum power point (MPP), mis-
matched environmental condition, series connected mod-
power from each PV module during MEC, attempts have been
ule, single phase, transformerless. made to control each PV module in a PV array either by hav-
ing a power electronic equalizer [11] or by interfacing a dc to
I. INTRODUCTION dc converter [1], [12]–[14]. Schemes utilizing power electronic

T HE major concern of a photovoltaic (PV) system is to equalizer require large component count thereby increasing the
ensure optimum performance of individual PV modules cost and operation complexity of the system. The scheme pre-
in a PV array while the modules are exposed to different sented in [1] uses generation control circuit (GCC) to operate
environmental conditions arising due to difference in insolation each PV module at their respective MPP wherein the difference
level and/or difference in operating temperature. The presence in power between each module is only processed through the
of mismatch in operating condition of modules significantly GCC. Scheme presented in [12] uses shunt current compensa-
reduces the power output from the PV array [1]. The problem tion of each module as well as series voltage compensation of
with the mismatched environmental conditions (MECs) be- each PV string in a PV array to enhance power yield during
comes significant if the number of modules connected in series MEC. The schemes based on module integrated converter [13],
[14] use dedicated dc to dc converter integrated with each PV
Manuscript received July 12, 2017; revised October 10, 2017; ac- module. However, the efficiency of the aforesaid schemes are
cepted October 31, 2017. Date of publication November 17, 2017; date low due to the involvement of large number of converter stages,
of current version March 6, 2018. (Corresponding author: Subhendu
Dutta.)
and further in these schemes the component count is high, and
The authors are with the Department of Electrical Engineering, In- hence, they face similar limitations as that of power electronic
dian Institute of Technology Bombay, Mumbai 400076, India (e-mail: equalizer based scheme. Instead of ensuring MPP operation of
subdut87@gmail.com; kishore@ee.iitb.ac.in).
Color versions of one or more of the figures in this paper are available
each and every module, certain number of modules are con-
online at http://ieeexplore.ieee.org. nected in series to form a string and the so formed strings are
Digital Object Identifier 10.1109/TIE.2017.2774768 then made to operate under MPP in [15] and [16]. Even then,

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.
5562 IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS, VOL. 65, NO. 7, JULY 2018

there is not much reduction in overall component count and

control complexity [6].
In order to simplify the control configuration and to reduce the
component count, schemes reported in [17] and [18] combine
all the PV modules into two subarrays, and then, each of the
subarrays is made to operate at their respective MPP. However,
the reported overall efficiency of both the schemes are poor. By
introducing a buck and boost stage in SPGCT PV inverter, power
extraction during MEC is improved in [19]–[21]. Further, as a
consequence of the presence of the intermediate boost stage, the
requirement of series connected PV modules in a PV array has
become less. In the schemes presented in [19]–[21], the switches
of either the dc to dc converter stage or inverter stage operate
Fig. 1. Dual buck and boost based Inverter.
at high frequency, as a result there is a considerable reduction
in the size of the passive element count, thereby improving
the operating efficiency of these schemes. Further, the reported
efficiency of [20] and [21] is 1–2% higher than that of [19].
An effort has been made in this paper to divide the PV mod-
ules into two serially connected subarrays and controlling each
of the subarray by means of a buck and boost based inverter
so that optimum power evacuation from the subarrays is as-
certained during MEC. This process of segregation of input PV
array into two subarrays reduces the number of series connected
Fig. 2. Buck stage and boost stage of the proposed inverter.
modules in a subarray almost by half compared to that of the
schemes proposed in [20] and [21]. The topological structure
and control strategy of the proposed inverter ensure that the self-commutated switches, S1 along with its antiparallel body
magnitude of leakage current associated with the PV arrays re- diode D1 , S3 along with its antiparallel body diode D3 , the free
mains within the permissible limit. Further, the voltage stress wheeling diodes Df 1 , Df 3 , and the filter inductors and capaci-
across the active devices is reduced almost by half compared tors L1 , Cf 1 , and Co1 . Similarly, the segment, CONV2 consists
to that of the schemes presented in [20] and [21]; hence, very of the self-commutated switches, S2 along with its antiparallel
high-frequency operation without increasing the switching loss body diode D2 , S4 along with its antiparallel body diode D4 ,
is ensured. High-frequency operation also leads to the reduction the free wheeling diodes Df 2 , Df 4 , and the filter inductors and
in the size of the passive elements. As a result the operating capacitors L2 , Cf 2 , and Co2 . The inverting stage is consisting
efficiency of the proposed scheme is high. The measured peak of the self-commutated switches S5 , S6 , S7 , S8 and their corre-
efficiency and the European efficiency (ηeuro ) of the proposed sponding body diodes D5 , D6 , D7 , and D8 , respectively. The
scheme is found to be 97.65% and 97.02%, respectively. inverter stage is interfaced with the grid through the filter in-
The detailed operation of the proposed inverter with math- ductor, Lg . The PV array to the ground parasitic capacitance is
ematical validation is explained in Section II. Afterwards the modeled by the two capacitors Cpv1 and Cpv2 .
mathematical model of the proposed inverter has been derived Considering Fig. 2, CONV1 operates in buck mode when
in Section III followed by the philosophy of control strategy in Vpv1 ≥ vco1 , while CONV2 operates in buck mode when Vpv2 ≥
Section IV. The criteria to select the values of the output filter vco2 . Vpv1 , Vpv2 are the MPP voltages of PV1 and PV2 and vco1 ,
components are presented in Section V. The proposed scheme vco2 are the output voltages of CONV1 and CONV2 , respec-
is verified by performing extensive simulation studies and the tively. During buck mode duty ratios of the switches, S1 and
simulated performance is presented in Section VI. A 1.5 kW lab- S2 are varied sinusoidally to ensure sinusoidal grid current (ig )
oratory prototype of the proposed inverter has been fabricated to while S3 and S4 are kept OFF. When Vpv1 < vco1 , CONV1 oper-
carry out thorough experimental studies. The measured perfor- ates in boost mode while CONV2 operates in boost mode when
mances of the scheme, which confirm its viability are presented Vpv2 < vco2 . During boost mode duty ratios of the switches, S3
in Section VII. and S4 are varied sinusoidally to ensure sinusoidal ig while S1
and S2 are kept on throughout this mode. The sinusoidal switch-
ing pulses of the switches of CONV1 and CONV2 are synchro-
II. PROPOSED INVERTER AND ITS OPERATION nized with the grid voltage, vg to accomplish unity power factor
The schematic of the proposed dual buck and boost based operation. The switches S5 and S8 are kept ON, and switches S6
inverter (DBBI), which is depicted in Fig. 1, is comprised of and S7 are kept OFF permanently during the entire positive half-
a dc to dc converter stage followed by an inverting stage. The cycle (PHC) while during entire negative half-cycle (NHC), the
dc to dc converter stage has two dc to dc converter segments, switches, S6 and S7 are kept ON and switches, S5 and S8 are
CONV1 and CONV2 to service the two subarrays: PV1 and kept OFF permanently. All the operating states of the proposed
PV2 of the solar PV array. The segment CONV1 consists of the inverter are depicted in Fig. 3.
DUTTA AND CHATTERJEE: BUCK AND BOOST BASED GRID CONNECTED PV INVERTER MAXIMIZING POWER YIELD FROM TWO PV ARRAYS 5563

