Solar Energy 209 (2020) 42–56

Contents lists available at ScienceDirect

Solar Energy
journal homepage: www.elsevier.com/locate/solener

Three-level quadratic boost DC-DC converter associated to a SRM drive for

water pumping photovoltaic powered systems
Armando Cordeiro a, c, d, *, V. Fernão Pires b, c, d, Daniel Foito b, d, e, A.J. Pires b, d, e, J.F. Martins e
a
Instituto Politécnico de Lisboa, ISEL, ADEEEA, LCEC, Rua Conselheiro Emídio Navarro 1, Lisboa, Portugal
b
Instituto Politécnico de Setúbal, ESTSetúbalInstituto Politécnico de Setúbal, Setúbal, Portugal
c
INESC-ID Lisboa, Portugal
d
Sustain.RD, Instituto Politécnico de Setúbal, Setúbal, Portugal
e
CTS – Uninova, FCT – Universidade Nova de Lisboa, Lisboa, Portugal

A R T I C L E I N F O A B S T R A C T

Keywords: This paper presents a water pumping system powered by solar photovoltaic (PV) panels employing a switched
Solar photovoltaic panels reluctance motor (SRM) drive and a three-level quadratic Boost (3LQB) DC-DC converter for a dual output. The
Three-level quadratic Boost DC-DC converter DC-DC topology is characterized by quadratic voltage static gain, capability to ensure voltage balance in the
Asymmetrical half-bridge
output capacitors and reduced voltage stress across power switches and diodes. The three-level voltage of the DC-
SRM drive
DC converter allows to connect an asymmetrical half-bridge (AHB) converter integrated into the SRM drive. The
Water pumping system
DC-DC converter also operates in continuous conduction mode (CCM) which, combined with a proposed
maximum power point tracking (MPPT) algorithm, helps to optimize the power supplied by the PV panels. A
laboratory prototype was also developed and installed in a workbench to verify its practical implementation
considering all devices involved. The tests were performed considering different loads and solar irradiations to
verify the system performance under different situations. The PV panels were simulated using a remotely
controlled DC power source to impose the typical characteristic I-V curves of the PV panels according to solar
irradiation and the water pump behavior was emulated in the laboratory using a controlled torque load coupled
to the SRM drive. The experimental results indicate that the proposed solution allows to fully exploit the solar
energy to provide as much as possible a continuous water flow in isolated pumping solutions. The operating
limits of this solution should be determined by the minimum water flow rate required, necessary head (elevation)
and net positive suction head (NPSH).

1. Introduction valuable in many ways, especially when applied in remote places where
no electricity supply is available or the investment costs in such in­
Despite the intermittent nature of solar energy, there are many do­ frastructures are substantial (Gopal et al., 2013; Hamrouni et al., 2009).
mestic, commercial or industrial applications (for self-consumption, There are, in many places around the world, pumping systems using
essentially) where PV panels are viable as an alternative energy internal combustion engines (mainly diesel) with the intrinsic disad­
source, from decentralized power generation to supply isolated loads or vantages of fossil fuel prices (including transportation costs), environ­
even connected to the local distribution network. An important appli­ mental pollution (noise, smoke and fuel leaks), high maintenance costs,
cation where solar energy can be used is the water pumping systems for low reliability and low efficiency (Sarkar et al., 2015). To overcome
human consumption, animals or fauna and irrigation channels in agri­ some of these drawbacks and environmental problems, several solutions
culture, among other uses. Since water is one of the most important of water pumping systems powered by PV panels have emerged over the
needs for life in hearth and global sustainability of all countries, the last years (Sharma et al., 2017; Pires et al., 2018a; Koreboina et al.,
huge number of water pumping facilities around the world is under­ 2017; Mishra and Singh, 2016), providing autonomous operating ca­
standable. The use of photovoltaic solar systems for this purpose can be pacity, robustness, reliability and low operating and maintenance costs.

* Corresponding author at: Instituto Superior de Engenharia de Lisboa – ISEL, Electrical Engineering Automation and Energy Department-ADEEEA, Instituto
Politécnico de Lisboa – IPL, Rua Conselheiro Emídio Navarro 1, 1959-007 Lisboa, Portugal.
E-mail address: [email protected] (A. Cordeiro).

https://doi.org/10.1016/j.solener.2020.08.076
Received 17 January 2020; Received in revised form 15 June 2020; Accepted 24 August 2020
Available online 7 September 2020
0038-092X/© 2020 International Solar Energy Society. Published by Elsevier Ltd. All rights reserved.
A. Cordeiro et al. Solar Energy 209 (2020) 42–56

Some reviews about solar photovoltaic water pumping systems and advantage, which is the use of an asymmetrical half-bridge (AHB) con­
modeling of solar energy systems can be found in the literature (Aliyu verter in the SRM drive with half of the power semiconductors and
et al., 2018; Li et al., 2017; Gopal et al., 2013; Elsheikh et al., 2019; without impact in the machine operation. In general, the proposed so­
Houshmand et al., 2013). Typically, the classical solutions for this lution, resulting from the combination of all the described elements,
purpose are driven by direct current (DC) motors (Vijayakumar and allows to improve the performance of water pumping solutions powered
Amrutha, 2016; Gonzalez-Llorente et al., 2010), alternating current by photovoltaic systems in different ways. The validation of the theo­
(AC) motors (Singh and Shukla, 2018; Ebrahim et al., 2010; Chan­ retical concepts presented in this paper will be supported by several
drasekaran and Thyagarajah, 2012) or brushless DC (BLDC) motors simulation and experimental results.
(Kumar and Singh, 2019, 2018; Singh and Kumar, 2016). Such classical Besides this introduction, the proposed solution will be presented in
solutions present some disadvantages related to the type of motor (e.g. Section 2, with the most important details about the different elements
maintenance of DC motor) or the number of power semiconductors that compose the system and control aspects. Section 3 is dedicated to
necessary to drive the motor. In recent years, there has been a growing showing some simulation results of the proposed solution. The experi­
interest in the use of switched reluctance motors (SRM) for several ap­ mental results will be presented in Section 4 and the conclusions in
plications including water pumping systems. This interest is due to as­ Section 5.
pects of robustness, low inertia, low cost, simplicity of operation and
potential to work in hostile environments (Cordeiro et al., 2019; Zhang 2. Proposed solution
et al., 2019; Han et al., 2019). In addition, this type of motor can be
driven by a simple converter powered by a DC power source such as PV The proposed solution combines several systems, namely a solar PV
panels. array, a DC-DC converter with MPPT control algorithm, an asymmetrical
It is well known that power electronic circuits are essential to pro­ half-bridge (AHB) converter and a centrifugal pump powered by a four-
vide an efficient energy conversion and overcome different integration phase 8/6 SRM, as presented in Fig. 1. Due to the proposed DC-DC
problems. Among all power converters, the DC-DC converter is currently converter, the four phases of the 8/6 SRM can have a common star
one of the most important electronic circuits regarding PV applications, winding connection, and consequently the AHB converter is composed
allowing to regulate power and stabilize the output voltage regardless of by four branches only.
the input voltage variation and output load conditions (up to certain
limits). Additionally, for most PV applications, continuous input current
2.1. Solar PV panels
and high voltage gain are highly desirable. Many DC-DC converters have
been proposed for integration of PV applications or other renewable
A solar photovoltaic panel cell is basically composed by a p-n semi­
energy sources into the grid (Fang et al., 2019; Pires et al., 2019, 2018b;
conductor junction which generates a DC current when exposed to
Wu et al., 2014; Das and Agarwal, 2015). Undesirably, most of these
sunlight. Silicon PV cells are composed of a thin layer of bulk Si or a thin
converters have inadequate voltage gains and do not allow to explore
Si film connected to electric terminals. One of the sides of the Si layer is
the full potential of photovoltaic systems, especially for water pumping
doped to form the p-n junction (Asahi and Horikoshi, 2019a; Villalva
applications. Additionally, most solutions propose the classical MPPT
et al., 2009). The study of the solar radiation effect on PV devices is
(maximum power point tracking) controller such as perturb & observe
difficult because the spectrum of the sunlight on the Earth’s surface is
(Kumar et al., 2019) and incremental conductance (Tey and Mekhilef,
influenced by several factors such as the variation of the temperature
2014). These methods are well known and provide reasonably dynamic
and the influence of the atmosphere. On the Earth’s surface, the irra­
and steady tracking, but there is still potential for the development of
diation is approximately 1 kW/m2 (this is a reference value only, as the
new, robust, and simple MPPT solutions.
net irradiation on the Earth’s surface depends on many factors) (Guechi
For all these reasons, new proposals are needed. The proposed
and Chegaar, 2007). The equivalent electrical circuit of an ideal solar
research paper intends to overcome most of the existing limitations of
cell can be treated as a current source in parallel with a diode (Villalva
other solutions. In this way, the proposed solution uses a three-level
et al., 2009) as shown in Fig. 2, where IPV_SC denotes the current due to
quadratic Boost (3LQB) DC-DC converter to supply a four-phase 8/6
the solar irradiation G and module cell temperature T. The series and
SRM drive responsible for driving a water pumping system. Since PV
parallel equivalent resistances are represented by RS and RP,
panels are characterized by a huge voltage variation, as a function of the
respectively.
solar irradiation, usually it is necessary to use DC-DC Boost converters
From the analysis of the electric equivalent circuit, using Kirchhoff’s
with high voltage gain to harness their energy when solar irradiation is
laws, the following relationship between currents can be obtained for an
reduced. The proposed converter provides quadratic voltage gain and
ideal solar cell:
continuous input current which, allied to a new design in the MPPT
control, presented in this paper, allow to minimize the impact of the VD
IPVS C = ID + + IPV (1)
reduced solar irradiation in the water pumping application. The pro­ RP
posed MPPT control algorithm is based on the variation of the power and In this model, the diode represents the p-n junction and its (dark)
voltage with respect to time, resulting in a robust and very easy practical current ID (2) which is dependent on the output voltage VD, where VT (3)
implementation. A class of DC-DC converter topologies with quadratic is the thermal voltage and IO is the saturation current; n is the ideality
characteristics have been used with success in other applications (No­ factor of the diode, μV and μI are the voltage and current temperature
vaes et al., 2007; Amir et al., 2019; Bottarelli et al., 2007; Cabral et al., coefficients, respectively; VOC_STC and ISC_STC are the open-circuit voltage
2013). Thus, their study and implementation for water pumping PV and short-circuit current under standard test conditions (STC),
powered applications can be a very interesting solution. The high respectively.
voltage gain of the quadratic Boost converter allows to maintain the ( ( ) )
output voltage in a level at which the SRM drive can still operate and, VD
ID = IO exp − 1 (2)
consequently, the water pump system also. The proposed solution is nVT
ideal for water pumping systems with storage tanks, taking advantage of
all the water flow now available, as a result of the floating nature of solar ISC TC + μI .(T − 298, 15)
IO = ⎛ (S ) ⎞ (3)
energy and consequent SRM speed variation. Additionally, the three- VOC TC +μV .(T− 298,15)
S

level operation of the converter reduces the size of the storage ele­ ⎜ ⎟
nVT
⎝e − 1⎠
ments and the voltage stress on the power semiconductors and provides
dual-output voltage. The dual-output voltage also introduces an

43
A. Cordeiro et al. Solar Energy 209 (2020) 42–56

Fig. 1. Detailed diagram of the proposed solution.

Table 1
Characteristics of the PV panel (Sungoldsolar, 2019).
Model SGP-320W36V

Maximum power (Pmax) 320 W

Voltage at Pmax (VMP) 35.6 V
Current at Pmax (IMP) 8.99 A
Open-circuit voltage (VOC) 42.7 V
Short-circuit current (ISC) 9.71 A
Cell efficiency (%) 18.30%
Fig. 2. Electric equivalent circuit of an ideal solar cell including the series and Power temp. coefficient − 0.41%/◦ C
Voltage temp. coefficient − 0.32%/◦ C
parallel resistances.
Current temp. coefficient +0.05%/◦ C
Output power tolerance ±3%
The thermal voltage for a PV with Ncell modules is given by (4), where Number of cells 72 (6 × 12)
k is the Boltzmann constant, q is the elementary charge, and T is the Standard testing conditions (STC):1000 W/m2; AM 1.5.
temperature. The output voltage of the solar PV with Ncell modules is
expressed by (5).
2.2. DC-DC converter
k.TNcell
VT = (4)
q Photovoltaic cells present a high voltage variation as a function of
the solar irradiation. In this way, to obtain a better harnessing of the
VPV = VD − (RS IPV ) (5) energy generated by the solar panels, a DC-DC converter that could be
able to handle an input high voltage gain is required. Besides that, when
The combination of Eqs. (1) to (5) leads to the I-V characteristic Eq.
the solar irradiation is reduced, the solar panel output voltage becomes
(6). In this model, the VPV(n− 1) and IPV(n− 1) represent the voltage and the
very low, by which a converter with a high voltage gain capability is also
current of the previous iteration.
⎛ ( ⎞ required. The use of conventional converters like the Boost is not usually
)
(VPV(n− 1) +RS IPV(n− 1) )q ( ) the best option since their static voltage gain is limited by the parasitic
⎜
IPV(n) = − IO ⎜
nkTNCell ⎟
− 1⎟
VPV(n− 1) + RS IPV(n− 1) elements (Tofoli et al., 2015). To solve this problem, converters with
⎝e ⎠− + IPVS C
RP very high static voltage gain are required. Another issue associated to
this kind of converters is the possibility to use topologies with more than
(6) two levels (Dusmez et al., 2015; Vorobei et al., 2015). In fact, compared
The influence of solar irradiation and temperature in the open-circuit with the classical two levels, they present several advantages, such as
voltage and short-circuit current are given by (7) and (8). These values reduction in the size of the storage elements and the voltage stress of the
are specified by the PV module manufacturer (see Table 1). switches. In this way, for this application, a three-level quadratic Boost
(3LQB) DC-DC converter with dual output that provides quadratic static
VPVO C =
(VOCS TC + μV .(T − 298, 15) )G
(7) gain as a function of the duty-cycle ratio and reduced number of active
1000W/m2 and passive components is proposed. So, besides the advantage of
providing high voltage gain and being characterized by three levels
IPVS C =
(ISCS TC + μI .(T − 298, 15) )G
(8) (with the advantages already referred), it is also very well adapted to the
1000W/m2 SRM drive. In fact, since a drive with a reduced number of legs will be
Usually, several cells are associated to increase the voltage, creating used, by which they will be supplied by two DC sources, this topology
a PV array, because the DC voltage produced by a single solar panel is will intrinsically provide these two voltages. Fig. 3 illustrates the to­
typically reduced. pology of the proposed dual-output DC-DC converter.
This converter can create a dual-output voltage with opposite