Hence, the instantaneous injected power to the grid, pg can be

written as
pg = vg ig = (vco1 + vco2 )ig (4)
wherein vco1 and vco2 denote the instantaneous quantities of Vco1
and Vco2 , respectively. As ig is in-phase with vg
Pg
Ig = (5)
Vg
wherein Vg and Ig denote rms values of vg and ig , respectively.
The power injected to the grid can be expressed as
!
1 π
Pg = pg d(ωt)
π 0
! !
1 π 1 π
= vco1 ig d(ωt) + vco2 ig d(ωt) (6)
π 0 π 0
= Pco1 + Pco2 . (7)
As vco1 and vco2 are synchronized with vg . Hence
!
1 π
Pco1 = Vco1m sin(ωt) Ig m sin(ωt) d(ωt)
π 0
Vco1m Ig m
= . (8)
2
Similarly
Vco2m Ig m
Pco2 = (9)
2
wherein the amplitudes of vco1 , vco2 , and ig are denoted
as Vco1m , Vco2m , and Ig m , respectively. Combining (1), (8),
and (9)
√ √
2Ppv1 2Ppv1 2Ppv1
Vco1m = = = (10)
Ig m Ig Pg /Vg
√ √
Fig. 3. Operating states of DBBI. (a) Active and (b) freewheeling states 2Ppv2 2Ppv2 2Ppv2
in buck mode of PHC, (c) active and (d) freewheeling states in buck mode Vco2m = = = . (11)
of NHC, (e) active and (f) freewheeling states in boost mode of PHC, Ig m Ig Pg /Vg
and (g) active and (h) freewheeling states in boost mode of NHC.
Similarly by combining (2), (10), and (11)
When the insolation level and ambient temperature of subar- Vm Ppv1 Vm Ppv2
Vco1m = and Vco2m = . (12)
ray PV1 are different from that of PV2 , the MPP parameters of Ppv1 + Ppv2 Ppv1 + Ppv2
the two subarrays Vpv1 and Vpv2 , MPP current Ipv1 and Ipv2 cor-
The voltage templates of vco1 and vco2 appear as full wave rec-
respond to PV1 and PV2 , respectively, and power at MPP Ppv1
tified sinusoidal waveform with amplitudes, Vco1m and Vco2m ,
and Ppv2 correspond to PV1 and PV2 , respectively, differ from
respectively. Vm is the amplitude of vg . It can be deduced from
each other. By considering that both the subarrays are operating
(12) that the magnitudes of Vco1m and Vco2m are decided by
at their respective MPP and neglecting the losses incurred in
the power extracted from each of the subarray. If the power ex-
power processing stages, the average power involved with Co1
tracted from PV1 is less than PV2 , then Vco1m < Vco2m , whereas
and Co2 , Pco1 and Pco2 over a half-cycle can be assumed equal
Vco2m < Vco1m if power extracted from PV2 is less than PV1 .
to the power extracted from PV1 and PV2 . Therefore
During buck mode, the duty ratios, d1 of S1 and d2 of S2 vary
Pco1 = Ppv1 and Pco2 = Ppv2 . (1) sinusoidally with an amplitude d1m and d2m , wherein
The power injected to the grid averaged over a half-cycle, Pg Vco1m Vco2m
d1m = and d2m = (13)
can be written as Vpv1 Vpv2
Pg = Ppv1 + Ppv2 . (2) while during boost mode the duty ratios, d3 of S3 and d4 of S4
vary sinusoidally with amplitude d3m and d4m , wherein
Further, at any half-cycle
Vpv1 Vpv2
vg = vco1 + vco2 . (3) d3m = 1 − and d4m = 1 − . (14)
Vco1m Vco2m
5564 IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS, VOL. 65, NO. 7, JULY 2018

The CONV1 and CONV2 have the same output current ig .