44
A. Cordeiro et al. Solar Energy 209 (2020) 42–56

polarity considering a common connection point which can be used to voltage gain of this converter considering CCM operation. Considering
connect to the middle point of the SRM with a star winding connection, that the average voltages of inductors L1 (9) and L2 (10) in the operating
which will be described later. This converter is composed by two power modes I and IV (which are the main operating modes) are equal, the
devices to equalize the dual-output voltage whenever necessary. The following relationships are obtained considering balanced output volt­
proposed DC-DC topology is analyzed based on the assumption that the ages (VC2 = VC3):
converter operates in continuous conduction mode (CCM) and all
δ Vin + (1 − δ) (Vin − VC1 ) = 0 (9)
components are ideal. According to this, the converter can have four
different operating modes which can be combined to achieve the desired
δ VC1 + (1 − δ) (VC1 − Vout ) = 0 (10)
operation. In this converter, it is possible to combine the operating
modes I (Fig. 4a)) and IV (Fig. 4d)) when the voltages of the output Vout = VC2 + VC3 (11)
capacitors are balanced (VC2 = VC3). Despite the balanced load, there are
always small differences in components and connections which lead From the partial voltage gains, obtained from the previous Eqs. (9)
normally to small imbalances, affecting the correct operation of the and (10), the input-to-output voltage gain is given by (12):
application. Using this converter, it is possible to mitigate such possible Vout 1
imbalance due to the use of two power devices connected to the middle = (12)
Vin (1 − δ)2
point of the capacitors. Thus, when VC3 < VC2, it is necessary to combine
the operating modes I (Fig. 4a)), II (Fig. 4b)) and IV (Fig. 4d)) to achieve The theoretical CCM waveforms of the proposed DC-DC converter,
balance. Similarly, when VC2 < VC3, it is necessary to combine the considering balanced output voltage, are shown in Fig. 5. In order to
operating modes I (Fig. 4a)), III (Fig. 4c)) and IV (Fig. 4d)) to achieve simplify the analysis, it is assumed in this figure that the solar PV panels
balance. The output voltage balance is accomplished using a closed-loop provide a constant voltage and components are considered ideal.
voltage controller which will be described later. The operating modes Based on the analysis of these two operating modes (I and IV), it is
are: also possible to obtain the following relationships:
Vin
• Operating mode I (Fig. 4a)): This operating mode is effective when VC1 = (13)
(1 − δ)
the power switches S1 and S2 are turned on. In this mode, the in­
ductors L1 and L2 are in charging mode receiving energy from the PV
Io = IL1 (1 − δ)2 = IL2 (1 − δ) (14)
array (represented by a constant DC power supply) and capacitor C1,
respectively. C2 and C3 are in discharging mode over the load (which Performing a similar analysis considering that VC2 < VC3 (which
is represented by a current source). Diodes D1, D3 and D4 are blocked corresponds to operating modes I, III and IV), the following relationships
due to reverse voltages. can be obtained (δ = δ − δ ′ ):
′ ′

• Operating mode II (Fig. 4b)): This mode starts when the power
(15)
′ ′ ′

switch S2 turns off after the operating mode I (S1 remains turned on). δ Vin + δ ′ (Vin − VC2 ) + (1 − δ − δ ′ ) (Vin − VC1 ) = 0
The inductors L1 and L2 remain in charging mode, similar to mode I,
(16)
′ ′ ′

but in this case the charge rate is different since capacitor C3 is also in δ VC1 + δ ′ (VC1 − VC2 ) + (1 − δ − δ ′ ) (VC1 − Vout ) = 0
charging mode, allowing to increase the voltage over this capacitor. From the partial voltage gains, obtained from Eqs. (15) and (16), the
Capacitor C2 is in discharging mode over the load. In this condition, input-to-output voltage gain is given by (17):
D4 turns on.
• Operating mode III (Fig. 4c)): This mode is similar to mode II, but in Vout
=
1
(17)
Vin (1 − δ ′ )2 − VC2 δ ′ [2(1 − δ ′ ) − δ] (1 − δ − δ ′ )2
′ ′ ′ ′
this case the power switch S1 turns off and S2 remains turned on. The
capacitor C2 is in charging mode, and C3 is in discharging mode. In The theoretical CCM waveforms of the proposed DC-DC converter,
this condition, D3 turns on. considering unbalanced output voltage VC2 < VC3, can be seen in Fig. 6.
• Operating mode IV (Fig. 4d)): In this condition, both power switches Similarly, when VC3 < VC2 (modes I, III and IV), the input-to-output
are turned off and the inductors L1 and L2 are in discharge mode. In voltage gain is given by (18):
this mode, all the capacitors are in charging mode. In this condition,
D1, D3 and D4 turn on. Vout
=
1
(18)
Vin (1 − δ ′ )2 − VC3 δ ′ [2(1 − δ ′ ) − δ] (1 − δ − δ ′ )2
′ ′ ′ ′

From the analysis of the operating modes, it is possible to obtain the

Fig. 3. Three-level quadratic Boost DC-DC converter with a dual output.

45
A. Cordeiro et al. Solar Energy 209 (2020) 42–56

Fig. 4. Converter operating modes: a) Mode I (principal charging mode), b) Mode II (secondary charging mode over C3), c) Mode III (secondary charging mode over
C2), d) Mode IV (discharging mode).

Notice that Eqs. (17) and (18), when δ ′ = 0⇒δ = δ, have the same
′ ′

result of Eq. (12).

2.3. MPPT control mode

MPPT is currently an essential part of any PV system. The aim of

MPPT techniques is to automatically find the voltage VMP or current IMP
at which a PV array should operate to obtain the maximum power
output under a given temperature and irradiation (Subudhi and Prad­
han, 2013). Due to the extreme importance of assuring that the PV
panels operate at this point, many publications have presented different
proposals regarding MPPT. However, there are two that are considered
the most well-known and used. Those two algorithms are designated as
perturb & observe (Kumar et al., 2019) and incremental conductance
(Tey and Mekhilef, 2014). In this paper, a different MPPT control system
that results in a robust and very easy practical implementation is
proposed.
Usually, the MPPT algorithms are based on the analysis of the P-V
characteristic of the solar cells. Indeed, through that analysis, it is
concluded that the MPP of a solar cell is achieved when the derivative of
the solar cell power with respect to its voltage is equal to zero. However,
using an algorithm based on this concept, singularities of the system
(Roncero-Clemente et al., 2013b) could originate. Thus, the proposed
MPPT method will be based on a different approach. Considering (19)
and the typical characteristic curve of the PV cells (see Fig. 7), the MPPT
system can be analyzed in order of time. However, in this case, the de­
rivative of output power and the derivative of output voltage with
respect to time must be considered. Analyzing the conditions dP/dt >
0 and dV/dt > 0, it is possible to verify that this results in a trajectory
toward the MPP (Fig. 7). However, the conditions dP/dt > 0 and dV/dt <
0 also result in a trajectory toward the MPP since now it is located on the
right side of that point. If dP/dt = 0, then the system is at the MPP.
According to this, the algorithm is based on these conditions, as pre­
sented in (20).
Fig. 5. Theoretical CCM waveforms of the proposed DC-DC converter consid­
ering balanced output voltage (modes I and IV).