Hence, the input side currents before getting filtered by in-
put filter capacitors of CONV1 , isw1 and CONV2 , isw2 can be
related with ig in the buck mode by considering the switching
cycle average of corresponding quantities as follows:
⟨isw1 ⟩T s = ⟨d1 ⟩T s ⟨ig ⟩T s (15)
⟨isw2 ⟩T s = ⟨d2 ⟩T s ⟨ig ⟩T s . (16)
Similarly by considering switching cycle average of correspond-
ing quantities the relation between isw1 , isw2 , and ig can be
deduced during boost mode as
" #
1
⟨isw1 ⟩T s = ⟨ig ⟩T s (17)
1 − d3 T s
" #
1
⟨isw2 ⟩T s = ⟨ig ⟩T s . (18)
1 − d4 T s
Fig. 4. Equivalent circuit in buck mode (a) S 1 , S 2 are ON, (b) S 1 , S 2
are OFF in boost mode, (c) S 3 , S 4 are ON, and (d) S 3 , S 4 are OFF.
Therefore, it can be inferred from (12) and (13) that if the
insolation level of PV1 is lower than that of PV2 , during
buck mode, d1m < d2m , thereby ⟨d1 ⟩T s < ⟨d2 ⟩T s whereas dur- filter capacitors is neglected. The values of the system param-
ing boost mode as per (12) and (14), d3m < d4m , thereby eters are considered to be as follows: RL 1 = RL 2 = 0.12 Ω,
⟨d3 ⟩T s < ⟨d4 ⟩T s . Hence, it can be concluded from (15)–(17), Rg = 0.04 Ω, Rco1 = Rco2 = 0.26 Ω, Vpv1 = Vpv2 = 130 V.
and (18) that in any operating mode, ⟨isw1 ⟩T s < ⟨isw2 ⟩T s ; there- Considering symmetry in operation of CONV1 and CONV2 ,
fore, Ipv1 < Ipv2 . Following the same argument, Ipv1 > Ipv2 if and by applying state-space averaging technique to (19)–(21),
the insolation level of PV1 is higher than that of PV2 . the simplified transfer functions of ig (s)/d(s), iL 1 (s)/d(s), and
Considering Fig. 1 it can be noted that during operation vco1 (s)/d(s) in s-domain during buck mode are obtained as
in PHC, vcpv1 = vco2 + Vpv1 , vcpv2 = vco2 − Vpv2 while dur- ig (s) 2.87 × 108 s + 2.2 × 1014
ing NHC vcpv1 = −vco1 + Vpv1 , vcpv2 = −vco1 − Vpv2 , wherein = 3
d(s) s + 2267 × s2 + 1.33 × 109 s + 3 × 1011
vcpv1 and vcpv2 are the voltages impressed across Cpv1 and Cpv2 ,
respectively. Hence, the voltages across Cpv1 and Cpv2 contain (22)
significant amount of dc and low-frequency components, which iL 1 (s) 5 2
2.17 × 10 s + 3.52 × 10 s + 2.2 × 108 14
also ensures that the magnitude of the leakage current is main- = 3
d(s) s + 2267 × s2 + 1.33 × 109 s + 3 × 1011
tained within the limit specified in the standard VDE 0126-1-1,
and also cited in [23]. (23)
10 13
vco1 (s) 4.33 × 10 s + 1.3 × 10
= 3
III. MATHEMATICAL MODEL OF THE PROPOSED SCHEME d(s) s + 2267 × s2 + 1.33 × 109 s + 3 × 1011
A small signal modeling of the proposed inverter has been (24)
carried out for buck mode and boost mode of operation. Fig. 4(a)
and (b) represent the equivalent circuit of the proposed inverter where in d is the duty ratio of component converters. Similarly,
while it operates in buck mode, whereas Fig. 4(c) and (d) rep- the simplified transfer functions of ig (s)/d(s), iL 1 (s)/d(s), and
resent the equivalent circuit of the converter while it operates vco1 (s)/d(s) in s-domain during boost mode are obtained as
in boost mode. RL 1 , RL 2 , Rg , Rco1 , and Rco2 are the parasitic ig (s) −9487 s2 − 9.7 × 109 s + 2.1 × 1014
resistances of L1 , L2 , Lg , Co1 , and Co2 , respectively. As indi- = 3
d(s) s + 2018 × s2 + 1.26 × 109 s + 2.7 × 1011
rect grid current control method [21] is adopted to control ig ,
the quantities iL 1 , iL 2 , vco1 , vco2 , and ig are considered to be (25)
5 2 9 14
the state variables. The state equations representing the buck iL 1 (s) 2.4 × 10 s + 3.3 × 10 s + 2.4 × 10
mode of operation of the inverter are derived to be (19) and = 3
d(s) s + 2018 × s2 + 1.26 × 109 s + 2.7 × 1011
(20), shown at the bottom of the next page, by considering the
(26)
equivalent circuits of Fig. 4(a) and (b), while the state equations
6 2 10 12
for boost mode are derived to be (21) and (19), also shown at the vco1 (s) −2 × 10 s + 4.1 × 10 s + 4 × 10
= 3 .
bottom of the next page, by considering the equivalent circuits d(s) s + 2018 × s2 + 1.26 × 109 s + 2.7 × 1011
of Fig. 4(c) and (d).
(27)
The state-space averaging based technique is adopted as the
grid frequency, fg is adequately lower than the switching fre- Due to the existence of symmetry, transfer functions for
quency, fs . In order to simplify the analysis, Vpv1 , Vpv2 , and vg iL 2 (s)/d(s) and vco2 (s)/d(s) remain the same as that of (23),
are considered as stiff voltage sources, and the effect of input (24) in buck mode and (26) and (27) in boost mode, respec-
DUTTA AND CHATTERJEE: BUCK AND BOOST BASED GRID CONNECTED PV INVERTER MAXIMIZING POWER YIELD FROM TWO PV ARRAYS 5565

tively. Based on the derived transfer functions of the system,

compensators are designed to achieve the phase margin of 90◦
for both the plants, and at the same time to maintain the desired
total harmonic distortion (THD) for the grid current.

IV. CONTROL STRATEGY OF THE PROPOSED SCHEME

The control strategy of the proposed scheme is depicted in
Fig. 5. The controller is designed to fulfill the following objec-
tives.
1) Both subarrays operate at their corresponding MPP si-
multaneously.
2) Sensing of output voltages vco1 and vco2 is not required.
Fig. 5. Control configuration of the proposed inverter.
3) ig is sinusoidal and is in-phase with vg throughout the
operating range.
Two separate MPP trackers and two proportional integral
(PI) controllers are employed to determine the value of Ppv1 and fied version of a unity sinusoidal function, R is generated from
Ppv2 , which are required to estimate Vco1m and Vco2m . Using a unity sinusoidal function, X, synchronized with vg , and is
(12), Vco1m and Vco2m are determined where the information obtained from the same PLL. R is multiplied with Vco1m and
of Vm is obtained from the phase locked loop (PLL). A recti- Vco2m to estimate vco1 and vco2 . Hence, two voltage sensors

⎡ ⎤
RL 1 + Rco1 1 Rco1 ⎡ 1 ⎤
⎡ ⎤ − 0 − 0
i̇L 1 (t) ⎢ L1 L1 L1 ⎥⎡ ⎤ 0 0
⎢ ⎥ iL 1 (t) ⎢ L1 ⎥
⎢ ⎥ ⎢ RL 2 + Rco2 1 Rco2 ⎥ ⎢ ⎥⎡ ⎤
⎢ ⎥ ⎢ 0 − 0 − ⎥⎢ ⎥ ⎢ 1 ⎥ vpv1 (t)
⎢ i̇L 2 (t) ⎥ ⎢ L2 L2 L2 ⎥⎢ iL 2 (t) ⎥ ⎢ 0 0 ⎥
⎢ ⎥ ⎢ 1 1 ⎥⎢ ⎥ ⎢ L2 ⎥⎢ ⎥
⎢ ⎥ ⎢ 0 0 0 ⎥⎢ ⎥ ⎢ ⎥⎢ ⎥
⎢ v̇co1 (t) ⎥ = ⎢ − ⎥⎢ v (t) ⎥ + ⎢ 0 0 0 ⎥⎢ vpv2 (t) ⎥
⎢ ⎥ ⎢ Co1 Co1 ⎥⎢ co1 ⎥ ⎢ ⎥⎣ ⎦
⎢ ⎥ ⎢ 1 1 ⎥⎢ ⎥ ⎢ ⎥
⎢ v̇co2 (t) ⎥ ⎢ 0 0 0 − ⎥⎣ vco2 (t) ⎦ ⎢ 0 0 0 ⎥ vg (t)
⎣ ⎦ ⎢ C C ⎥ ⎢ ⎥
⎢ o2 o2 ⎥ i (t) ⎣ 1 ⎦
i̇g (t) ⎣ Rco1 Rco2 1 1 Rco1 + Rco2 + Rg ⎦ g 0 0 −
− Lg
Lg Lg Lg Lg Lg
(19)
⎡ ⎤
RL 1 + Rco1 1 Rco1
⎡ ⎤ − 0 − 0 ⎤ ⎡0 0 0 ⎤
⎢ L1 L1 L1 ⎥⎡
i̇L 1 (t) ⎢ ⎥ iL 1 (t)
⎢ ⎥ ⎢ RL 2 + Rco2 1 Rco2 ⎥⎢ ⎥ ⎢ ⎥
⎢ ⎥ ⎢ 0 − 0 − ⎥⎢
⎥⎢ iL 2 (t) ⎥
⎢0 0 0 ⎥⎡ ⎤
⎢ i̇L 2 (t) ⎥ ⎢ L2 L2 L2 ⎥ ⎢ ⎥ vpv1 (t)
⎢ ⎥ ⎢ ⎥⎢ ⎥ ⎢ ⎥
⎢ ⎥ ⎢ 1 1 ⎥⎢ ⎥ ⎢⎢0 0 0
⎥⎢ ⎥
⎢ v̇co1 (t) ⎥ = ⎢ 0 0 0 ⎥⎢
⎥⎢ vco1 (t) ⎥
⎥⎢ ⎥
⎥⎢ vpv2 (t) ⎥
⎢ ⎥ ⎢ − ⎥+ ⎢
⎢ ⎥ ⎢ Co1 Co1 ⎥⎢ ⎥ ⎢ ⎥⎣ ⎦
⎢ ⎥ ⎢ 1 1
⎥⎢
⎥⎢ vco2 (t) ⎥
⎢0 0 0 ⎥
⎢ v̇co2 (t) ⎥ ⎢ 0 0 0 ⎥ ⎢ ⎥ vg (t)
⎣ ⎦ ⎢⎢ Co2
−
Co2
⎥⎣
⎥ ⎦ ⎢⎣
⎥
⎦
⎢ ⎥ 1
i̇g (t) ⎣ Rco1 Rco2 1 1 Rco1 + Rco2 + Rg ⎦ ig (t) 00−
− Lg
Lg Lg Lg Lg Lg
(20)
⎡ ⎤
RL 1 ⎡ 1 ⎤
⎡ ⎤ − 0 0 0 0
⎢ L1 ⎥⎡ ⎤ 0 0
i̇L 1 (t) ⎢ ⎥ iL 1 (t) ⎢ L1 ⎥
⎢ ⎥ ⎢ RL 2 ⎥⎢ ⎥ ⎢ 1
⎥
⎢ ⎥ ⎢ 0 − 0 0 0 ⎥⎢
⎥⎢ iL 2 (t) ⎥
⎢
0 0
⎥⎡ ⎤
⎢ i̇L 2 (t) ⎥ ⎢ L2 ⎥ ⎢ ⎥ vpv1 (t)
⎢ ⎥ ⎢ ⎥⎢ ⎥ ⎢ L2 ⎥
⎢ ⎥ ⎢ 1 ⎥⎢ ⎥ ⎢ ⎥⎢ ⎥
⎢ v̇co1 (t) ⎥ = ⎢ 0 0 0 0 ⎥⎢
⎥⎢ vco1 (t) ⎥
⎢ 0 0 0 ⎥⎢ ⎥
⎥⎢ vpv2 (t) ⎥.
⎢ ⎥ ⎢ − ⎥+ ⎢ (21)
⎢ ⎥ ⎢ Co1 ⎥⎢ ⎥ ⎢ ⎥⎣ ⎦
⎢ ⎥ ⎢ 1
⎥⎢
⎥⎢ vco2 (t) ⎥
⎢
⎢ 0 0 0
⎥
⎢ v̇co2 (t) ⎥ ⎢ 0 0 0 0 ⎥ ⎥ vg (t)
⎣ ⎦ ⎢⎢
−
Co2
⎥⎣
⎥ ⎦ ⎢⎢
⎥
⎥
⎢ ⎥ ⎣ 1 ⎦
i̇g (t) ⎣ 1 1 Rco1 + Rco2 + Rg ⎦ ig (t) 0 0 −
0 0 − Lg
Lg Lg Lg
5566 IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS, VOL. 65, NO. 7, JULY 2018

that otherwise would have been required to determine vco1 and TABLE I
EMPLOYED PARAMETERS/ELEMENTS FOR SIMULATION AND
vco2 get eliminated. Vpv1 and vco1 are compared to decide about EXPERIMENTAL PURPOSE
the mode of operation (buck mode or boost mode) of CONV1 ,
while Vpv2 and vco2 are compared to determine the mode of op-
Parameter/elements Value
eration of CONV2 . RMS values of vco1 and vco2 are estimated,
which are then subsequently squared and are then divided by V g and fg 220 V and 50 Hz
L 1 , L 2 , L g , and C o 1 , C o 2 0.6 mH, 0.6 mH, 0.4 mH, and 5 µF, 5 µF
Ppv1 and Ppv2 to obtain the emulated effective resistances, Rpco1 C pv1 and C pv2 0.1 µF
and Rpco2 of the two component converters. Subsequently the MPPT Algorithm Incremental Conductance
reference current, iL 1ref of L1 , and the reference current, iL 2ref MOSFETS (S 1 –S 8 ) IPW60R041C6
Diodes (D f 1 –D f 4 ) MBR40250
of L2 , are synthesized by utilizing (28) in the buck mode [21] fs of S 1 –S 4 and fs of S 5 –S 8 50 kHz and 50 Hz
vco1 vco2 Digital signal controller TMS320F28335
iL 1ref = and iL 2ref = (28)
Rpco1 Rpco2
while for boost mode (29), it is used to generate iL 1ref and TABLE II
iL 2ref [21]. ESTIMATED VARIATIONS OF DIFFERENT QUANTITIES DURING APPLIED
VARIATIONS ON INSOLATION AND TEMPERATURE OF TWO SUBARRAYS
2 2
vco1 vco2
iL 1ref = and iL 2ref = . (29)
Rpco1 Vpv1 Rpco2 Vpv2 Time in Second 0–1 1–2 2–3 3–4 4–5 5–6 6–7 7–8
The sensed inductor currents, iL 1 and iL 2 are compared with Insol. in PV1 (kW/m2 ) 0.5 0.6 0.7 0.8 0.9 1.0 1.0 1.0
their corresponding references iL 1ref and iL 2ref . The errors so Insol. in PV2 (kW/m2 ) 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8
obtained are processed through two separate PI controllers to Temp. in PV1 (◦ C) 25 25 25 25 25 25 30 35
Temp. in PV2 (◦ C) 25 25 25 25 25 25 25 25
generate the required sinusoidal duty ratios for the switches, P pv1 (W) 331 397 463 529 595 661 638 621
S1 and S2 during buck mode. Similarly, two separate PI con- P pv2 (W) 529 529 529 529 529 529 529 529
trollers are engaged to process the generated errors to synthesize Ig m (A) 5.5 6.0 6.4 6.8 7.2 7.7 7.5 7.4
V co1 m (V) 120 133 147 155 165 173 170 168
required sinusoidal duty ratios for switches S3 and S4 during V co2 m (V) 191 178 164 156 146 138 141 143
boost mode. Signal Y is used to generate gating signals for S5 , IL 1 m (A) 5.7 7 8.1 9 10.3 11.4 11 10.7
S8 while signal Z is used to generate gating signals for S6 , S7 IL 2 m (A) 9 9 9 9 9 9 9 9
of the grid frequency unfolding inverter.