46
A. Cordeiro et al. Solar Energy 209 (2020) 42–56

dP dP
= dtdV (19)
dV dt

⎧
⎪
⎪ dP
⎪
⎪ =0 , at MPP
⎪
⎪ dt
⎪
⎪ ( ) ( )
⎨ dP dV dP dV
> 0 and > 0 or < 0 and < 0 , left of MPP
⎪ dt
⎪ dt dt dt
⎪
⎪ ( ) ( )
⎪
⎪ dP dV dP dV
⎪
⎪
⎩ > 0 and < 0 or < 0 and > 0 , right of MPP
dt dt dt dt
(20)
To guarantee that the PV panel operates at MPP, it is necessary to
correctly control the DC-DC converter. So, the MPPT algorithm should
increase or decrease the panel voltage depending on the exact location
(left side or right side) regarding the MPP. However, this increase, or
decrease, must be a function of the PV output power and the derivative
of the PV output voltage with respect to time. To assure that, a control
law is established in accordance with (21), in which the duty-cycle
reference (δref) for the switches of the DC-DC converter is defined.
∫
dP dV
δref = k dt (21)
dt dt
The control law defined by (21) ensures that the PV panels operate at
their MPP. This control law can also be expressed as function of the sign
of the derivatives. In accordance with this, it is possible to obtain a new
control law, which is now defined by:
∫ ( ) ( )
dP dV
δref = k sign sign dt (22)
dt dt
The control law presented in Eq. (22) will be used as the MPPT
control algorithm of the proposed system. This algorithm can also be
seen in Fig. 8. Analyzing this algorithm, it is possible to see that is quite
simple. Indeed, it can easily be implemented and in a very cheap way
since it only requires a simple microcontroller or even a simple analog
circuit.
Fig. 6. Theoretical CCM waveforms of the proposed DC-DC converter consid­
ering unbalanced (VC2 < VC3) output voltage (modes I, III and IV). 2.4. Output voltage balance control

The output voltage balance control algorithm is responsible for

equalizing the voltage of the output capacitors. Thus, to develop this
control algorithm, different duty cycles for the switches (δ1 and δ2) are
defined, in accordance with:
{ ′
δ1 = δref + δ ′
′′ (23)
δ2 = δref − δ

where δref is the duty cycle generated by the MPPT, and ∂ is the
correction variable that is used to balance the capacitor voltages (C2 and
C3).
The correction of the duty cycles of each switch is a function of the
voltage error between both outputs. This can be achieved by a PI
compensator, in accordance with:
∫
( ) ( )
(24)
′
δ ′ = K P V C 2 − V C3 + K I VC2 − VC3 dt

Fig. 7. Illustrative cases of MPP and dP/dt slope with respect with dV/dt > 0.

Fig. 8. Proposed MPPT control algorithm associated to the DC-DC converter.

47
A. Cordeiro et al. Solar Energy 209 (2020) 42–56

Fig. 9. Proposed output voltage balance control.

Fig. 9 shows the implementation scheme of the output voltage bal­ (PI) with gains KP = Tz/Tp and KI = 1/Tp, is represented in Fig. 10 as a
ance control. closed-loop system block diagram.
The determination of the PI parameters is realized through the To obtain the PI compensator parameters Tz and Tp, the closed-loop
analysis of the system in the closed-loop function of the voltage imbal­ transfer function may be obtained from the block diagram of the system
ance. Thus, for this analysis, each output capacitor voltage dynamics, in a closed loop:
VC2 and VC3, must be considered, which are given by: (1+sTz )
K ΔvC *
Td Tp C G
− sKG (1+sT d)
Td C o
I (s)
dvC2 ΔvC (s) = (32)
iC2 = C2 = iD3 − io (25) s3 + T1d s2 + TKdGTTp Cz s + TdKTGp C
dt
Considering ΔvC * = 0, the transfer function can be simplified as:
dvC3
iC3 = C3 = iD4 − io (26)
dt − sKG (1+sT d)
I (s)
(33)
Td C o
ΔvC (s) =
The dynamics of the currents iD3 and iD4 can be related to the current s3 + T1d s2 + TKdGTTp Cz s + TdKTGp C
iL1 with a gain and a time delay first-order model (27)–(28), where KG2
and KG3 are converter gains derived from (17) and (18), considering The parameters of the PI compensator will be determined in accor­
power conservation. The time constant Td approximates the quadratic dance with the ITAE criteria (Ogata, 1997), by comparing the denomi­
converter internal dynamics of iL2 and the output voltage. nator of (33) to the third-order ITAE polynomial in (34):

KG2 P(s) = s3 + 1.75ωo s2 + 2.15ω2o s + ω3o (34)

iD3 = iL (27)
Td s + 1 1
The PI compensator parameters Tz and Tp are:
KG3 ⎧
iD4 = iL (28) ⎪
⎪
⎪ Tz =
CVout
Td s + 1 1 ⎨ Io
(35)
Applying the Laplace transform to (25) and (26) and replacing the ⎪
⎪
⎪ 4KG Td Vout
⎩ Tp =
currents iD3 and iD4 from (27) and (28), the output voltages are now Io
obtained by:
( ) where Vout is the total output voltage of the converter. Since KP = Tz/Tp
1 KG2 and KI = 1/Tp, the PI gains can be computed from (36) as:
vC2 (s) = iL1 − io (29)
sC2 Td s + 1
⎧ Tz 2.15 C
( ) ⎪
⎪ K = =
1 KG3 ⎪
⎨ P Tp 1.752 Td
vC3 (s) = iL1 − io (30) (36)
sC3 Td s + 1 ⎪
⎪
⎪ 1 C
⎩ KI = =
It follows that the voltage imbalance ΔvC = vC2 − vC3 can be Tp 1.753 KG Td2
expressed as (where it is considered that C2 = C3 = C):
( ) ( )
1 KG2 − KG3 1 KG 2.5. SRM drive
ΔvC (s) = iL1 − io = iL1 − io (31)
sC Td s + 1 sC Td s + 1
The equivalent voltage balancing loop, using a proportional-integral To describe the SRM operating principle, it is necessary to recall that
the reluctance of the machine magnetic circuit and its variations are

Fig. 10. Closed-loop output voltage balance control diagram.

48
A. Cordeiro et al. Solar Energy 209 (2020) 42–56

dependent on the rotor position. The variation of the magnetic reluc­ The equivalent circuit structure of PV panels can be modeled as an
tance in the SRM is essential to have torque (reluctance torque), and it equivalent impedance circuit (Houshmand et al., 2013). Considering the
can be properly controlled by accurately controlling the time of ener­ equivalent circuit structure of the PV panel, in which this circuit consists
gizing and de-energizing the stator phases. In this machine, the torque usually in a voltage source, resistors and capacitors, since at the input of
(Tj) developed by each phase can be expressed by (37) as a function of the DC-DC converter there is an inductor, which is connected to a switch
the phase current (ij) and the self-inductance coefficient (Lj), which that will short-circuit or not that inductor, it can be considered that from
depends on the rotor position (Zhang et al., 2019; Han et al., 2019). this point of view, the input converter will act as a current source. That
current source can be increased or decreased depending on the duty
( ) 1 dLj (θr ) 2
Tj θr , ij = i (37) cycle of the switches. So, the model will only slightly affect the growth or
2 dθr j
decrease time. Once the output of the DC-DC converter is composed by
The methodology adopted to produce torque in this machine is based large capacitors, then the DC-DC converter can be modeled as a voltage
on a simple flat-top PWM current controller, as presented in Fig. 11. source, which means that the SRM drive, connected to the output of the
The current reference (iref) of the controller is the nominal value of DC-DC converter, is supplied by an equivalent DC power source. The
the SRM machine nameplate. In this way, all the energy produced by the current control of the SRM drive requires intermittent current from both
solar PV array in each moment can be used to produce torque in the SRM output capacitors and can be considered as an equivalent variable cur­
machine up to the nominal value. Even when solar irradiation decreases, rent source, connected to an equivalent DC power source (output ca­
the MPPT is used to obtain the maximum power output in each specific pacitors of the converter). The DC-DC converter being controlled to
moment. The available power output will be used to maintain the SRM explore the MPP of the PV panels, the SRM drive practically does not
machine running (and the water pump) even at reduced speeds. The influence the input current since the current in the inductor L1 remains
operating limits of this solution for water pumping purposes should be almost constant and shows a high degree of decoupling from the SRM
determined by the minimum water flow rate required, necessary head drive operation while the current in inductor L2 is quite dependent on
(elevation) and net positive suction head (NPSH). the SRM drive operation.
The selected values for the passive components of the DC-DC con­
3. Simulation results verter were L1 = L2 = 10 mH, C1 = 20 µF and C2 = C3 = 470 µF. The first
simulation result can be seen in Fig. 12. This figure presents the behavior
The proposed solution was simulated using MATLAB/Simulink soft­ of the adopted MPPT control regarding the solar PV panels (4 × 320 W)
ware. The ideal model described in Section 2.1 was chosen for the and load variation. In this simulation, the DC-DC converter was supplied
simulation of the solar PV array. The data for this model were obtained by the equivalent model of the solar PV panels (T = 25 ◦ C and G = 1000
from the commercial PV SGP-320W36V Poly Solar Panel (see Table 1), W/m2) and connected to a resistive load to firstly test the capability to
and four PV panels were considered in the presented simulations. balance the output voltages. In this situation, the voltage gain of the DC-
The proposed DC-DC converter was simulated in dynamic operation DC converter is about 2.5, but it is possible to achieve higher voltage
using the PV model as power source and the SRM drive as load. A typical gains in case of reduced values of solar irradiation. At t = 0.2 s, a sudden
quadratic torque load was adopted in order to simulate the water pump change in the resistive load was introduced and removed at t = 0.35 s.
(Vodovozov et al., 2014). The MPPT algorithm was combined with the Fig. 12 demonstrates that the output voltage (VPV) and the output cur­
output voltage balance control to adjust the duty cycle of the power rent (IPV) of the solar PV panels remain at the same levels despite this
devices of the DC-DC converter and to provide balanced voltages. load variation. The output power (P0) and output currents (io1, io2) of the

Fig. 11. PWM current control solution of SRM drive.

49
A. Cordeiro et al. Solar Energy 209 (2020) 42–56

Fig. 12. Simulation result of the behavior of the MPPT control adopted regarding the solar PV panels (4 × 320 W) and load variation.

DC-DC converter can also be seen in Fig. 11, showing that the MPPT (io1, io2), the voltages on both output capacitors (VC2, VC3) and the gate
control adjusts properly to the respective load variations. drive signals (VG1, VG2), considering an unbalanced resistive load
Another simulation result can be seen in Fig. 13, which intends to (modes I, III and IV) in the DC-DC converter (VC2 < VC3). In this situa­
demonstrate the behavior of the output voltage balance control of the tion, the duty cycle of power devices is different as well as the electrical
DC-DC converter. In this simulation, a different resistive load was con­ waveforms of both inductors and output capacitors (voltages and cur­
nected in parallel with each output capacitor (C2 and C3) of the DC-DC rents). Despite such differences, the adopted voltage control can main­
converter at t = 0.2 s, to produce an effective imbalance in each output tain the voltage balance on both output capacitors. This simulation
capacitor. At t = 0.35 s, the initial resistive loads were introduced again result indicates that the output voltage balance control is always
to return to the original condition (equal resistive loads). This simula­ adjusting the necessary duty cycles to equalize the output voltage
tion result was performed with the output voltage balance control values. Consequently, the currents in the inductors increase or decrease
disabled. In this figure, the dynamic behavior of the total output voltage to fulfill this requirement.
(Vout), output currents (io1, io2) and the respective individual voltages of After the confirmation of the DC-DC converter regarding the high
capacitors C2 and C3 can be seen, showing a clear imbalance at t = 0.2 s voltage gain and the capability to balance the two output voltages,
and t = 0.35 s. Due to the MPPT control algorithm, when the load in­ additional simulations were performed in order to demonstrate that the
creases, the output voltage of the DC-DC converter must decrease pro­ DC-DC converter can provide energy to the four-phase 8/6 SRM drive
portionally to maintain a constant PV output power. This effect can also and water pump as electromechanical load. An example of the behavior
be seen in this figure. of the DC-DC converter in this condition can be seen in Fig. 16. In this
The simulation presented in Fig. 14 illustrates the performance of the simulation test, the DC-DC converter is always regulating the MPP from
output voltage balance control of the proposed DC-DC converter, PV panels and the voltage of both output capacitors according to load
showing the balance capability between t = 0.2 s and t = 0.35 s despite demand. Note that the SRM drive circuit is always requesting intermit­
different resistive loads connected to each capacitor. In this figure, it is tent current from the DC-DC output capacitors, depending on the SRM
also possible to see the fast dynamic response of the output voltage rotor position, speed, and water pump mechanical load. Due to this, the
balance control where both capacitors maintain approximately the same voltages VC2 and VC3 and currents IL1 and IL2 of the DC-DC converter are
voltage. always oscillating according with the demand of the SRM drive. In this
As mentioned in Section 2.2, there are four operating modes for the figure, it is possible to see the continuous variations in the duty cycles of
proposed DC-DC converter regarding the output voltage balance re­ both power devices, reflected in the waveforms of VL1 and VL2 as a
quirements. Fig. 15 shows the simulation result of the voltages and consequence of the output voltage control. It is also possible to realize
currents on both inductors (VL1, VL2) and (iL1, iL2), the output currents that the magnitude of VL1 remains constant despite the load variation.

50
A. Cordeiro et al. Solar Energy 209 (2020) 42–56

Fig. 13. Simulation result of the behavior of the total output voltage (Vout), output currents (io1, io2) and individual voltages (VC2, VC3) of capacitors C2 and C3
considering the output voltage balance control disabled.

Fig. 14. Simulation result of the behavior of the total output voltage (Vout), output currents (io1, io2) and individual voltages (VC2, VC3) of capacitors C2 and C3
considering the output voltage balance control enabled.

This is because the inductor voltage depends on the output voltage of the simulation test, it was considered that the PV panels are operating at T =
PV panel, which is approximately constant due to the MPPT algorithm, 25 ◦ C and G = 1000 W/m2.
as can be seen in Fig. 12. Due to this aspect, the current IL1 presents very Another simulation test was performed regarding the specific oper­
small variations whereas the current IL2 presents high variations. In this ation of the SRM drive during the startup acceleration of the SRM drive.

51
A. Cordeiro et al. Solar Energy 209 (2020) 42–56

Fig. 15. Simulation result of the voltages and currents on both inductors (VL1, VL2) and (iL1, iL2), output current (io1), voltages on both output capacitor (VC2, VC3)
and gate drive signals (VG1, VG2) considering an unbalanced resistive load in the DC-DC converter.