V. SELECTION OF L1 , L2 , Lg , AND Co1 , Co2 follows: Vpv1 = Vpv2 = 116 V, Ipv1 = Ipv2 = 5.7 A, and Ppv1
In order to select the value of the filter elements, L1 , L2 , Lg , = Ppv2 = 661 W. The parameters that are used to simulate the
and Co1 , Co2 the design principle given in [24] is followed and proposed inverter are indicated in Table I. MATLAB-Simulink
the buck mode of operation for the inverter is considered. Values platform is utilized to simulate the performance of the proposed
of L1 and L2 are obtained from the expression given in [24] inverter.
The variation in insolation level and temperature with respect
Vpv1 Vpv2 to time, which is considered for the two subarrays to demon-
L1 = and L2 = (30)
4∆IL 1 fs 4∆IL 2 fs strate the effectiveness of the proposed inverter are tabulated in
where in Vpv1 = Vpv2 = 200 V, and percentage peak to peak Table II. Estimated variation of Ppv1 , Ppv2 along with the other
ripple of iL 1 and iL 2 , ∆IL 1 and ∆IL 2 are considered as 15% of parameters Ig m , Vco1m , Vco2m , peak of iL 1 (IL 1m ) and peak of
rated peak current. iL 2 (IL 2m ) are also indicated in the same table. Fig. 6(a)–(c)
The values of Co1 and Co2 are obtained from the expression represents the variation of Ppv1 , Ppv2 , Vpv1 , Vpv2 , Ipv1 , Ipv2 of
given in [24] the two subarrays and also demonstrate the ability of the pro-
posed inverter to operate the two subarrays simultaneously at
xPco1 xPco2
Co1 = 2 and Co2 = 2 (31) their respective MPP. Variation in ig , iL 1 , iL 2 , vco1 , and vco2
2πfg Vco1 2πfg Vco2 along with their magnified versions for two different insolation
where in Vco1 = Vco2 = 110 V, Pco1 = Pco2 = 750 W, and factor levels are depicted in Figs. 7–9. The estimated values of the
x = 2.5%. aforementioned quantities as tabulated in Table II conform to
In order to achieve wide stability margin and large control that of obtained through simulation studies, thereby ensuring
bandwidth a value, which is less than L1 or L2 is selected for the viability of the proposed scheme.
Lg [24].
VII. EXPERIMENTAL VERIFICATION
VI. SIMULATION STUDY A 1.5 kW laboratory prototype of the proposed inverter is
To demonstrate the efficacy of the proposed inverter a PV fabricated and detailed experimental studies have been carried
array consisting of two PV subarrays while each of the subar- out to demonstrate the effectiveness of the proposed scheme.
ray having four series connected Canadian solar polycrystalline The parameters as mentioned in Table I are used to realize the
modules “CS6P-165PE” [25] is considered. The MPP param- laboratory prototype of the inverter. In order to realize PV1 and
eters of each subarray at standard test condition (STC) are as PV2 two programmable EPS PSI9360-15 power supplies having
DUTTA AND CHATTERJEE: BUCK AND BOOST BASED GRID CONNECTED PV INVERTER MAXIMIZING POWER YIELD FROM TWO PV ARRAYS 5567

Fig. 8. Simulated waveform. iL 1 and iL 2 and their magnified views.

Fig. 6. Simulated waveform. Variation in (a) p pv1 and p pv2 , (b) v pv1 and
v pv2 , and (c) ipv1 and ipv2 during entire range of operation.

Fig. 9. Simulated waveform. v co1 and v co2 and their magnified views.

Fig. 7. Simulated waveform. v g and ig and their magnified views.

solar PV emulation feature are utilized. The photograph of the

experimental prototype is shown in Fig. 10.
The EPS PSI9306-15 power supply has the provision to Fig. 10. Experimental prototype of the proposed inverter.
change only the effect of insolation level while the option to
change the effect of temperature is unavailable. In order to em- TABLE III
ulate simultaneous variation in temperature and level of inso- ESTIMATED VARIATION IN IPV1 , IPV2 , P PV1 , P PV2 , V CO1 m , V CO2 m , Ig m ,
lation, the MPP parameters of the two solar emulators (solar IL 1 m , IL 2 m DURING PV1 INSOLATION VARIATION
emulator 1 as PV1 and solar emulator 2 as PV2 ) are set as fol-
lows at STC: Vpv1 = 130 V, Ipv1 = 5 A, and Vpv2 = 120 V, Ipv2 % Insol. of PV1 40 50 60 70 80 90 100
% Insol. of PV2 80 80 80 80 80 80 80
= 5 A. The variation in insolation level of PV1 is indicated in
Table III while the insolation level of PV2 is maintained at 80%. Ipv1 (A) 2 2.5 3 3.5 4 4.5 5
The expected values of Ipv1 , Ipv2 , Ppv1 , Ppv2 , Vco1m , Vco2m , Ig m , Ipv2 (A) 4 4 4 4 4 4 4
P pv1 (W) 260 325 390 455 520 585 650
IL 1m , and IL 2m for the entire operating range are tabulated in P pv2 (W) 480 480 480 480 480 480 480
the Table III. Fig. 11 depicts the change in ig , Ipv1 , Ipv2 , Ppv1 , V co1 m (V) 109 126 140 151 162 171 179
and Ppv2 throughout the range of variation in the level of insola- V co2 m (V) 202 185 171 160 149 140 132
Ig m (A) 4.6 5 5.4 5.8 6.2 6.6 7
tion as specified in Table III. Magnified version of the responses IL 1 m (A) 4.6 5 5.8 6.8 7.8 8.7 10
of vco1 , vco2 , iL 1 , and iL 2 , along with vg , ig , are also shown IL 2 m (A) 7.7 7.7 7.7 7.7 7.7 7.7 7.7
in Fig. 12(a)–(f) for two different insolation levels of PV1 . The
5568 IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS, VOL. 65, NO. 7, JULY 2018

Fig. 13. Experimental waveform. FFT of ig .