This simulation, presented in Fig. 17, shows the current in the four converter, considering unbalanced resistive loads. In this experimental
phases (ia, ib, ic and id), the electromagnetic torque (Te) and the rotor test, different resistive loads were introduced in each output capacitor
speed (ωm), considering that the PV panels are operating at T = 25 ◦ C terminal (33 Ω @ C2, 50 Ω @ C3) through the available common
and G = 1000 W/m2. This simulation result shows the controlled cur­ connection point. The objective of this experimental test was to deter­
rents in the four phases of the SRM drive and the duration of the control mine the operation of the output voltage balance control considering
in each phase function of the motor speed. The higher the speed, the different resistive loads. This result shows that the control is operating
shorter the control duration of each phase in the SRM. The speed, the correctly since both capacitors have equal voltages although each one
torque and the current waveforms are typical for this type of motor. supplies a different resistive load, as can be seen by the output currents
io1 (≈ 5 A) and io2 (≈ 3 A).
4. Experimental results Another experimental result, performed in the same conditions, can
be seen in Fig. 19. The objective of this experimental test is to analyse
The experimental results were obtained from a developed prototype the currents in both inductors (iL1, iL2), the voltage on both inductors
to confirm the operation of the proposed solution. Since PV panels were (VL1, VL2) and the voltage gate signals (VG1, VG2) on both power devices,
not available, a remotely controlled DC power source (EA-PS 8160-60) regarding the same unbalanced resistive loads used in the experimental
was used to simulate the behavior of such devices. The DC power sup­ test of Fig. 18. In this figure, it is possible to confirm that the power
ply was controlled by a dSPACE device (DS1104) considering the devices have different duty cycles as a result of the output voltage bal­
equations presented in Section 2.1, allowing to emulate the I-V curve of ance control operation. Notice that in this experimental test, capacitor
the PV panels. The passive components of the DC-DC prototype are C2 has a small resistive load in parallel, and therefore the duty cycle of
similar to those adopted in the simulation tests. The SRM drive is power transistor S2 is higher than that of S1. This difference allows to
composed by an asymmetrical half-bridge converter and a four-phase transfer energy to the output capacitor that requires it most. In this
1.2 kW, 220 V, 1550 rpm, 8/6 SRM. The MPPT algorithm, output figure, it is also possible to see the synchronization between currents and
voltage balance control and PWM current controller of the SRM drive voltages on both inductors. The experimental waveforms of inductors
were implemented in an FPGA development board (Xilinx Zybo board). and power semiconductors are in accordance with the theoretical
All the analog-to-digital signal conversion was done using the evaluation considerations.
board AD7606BFMCZ. Fig. 20 shows another experimental result performed with the DC-DC
Fig. 18 shows an experimental result of the voltage on both output converter considering a change from balanced to unbalanced load. In
capacitors (VC2, VC3) and output currents (io1, io2) of the DC-DC this figure, it is now possible to confirm the transient waveforms of the

52
A. Cordeiro et al. Solar Energy 209 (2020) 42–56

Fig. 16. Simulation result of the voltages (VL1, VL2) and currents (iL1, iL2) on both inductors and the voltages on both output capacitor (VC2, VC3) using the DC-DC
converter to supply the 8/6 SRM drive and water pumping system.

Fig. 18. Experimental results of the output current (io1, io2) (Ch3 = Ch4 = 5 A/
Fig. 17. Simulation result of the currents in the four phases (ia, ib, ic and id), div) and voltages on both output capacitors (VC2, VC3) (Ch1 = Ch2 = 100 V/
electromagnetic torque (Te) and rotor speed (ωm) considering the 8/6 SRM div) in the proposed DC-DC converter considering an unbalanced resistive load
during startup acceleration. (33 Ω @ C2 terminals and 50 Ω @ C3 terminals).

voltages on both output capacitors (VC2, VC3) and the output current (io1) system. In this figure, it is possible to see that under this condition the
because of the output voltage control based on the classical PI controller. currents in both inductors are not stationary, especially in the inductor
This controller takes about 10 to 15 ms to stabilize the new values. Since L2, due to the operation of the SRM (which depends on the mechanical
the MPPT control algorithm is operating to provide constantly the load imposed by the centrifugal pump) and the duty cycle of the power
maximum output power, when the load in the DC-DC converter is semiconductors, as a consequence of the closed-loop output voltage
higher, the output voltages of the converter decrease, as presented in control. Obviously, the load dynamics produce higher voltage variations
this figure, according to the PI voltage control algorithm. Again, this is in in the output capacitors and, consequently, higher variation in the
accordance with the theoretical considerations. current of inductor L2. Once the input voltage of the DC-DC converter is
Finally, some experimental results were performed considering the approximately constant, the current variation in L1 is only dependent on
entire system. Thus, as mentioned before, a PV simulator (by a the duty cycle of the power semiconductors. Fig. 21b) shows the voltage
controlled voltage source) was used, the controlled DC-DC converter on both output capacitors of the DC-DC converter in the same condi­
connected to the controlled SRM drive and the motor. Fig. 21a) shows tions. In this figure, it is possible to see that both capacitors have
the current on both inductors of the DC-DC converter for the entire approximately the same voltage with small variation. In this situation,

53
A. Cordeiro et al. Solar Energy 209 (2020) 42–56

Fig. 19. Experimental results of the proposed DC-DC converter considering an unbalanced resistive load (33 Ω @ C2 terminals and 50 Ω @ C3 terminals): a) Currents
on both inductors (iL1, iL2) (Ch3 = Ch4 = 5 A/div) and voltages on both inductors (VL1, VL2) (Ch1 = Ch2 = 200 V/div), b) Voltage gate signals on both power devices
(VG1, VG2) (Ch1 = Ch2 = 10 V/div).

the closed-loop output voltage control is always operating to equalize

the voltage of both output capacitors.
Some additional experimental results of the 8/6 SRM drive operation
supplied by the DC-DC converter are presented in Fig. 22. In this figure,
the current waveforms (ia, ib, ic and id) in the four phases of the SRM
when the DC-DC converter is connected to the PV panel simulator (T =
25 ◦ C) with an equivalent G = 1000 W/m2 (Fig. 22a)) and with an
equivalent G = 250 W/m2 (Fig. 22b)) can be seen. This experimental
result allows to conclude that, when the solar irradiation is significant,
the SRM drive can impose the reference currents into the SRM drive and
the motor runs as expected (Fig. 22a)). The reference currents in the
SRM drive were defined to be 6 A. Nevertheless, when the solar irradi­
ation is reduced and the energy generated by the PV panel is reduced, it
is difficult to impose the reference currents (Fig. 22b)) in the SRM drive.
This happens because the output voltage of the DC-DC converter is too
low to impose the necessary currents in the SRM drive. In this figure, it is
possible to see that the currents do not follow the reference value,
reducing the motor torque and consequently the speed. Nevertheless, it
Fig. 20. Experimental results of the voltages on both output capacitors (VC2,
is possible to see that the motor is still running at low speed and the
Vc3) (Ch1 = Ch2 = 50 V/div) and output current (io) (Ch3 = 5 A/div)
capability to supply water flow depends on the water pump character­
considering the transient change from balanced to unbalanced load.
istics and operating conditions such as the necessary pump head.
The currents at the four phases and SRM rotor speed (ωm) during the

Fig. 21. Experimental result of the proposed DC-DC converter when connected to the SRM drive: a) Currents on both inductors (iL1, iL2) (Ch3 = Ch4 = 2 A/div), b)
Voltages on both output capacitors (VC2, VC3) (Ch1 = Ch2 = 50 V/div).

54
A. Cordeiro et al. Solar Energy 209 (2020) 42–56

Fig. 22. Experimental result of the currents in the four phases (ia, ib, ic and id) (5 A/div) of the 8/6 SRM considering the DC-DC converter connected to the PV panel
simulator (T = 25 ◦ C) with: a) G = 1000 W/m2, b) G = 250 W/m2.

startup acceleration, considering the DC-DC converter connected to the change of the PV panel parameters only affects the slope of that varia­
PV panel simulator with G = 1000 W/m2, can be seen in Fig. 23. The tion, this will not affect the MPPT since the algorithm only analyzes if
water pump load characteristic was emulated in the laboratory using a this variation is positive, negative or zero. Regarding the output balance
controlled torque load. These experimental results show that the voltage control, the uncertainties of the DC-DC model parameters and
behavior of the 8/6 SRM drive is similar to that presented in the simu­ imposed load could affect the controller since it is a function of those
lation result (see Fig. 17) and according to what was expected, consid­ parameters. The description of the method used to calculate the gains of
ering this type of load. this controller, according to the model and load variation, was presented
These results show that the uncertainties associated to the experi­ in Section 2.4. However, as shown by the results, using the parameters
mental system did not affect their performance, maintaining a similar presented in that section, the results were similar to the simulation ones.
behavior when compared with the simulation results. The proposed The machine current control is a nonlinear hysteretic control which can
solution is composed by three different control systems, namely: the deal with the uncertainties produced by the operation of the SRM drive.
MPPT control algorithm, the output balance voltage control of the DC- The change of the machine parameters only affects the slope of the
DC converter and the PWM current control of the SRM drive. Since the current variation. Since the current is controlled by a hysteretic circuit
uncertainties could affect the operation of the control systems, they must control, that variation only affects the switching frequency but not the
be designed to withstand such uncertainties. Thus, the MPPT control is required value around the reference.
responsible for automatically finding the voltage VMP or current IMP at
which the PV panels should operate to obtain the maximum power 5. Conclusions
output under a given temperature and irradiation. The MPPT is only a
function of the variation of the PV panel voltage and current and is in­ A water pumping system, powered by solar photovoltaic panels,
dependent of the load connected to the DC-DC converter. So, since the employing a switched reluctance motor (SRM) drive and a three-level

Fig. 23. Experimental results of the currents in the four phases (ia, ib, ic and id) (5 A/div) and rotor speed (ωm) (5 V/500 rpm) of the 8/6 SRM during startup
acceleration considering the DC-DC converter connected to the PV panel simulator with G = 1000 W/m2.