Fig. 11. Experimental waveforms. v pv1 , v pv2 , ig , v g , ipv1 , ipv2 , p pv1 ,

and p pv2 throughout the entire operating range.

Fig. 14. Efficiency curves of the proposed inverter.

depending on the requirement. Thus, it can be inferred that

the two converter segments are able to operate in a decoupled
fashion. The measured variables, Ipv1 , Ipv2 , Ppv1 , Ppv2 , Vco1m ,
Vco2m , Ig m , IL 1m , and IL 2m , as depicted in Figs. 11 and 12,
are more or less same as that of the estimated ones presented in
Table III, and this validates the ability of the proposed inverter
to extract maximum power from two subarrays operating under
MEC.
Fig. 13 depicts the fast Fourier transform (FFT) of ig . The
THD of ig is found to be 4.61%, which is below the limit of
5%, as specified in the standards, IEEE 1574/IEC 61727 [22].
It may be noted that the measured THD of vg is found to be
Fig. 12. Experimental waveforms. Magnified version of ig , v g when
(a) insolation of PV1 = 40% and insolation of PV2 = 80%, (b) insolation 2.12%, and hence, the contribution to THD from the inverter is
of PV1 = 100% and insolation of PV2 = 80%, magnified version of much less than 4.61%.
v pv1 , v pv2 , v co1 , v co2 when (c) insolation of PV1 = 40% and insolation The measured and estimated efficiency curves of the proposed
of PV2 = 80%, (d) insolation of PV1 = 100% and insolation of PV2 =
80%, magnified version of iL 1 , iL 2 when (e) insolation of PV1 = 40% inverter are shown Fig. 14. In order to measure the efficiency
and insolation of PV2 = 80%, and (f) insolation of PV1 = 100% and of the proposed inverter the Yokogawa make power analyzer,
insolation of PV2 = 80%. WT1800 is used and further, the losses incurred in the active
and passive elements of the power circuit is considered while
the losses involved with the control circuit are neglected. The
figures, Fig. 12(a) and (b), ensure that ig remains to be sinusoidal efficiency is determined while both Vpv1 and Vpv2 are set at
and in-phase with vg in spite of having difference in the mag- 130 V. Measured peak efficiency is found to be 97.65% and the
nitude of power being extracted from the two subarrays. From measured European efficiency (ηeuro ) is obtained as 97.02%.
Fig. 12(c), it can be inferred that the converter associated with In order to measure the leakage current involved with the
PV1 operates completely in buck mode, whereas the converter proposed inverter, 0.1 µF polypropylene film capacitors are used
associated with PV2 operates in both buck and boost mode, to emulate Cpv1 and Cpv2 . Fig. 15 depicts the voltages that appear
DUTTA AND CHATTERJEE: BUCK AND BOOST BASED GRID CONNECTED PV INVERTER MAXIMIZING POWER YIELD FROM TWO PV ARRAYS 5569

TABLE IV
COMPARISON TABLE OF VARIOUS TRANSFORMERLESS SCHEMES

Schemes N PVR and V IN N MS and P SYS A PV E MEC

N PVC (V) N MT (kW) (m 2 )

NPC based [5] 1 and 1 > 2V m 28 and 28 4.6 44.8 Highest

H-Bridge based [2] 1 and 1 > Vm 14 and 14 2.3 22.4 High
Reported in [18] 2 and 2 < Vm 8 and 16 2.6 25.6 Low
Reported in [21] 1 and 1 < Vm 8 and 8 1.3 12.8 Low
Proposed DBBI 2 and 2 < Vm 4 and 8 1.3 12.8 Lowest

TABLE V
EFFECT OF MEC IN DIFFERENT TRANSFORMERLESS SCHEMES
Fig. 15. Experimental waveform. v g , v c pv1 , v c pv2 , ic pv1 , ic pv2 .
Mod1 in (%) 100 90 80 70 60 50
Schemes P avl (kW) 5.3 5.27 5.25 5.24 5.22 5.2

P ext (kW) 5.3 5.2 4.9 4.53 4.1 3.6

NPC based [5] P diff (kW) 0 0.07 0.35 0.71 1.12 1.6
P lost (%) 0 1.3 6.7 13.5 21.7 31.5
P ext (kW) 5.3 5.23 5.04 4.8 4.54 4.25
H-Bridge based [2] P diff (kW) 0 0.04 0.21 0.44 0.68 0.95
P lost (%) 0 0.8 4 8.2 13.1 18.3
P ext (kW) 5.3 5.25 5.15 5.03 4.90 4.75
Reported in [18] P diff (kW) 0 0.02 0.1 0.21 0.32 0.45
P lost (%) 0 0.4 2 4 6.1 8.6
P ext (kW) 5.3 5.25 5.15 5.03 4.90 4.75
Reported in [21] P diff (kW) 0 0.02 0.1 0.21 0.32 0.45
P lost (%) 0 0.4 2 4 6.1 8.6
P ext (kW) 5.3 5.26 5.21 5.14 5.08 5.01
Proposed DBBI P diff (kW) 0 0.01 0.04 0.1 0.14 0.19
P lost (%) 0 0.2 0.8 2 2.7 3.6

[21] has been performed and presented in Table IV. Following

issues are considered for carrying out this comparison.
Fig. 16. Experimental waveform. v g , ig , v co1 , v co2 , v pv1 , v pv2 when v g 1) Solar modules, i.e., Canadian solar “CS6P-165PE” [25],
is changed from 150 V to 220 V.
are utilized for the purpose.
2) Minimum input voltage requirement for NPC-based
scheme [5] and H-Bridge-based scheme [2] is taken to
across Cpv1 and Cpv2 , and the leakage currents, icpv1 and icpv2 be 800 V and 400 V, respectively, while minimum in-
flowing through Cpv1 and Cpv2 . The measured waveforms of put voltage requirement of the schemes presented in [18]
vcpv1 and vcpv2 show that they contain significant amount of and [21] and that of the proposed scheme is taken to be
dc and low-frequency components whereas presence of high- 230 V.
frequency components are negligible. The measured rms value 3) For simplicity, total area required for a system is de-
of total leakage current is found to be 80.7 mA, which is much termined by multiplying the total number of modules
lower than the limit 300 mA as specified in the standard, VDE required with the area of a single module.
0126-1-1, and also cited in [23]. The nomenclature used in Table IV is defined as follows:
In order to demonstrate the stability of the proposed scheme NPVR = required number of PV arrays/subarrays, NPVC = num-
in the event of a disturbance in vg , a step change of 70 V (150– ber of PV arrays controlled simultaneously, VIN = input voltage
220 V) is introduced in vg while Vpv1 and Vpv2 are kept at 130 requirement, NMS = number of modules connected in series in
V, and the references for the current controllers are purposely a PV string of a PV array/subarray, which is made with a single
set at a fixed value in each mode. The response of the system string, NMT = minimum number of modules required to design
during the aforesaid test condition where mode of operation of the PV system, PSYS = minimum power rating of the PV system,
the proposed inverter is shifted from buck mode to buck and APV = minimum area required to install all PV modules, EMEC
boost mode is shown in Fig. 16. It can be inferred from the = possibility to get affected by MEC, which is determined from
Fig. 16 that the system can effectively ride through situations Table V. Based on the objective comparison presented in the
arising due to the disturbances in vg . Table IV it can be inferred that the proposed inverter deals with
A comparison of various features of the proposed scheme with MEC in the most effective way.
existing transformerless schemes such as NPC-based scheme In order to compare the power extraction from PV array by
[5], H-Bridge-based scheme [2], schemes presented in [18] and various transformerless schemes as mentioned in Table IV while
5570 IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS, VOL. 65, NO. 7, JULY 2018

the schemes are operating under MEC, a 5.3 kW PV system at [3] B. Ji, J. Wang, and J. Zhao, “High-efficiency single-phase transformer-
STC, built with 32 “CS6P-165PE” Canadian solar modules [25] less PV H6 inverter with hybrid modulation method,” IEEE Trans. Ind.
Electron., vol. 60, no. 5, pp. 2104–2115, May 2013.
is considered. Depending on the minimum dc voltage require- [4] R. Gonzalez, E. Gubia, J. Lopez, and L. Marroyo, “Transformerless single
ment, type of input connection required and power rating of phase multilevel-based photovoltaic inverter,” IEEE Trans. Ind. Electron.,
the schemes, required number of PV modules are connected in vol. 55, no. 7, pp. 2694–2702, Jul. 2008.
[5] H. Xiao and S. Xie, “Transformerless split-inductor neutral point clamped
series-parallel combination to form PV array/subarrays. The PV three-level PV grid-connected inverter,” IEEE Trans. Power Electron.,
arrays/subarrays of the schemes are configured as follows: vol. 27, no. 4, pp. 1799–1808, Apr. 2012.
1) NMS = 32 in [5]; [6] A. Bidram, A. Davoudi, and R. S. Balog, “Control and circuit techniques to
mitigate partial shading effects in photo voltaic arrays,” IEEE J. Photovolt.,
2) NMS = 16 and NMP = 2 in [2], wherein NMP = number vol. 2, no. 4, pp. 532–546, Oct. 2012.
of strings connected in parallel in an array/subarray; [7] N. D. Kaushika and N. K. Gautam, “Energy yield simulations of inter-
3) NMS = 8 and NMP = 2 in [18]; connected solar PV arrays,” IEEE Trans. Energy Convers., vol. 18, no. 1,
pp. 127–134, Mar. 2003.
4) NMS = 8 and NMP = 4 in [21]; [8] H. Patel and V. Agarwal, “Maximum power point tracking scheme for PV
5) NMS = 4 and NMP = 4 in DBBI. systems operating under partially shaded conditions,” IEEE Trans. Ind.
Further, it is also assumed that no parallel diode is connected Electron., vol. 55, no. 4, pp. 1689–1698, Apr. 2008.
[9] D. Nguyen and B. Lehman, “An adaptive solar photovoltaic array us-
across PV modules. The insolation level of one module, Mod1in ing model-based reconfiguration algorithm,” IEEE Trans. Ind. Electron.,
is varied from 100% to 50% with a step of 10% while the vol. 55, no. 7, pp. 2644–2654, Jul. 2008.
insolation of rest of the 31 modules are kept at 100%, i.e., at [10] G. Velasco-Quesada, F. Guinjoan-Gispert, R. Pique-Lopez, M. Roman-
Lumbreras, and A. Conesa-Roca, “Electrical PV array reconfiguration
STC. The total estimated extracted power from the PV subarrays strategy for energy extraction improvement in grid-connected PV sys-
during MEC by any scheme, Pext = power of affected string + tems,” IEEE Trans. Ind. Electron., vol. 56, no. 11, pp. 4319–4331,
power of rest of the string, the actual available maximum power Nov. 2009.
[11] L. F. L. Villa, T.-P. Ho, J.-C. Crebier, and B. Raison, “A power electronics
in the PV array of any scheme, Pavl = 31× power of each equalizer application for partially shaded photovoltaic modules,” IEEE
module + power of affected module, their difference, Pdiff and Trans. Ind. Electron., vol. 60, no. 3, pp. 1179–1190, Mar. 2013.
the percentage loss of power due to MEC in any scheme, Plost [12] P. Sharma and V. Agarwal, “Maximum power extraction from a partially
shaded PV array using shunt-series compensation,” IEEE J. Photovolt.,
are tabulated in Table V. As the effect of MEC is less severe vol. 4, no. 4, pp. 1128–1137, Jul. 2014.
in parallel connected PV strings, the aforementioned effect is [13] N. Femia, G. Lisi, G. Petrone, G. Spagnuolo, and M. Vitelli, “Dis-
neglected to avoid complexity in calculation. From Table V it tributed maximum power point tracking of photovoltaic arrays: Novel
approach and system analysis,” IEEE Trans. Ind. Electron., vol. 55, no. 7,
can be concluded that the proposed inverter is the most effective pp. 2610–2621, Jul. 2008.
solution in extracting power during MEC. [14] C. Olalla, C. Deline, D. Clement, Y. Levron, M. Rodriguez, and D. Mak-
simovic, “Performance of power-limited differential power processing
architectures in mismatched PV systems,” IEEE Trans. Power Electron.,
VIII. CONCLUSION vol. 30, no. 2, pp. 618–630, Feb. 2015.
[15] E. Karatepe, T. Hiyama, M. Boztepe, and M. Çolak, “Voltage based
A single phase GCT buck and boost based PV inverter that can power compensation system for photo voltaic generation system under
operate two subarrays at their respective MPP was proposed in partially shaded insolation conditions,” Energy Convers. Manage., vol. 49,
this paper. The attractive features of this inverter were as follows. pp. 2307–2316, Aug. 2008.
[16] A. A. Elserougi, M. S. Diab, A. M. Massoud, A. S. Abdel-Khalik, and
1) The effect of MECs on the PV array could be dealt with S. Ahmed, “A switched PV approach for extracted maximum power en-
in an effective way. hancement of PV arrays during partial shading,” IEEE Trans. Sustain.
2) The operating efficiency achieved, ηeuro = 97.02% was Energy, vol. 6, no. 3, pp. 767–772, Jul. 2015.
[17] I. Patrao, G. Garcera, E. Figueres, and R. Gonzalez-Medina, “Grid-
high. tie inverter topology with maximum power extraction from two photo-
3) Decoupled control of component converters was possible. voltaic arrays,” IET Renewable Power Gener., vol. 8, no. 6, pp. 638–648,
4) A simple MPPT algorithm was employed to ensure MPP Aug. 2014.
[18] D. Debnath and K. Chatterjee, “Maximising power yield in a trans-
operation for the component converters. formerless single phase grid connected inverter servicing two sepa-
5) Leakage current associated with the PV arrays was within rate photovoltaic panels,” IET Renewable Power Gener., vol. 10, no. 8,
the limit mentioned in VDE 0126-1-1. pp. 1087–1095, 2016.
[19] N. A. Ahmed, H. W. Lee, and M. Nakaoka, “Dual-mode time-sharing
Mathematical analysis of the proposed inverter leading to sine wave-modulation soft switching boost full-bridge one-stage power
the development of its small signal model was carried out. The conditioner without electrolytic capacitor DC link,” IEEE Trans. Ind.
criterion to select the values of the output filter components was Appl., vol. 43, no. 3, pp. 805–813, May/Jun. 2007.
[20] Z. Zhao, M. Xu, Q. Chen, J. S. Lai, and Y. Cho, “Derivation, analysis,
presented. The scheme was validated by carrying out detailed and implementation of a Boost–Buck converter-based high-efficiency PV
simulation studies, and subsequently, the viability of the scheme inverter,” IEEE Trans. Power Electron., vol. 27, no. 3, pp. 1304–1313,
was ascertained by carrying out thorough experimental studies Mar. 2012.
[21] W. Wu, J. Ji, and F. Blaabjerg, “Aalborg inverter a new type of buck
on a 1.5 kW prototype of the inverter fabricated for the purpose. in buck, boost in boost grid-tied inverter,” IEEE Trans. Power Electron.,
vol. 30, no. 9, pp. 4784–4793, Sep. 2015.
REFERENCES [22] R. Teodorescu, M. Liserre, and P. Rodriguez, Grid Converters for Photo-
voltaic and Wind Power Systems. Hoboken, NJ, USA: Wiley, 2011.
[1] T. Shimizu, O. Hashimoto, and G. Kimura, “A novel high-performance [23] W. Li, Y. Gu, H. Luo, W. Cui, X. He, and C. Xia, “Topology review
utility-interactive photovoltaic inverter system,” IEEE Trans. Power Elec- and derivation methodology of single phase transformerless photovoltaic
tron., vol. 18, no. 2, pp. 704–711, Mar. 2003. inverters for leakage current suppression,” IEEE Trans. Ind. Electron.,
[2] S. V. Araujo, P. Zacharias, and R. Mallwitz, “Highly efficient single-phase vol. 62, no. 7, pp. 4537–4551, Jul. 2015.
transformerless inverters for grid-connected photovoltaic systems,” IEEE
Trans. Ind. Electron., vol. 57, no. 9, pp. 3118–3128, Sep. 2010.
DUTTA AND CHATTERJEE: BUCK AND BOOST BASED GRID CONNECTED PV INVERTER MAXIMIZING POWER YIELD FROM TWO PV ARRAYS 5571