55
A. Cordeiro et al. Solar Energy 209 (2020) 42–56

quadratic Boost (3LQB) DC-DC converter was presented in this paper. Gopal, C., Mohanraj, M., Chandramohan, P., Chandrasekar, P., 2013. Renewable energy
source water pumping systems—A literature review. Elsevier, Renewable
The proposed DC-DC converter provides quadratic static gain and
Sustainable Energy Rev. 25, 351–370.
reduced number of active and passive components. This converter can Guechi, A., Chegaar, M., 2007. Effects of diffuse spectral illumination on microcrystalline
create a dual-output balanced voltage with opposite polarity considering solar cells. J. Electron Dev. 5, 116–121.
a common connection point which is used to connect to the middle point Hamrouni, N., Jraidi, M., Chérif, A., 2009. Theoretical and experimental analysis of the
behavior of a photovoltaic pumping system. Sol. Energy 83 (8), 1335–1344.
of the SRM drive. The use of this DC-DC converter allows to use an AHB Han, G., Chen, H., Guan, G., 2019. Low-cost SRM drive system with reduced current
converter also with reduced number of active power switches. An MPPT sensors and position sensors. IET Electr. Power Appl. 13 (11), 853–862.
algorithm that is based in the variation of the power and voltage with Houshmand, M., Zandi, M., Gorji, N.E., 2013. Modeling the impedance of
Nanostructured PV in Simulink/MATLAB. Modern Phys. Lett. B, 27(20), 1350139
respect to time was also proposed, resulting in a robust and very easy (11 pages), World Scientific Publishing Company, doi: 10.1142/
practical implementation. The combination of all the proposed devices S021798491350139X, 1–11.
and systems allows to fully exploit the solar energy to provide as much as Kumar, R., Singh, B., 2018. Brushless DC motor-driven grid-interfaced solar water
pumping system. IET Power Electron. 11 (12), 1875–1885.
possible a continuous water flow in isolated pumping solutions. The Kumar, R., Singh, B., 2019. Solar PV powered-sensorless BLDC motor driven water pump.
performance of the proposed solution was tested through several com­ IET Renew. Power Gener. 13 (3), 389–398.
puter simulations and verified through some experimental tests using a Kumar, N., Singh, B., Panigrahi, B.K., 2019. Integration of solar PV with low-voltage
weak grid system: using maximize-M Kalman filter and self-tuned P&O algorithm.
laboratorial prototype developed for this purpose. The obtained results IEEE Trans. Ind. Electr. 66 (11), 9013–9022.
allowed to confirm the expected theoretical considerations. Koreboina, V.K., Narasimharaju, B.L., Kumar, D.M.V., 2017. Performance investigation
of simplified PWM MPPT approach for direct PV-fed switched reluctance motor in
water pumping system. IET Electr. Power Appl. 11 (9), 1645–1655.
Declaration of Competing Interest
Li, G., Jin, Y., Akram, M.W., Chen, X., 2017. Research and current status of the solar
photovoltaic water pumping system – A review. Renew. Sustain. Energy Rev. 79,
The authors declare that they have no known competing financial 440–458.
interests or personal relationships that could have appeared to influence Mishra, A.K., Singh, B., 2016. A single stage solar PV array-based water pumping system
using SRM drive. In: IEEE Industry Applications Society Annual Meeting, Portland,
the work reported in this paper. OR, USA, pp. 1–8.
Novaes, Y.R., Rufer, A., Barbi, I., 2007. A new quadratic, Three-Level, DC/DC converter
Acknowledgments suitable for fuel cell applications. In: Proc. Power Conversion Conference, Nagoya,
Japan, pp. 459–467.
Ogata, K., 1997. Modern Control Engineering, fourth ed.
This work was supported by Instituto Politécnico de Lisboa (IPL), Pires, V.F., Foito, D., Cordeiro, A., Silva, J.F., 2018a. A single-switch DC/DC buck-boost
Lisboa, Portugal. Projects IDI&CA 2019–SOLARPUMP. This work was converter with extended output voltage. In: Proc.7th Int. Conf. on Renewable Energy
Research and Appl. (ICRERA), Paris, France, pp. 791–796.
also supported by national funds through Fundação para a Ciencia e Pires, V.F., Foito, D., Cordeiro, A., 2018b. A DC–DC converter with Quadratic gain and
Tecnologia (FCT), Lisboa, Portugal, under contract UID/CEC/50021/ bidirectional capability for batteries/supercapacitors. IEEE Trans. Ind. Appl. 54 (1),
2019 and UID/EEA/00066/2019. 274–285.
Pires, V.F., Cordeiro, A., Foito, D., Silva, J.F., 2019. High step-up DC–DC converter for
fuel cell vehicles based on merged quadratic Boost-Ćuk. IEEE Trans. Vehicul.
References Technol. 68 (8), 7521–7530.
Roncero-Clemente, C., Husev, O., Minambres-Marcos, V., Stepenko, S., Romero-Cadaval,
Aliyu, M., Hassan, G., Said, S.A., Siddiqui, M.U., Alawami, A.T., Elamin, I.M., 2018. E., Vinnikov, D., 2013. Comparison of three MPPT algorithms for three-level neutral-
A review of solar-powered water pumping systems. Renew. Sustain. Energy Rev. 87, point-clamped qZ-source inverter. In: Proc. 8th Int. Conference-Workshop Comp.
61–76. And Power Electronics, Ljubljana, Slovenia, pp. 80–85.
Amir, Asim, Amir, Aamir, Che, H.S., Elkhateb, A., Rahim, M.A., 2019. Comparative Sarkar, N.I., Sifat, A.I., Rahim, N., Reza, S.M., 2015. Replacing diesel irrigation pumps
Analysis of high voltage gain DC-DC converter topologies for photovoltaic systems. with solar photovoltaic pumps for sustainable irrigation in Bangladesh: A feasibility
Elsevier, Renewable Energy 136, 1147–1163. study with HOMER. In: Proc. 2nd Int. Conf. on Electrical Inf. and Comm. Tech.
Asahi, H., Horikoshi, Y., 2019. MBE of III–V Semiconductors for Solar Cells. In: Molecular (EICT), Khulna, Bangladesh, pp. 498–503.
Beam Epitaxy: Materials and Applications for Electronics and Optoelectronics, 1st Sharma, U., Singh, B., Kumar, S., 2017. Intelligent grid interfaced solar water pumping
ed., Wiley Telecom, 2019, doi: 10.1002/9781119354987.ch16, pp. 265–278. system. IET Renew. Power Gener. 11 (5), 614–624.
Bottarelli, M.G., Barbi, I., Novaes, Y.R., Rufer, A., 2007. Three-level quadratic non- Singh, B., Kumar, R., 2016. Simple brushless DC motor drive for solar photovoltaic array
insulated basic DC-DC converters. In: Proc. Europ. Conf. on Power Electr. and Appl., fed water pumping system. IET Power Electron. 9 (7), 1487–1495.
Aalborg, Denmark, pp. 1–10. Singh, B., Shukla, S., 2018. Induction motor drive for PV water pumping with reduced
Cabral, J., Silva, T., Oliveira, S., Yales Rômulo de Novaes, Y.R., 2013. A new high gain sensors. IET Power Electron. 11 (12), 1903–1913.
non-isolated DC-DC boost converter for photovoltaic application. In: Proc. 2013 Subudhi, B., Pradhan, R., 2013. A comparative study on maximum power point tracking
Brazilian Power Electr. Conf., Gramado, Brazil, pp. 1–6. techniques for photovoltaic power systems. IEEE Trans. Sustainable Energy 4 (1),
Chandrasekaran, N., Thyagarajah, K., 2012. Modeling and performance study of single- 89–98.
phase induction motor in PV fed pumping system using MATLAB. Int. J. Elect. Eng. 5 Sungoldsolar, 2019. Sungoldsolar Solar Panel Manufacture. https://www.sungoldsolar.
(3), 305–316. com/SGP-320W36V-Poly-Solar-Panel-pd6089854.html, Accessed on 25 November
Cordeiro, A., Pires, Fernão V., Pires, A.J., Martins, J.F., Chen, H., 2019. Fault-tolerant 2019.
voltage-source-inverters for switched reluctance motor drives. In: Proc. IEEE 13th Tey, K., Mekhilef, S., 2014. Modified incremental conductance MPPT algorithm to
Int. Conf. on Compatibility, Power Electr. and Power Eng. (CPE-POWERENG), mitigate inaccurate responses under fast-changing solar irradiation level. Sol. Energy
Sonderborg, Denmark, pp. 1–6. 101, 333–342.
Das, M., Agarwal, V., 2015. Novel High-performance stand-alone solar PV system with Tofoli, F.L., Pereira, D.C., Josias de Paula, W., Oliveira Júnior, D.S., 2015. Survey on non-
high-gain high-efficiency DC–DC converter power stages. IEEE Trans. Ind. Appl. 51 isolated high-voltage step-up dc–dc topologies based on the boost converter. IET
(6), 4718–4728. Power Electron. 8(10), 2044–2057.
Dusmez, S., Hasanzadeh, A., Khaligh, A., 2015. Comparative analysis of bidirectional Vijayakumar, G., Amrutha, P.B., 2016. Fast dynamic response of SEPIC converter based
three-Level DC–DC converter for automotive applications. IEEE Trans. Ind. Electr. 62 photovoltaic DC motor drive for water pumping system. In: Proc. Int. Conf. on
(5), 3305–3315. Circuit, Power and Comp. Techn. (ICCPCT), Nagercoil, India, pp. 1–5.
Ebrahim, O.S., Badr, M.A., Elgendy, A.S., Jain, P.K., 2010. ANN-Based optimal energy Villalva, M.G., Gazoli, J.R., Filho, E.R., 2009. Comprehensive approach to modeling and
control of induction motor drive in pumping applications. IEEE Trans. Energy simulation of photovoltaic arrays. IEEE Trans. Power Electron. 24 (5), 1198–1208.
Convers. 25 (3), 652–660. Vodovozov, V., Gevorkov, L., Raud, Z., 2014. Modeling and analysis of pumping motor
Elsheikh, A.H., Sharshir, S.W., Elaziz, M.A., Kabeel, A.E., Guilan, W., Haiou, Z., 2019. drives in hardware-in-the-loop environment. J. Power Energy Eng. 2, 19–27.
Modeling of solar energy systems using artificial neural network: A comprehensive Vorobei, V., Zakis, J., Husev, O., Veligorskyi, O., Savenko, O., 2015. Simulation study of
review. Sol. Energy 180, 622–639. the three-level boost DCDC converter with full ZVS for PV application. In: Proc. 9th
Fang, X., Ding, X., Zhong, S., Tian, Y., 2019. Improved quasi-Y-source DC-DC converter Int. Conf. on Power Electronics-ECCE Asia, Seoul, Korea, pp. 2038–2043.
for renewable energy. CPSS Trans. Power Elec. and Appl. 4 (2), 163–170. Wu, H., Xu, P., Hu, H., Zhou, Z., Xing, Y., 2014. Multiport converters based on
Gonzalez-Llorente, J., Ortiz-Rivera, E.I., Salazar-Llinas, A., Jimenez-Brea, E., 2010. integration of full-bridge and bidirectional DC–DC topologies for renewable
Analyzing the optimal matching of dc motors to photovoltaic modules via dc-dc generation systems. IEEE Trans. Ind. Electron. 61 (2), 856–869.
converters. In: Proc. 25th Annu. IEEE Appl. Power Electron. Conf. Expo., Palm Zhang, W., Xu, A., Han, L., Wang, S., 2019. Minimizing torque ripple of SRM by applying
Springs, CA, USA, pp. 1062–1068. DB-DTFC. IET Electr. Power Appl. 13 (11), 1883–1890.

56