[24] W. Wu, Y. He, and F. Blaabjerg, “An LLCL power filter for single- Kishore Chatterjee (M’10) was born in Cal-
phase grid-tied inverter,” IEEE Trans. Power Electron., vol. 27, no. 2, cutta, India, in 1967. He received the B.E. degree
pp. 782–789, Feb. 2012. from Maulana Azad National Institute of Tech-
[25] Information on Canadian solar module CS6P-165PE. [Online]. Available: nology, Bhopal, India, in 1990; the M.E. degree
www.solarhub.com/product-catalog/pv-modules/124 from Indian Institute of Engineering Science and
Technology, Howrah, India, in 1992, both in elec-
trical engineering; and the Ph.D. degree from
the Indian Institute of Technology (IIT) Kanpur,
Kanpur, India, in 1998.
From 1997 to 1998, he was a Senior Re-
search Associate with the IIT Kanpur. Since
Subhendu Dutta was born in West Bengal, 1998, he has been with the Department of Electrical Engineering, IIT
India. He received the B.Tech. degree in elec- Bombay, Mumbai, India, where he is currently a Professor. He was a
trical engineering from West Bengal University Visiting Fellow with École de Technologie Supérieure, University of Que-
of Technology, Kolkata, India, in 2009 and the bec, Montreal, QC, Canada, in 2004. He has been leading the power
M.E. degree in electrical engineering from Ja- electronic group of the National Centre for Photovoltaic Research and
davpur University, Kolkata, India, in 2012. He is Education, hosted with IIT Bombay, since 2009. His current research in-
currently working toward the Ph.D. degree in the terests are power evacuation strategies from solar photovoltaic systems,
area of power electronics in the Department of modern VAR compensators, active power filters, utility-friendly converter
Electrical Engineering, Indian Institute of Tech- topologies, and induction motor drives.
nology Bombay, Mumbai, India.
He worked as an Assistant Professor in the
College of Engineering and Management, Kolaghat, India, from 2012 to
2013. His current research interests include the shading effect on solar
photo voltaic systems, the design and efficiency improvement of power
electronic converters for solar photovoltaic applications, and the design
of magnetic elements for power electronic systems.