energies

Article
Experimental Implementation of a Flexible PV Power
Control Mechanism in a DC Microgrid
Hongwei Wu , Fabrice Locment * and Manuela Sechilariu
Sorbonne University, Université de Technologie de Compiègne, EA 7284 AVENUES, 60203 Compiègne, France;
hongwei.wu@utc.fr (H.W.); manuela.sechilariu@utc.fr (M.S.)
* Correspondence: fabrice.locment@utc.fr; Tel.: +33-344-234-423




Received: 28 February 2019; Accepted: 26 March 2019; Published: 31 March 2019


Abstract: The intermittent and highly variable nature of photovoltaic (PV) sources is always the
major obstacle to the growth of their deployment. Research work is increasingly demonstrating that
PV generation should not only be maximized but also flexible based on the system requirements.
This article presents a simple and flexible PV control mechanism, which can seamlessly switch
between maximum power point tracking mode and power limiting mode. It can be integrated into a
DC microgrid for efficient energy management. The proposed mechanism has two configurations
that respectively converge to a lower and a higher PV panel voltage to perform PV shedding.
The experimental validation carried out in this study shows that this control can effectively adjust the
PV generation despite some physical constraints. The limitations of the control mechanism and the
energy efficiency are also analyzed. It can be concluded that each configuration can be particularly
useful depending on the different application scenarios.

Keywords: renewable energy; photovoltaic; DC microgrid; power control; PV curtailment

1. Introduction
Photovoltaic (PV) power generation has been proven to be a potential alternative to traditional
fossil fuel power sources [1]. It is renewable and free of greenhouse gas emission and, importantly, it can
be installed in urban areas which can lead to less power transmission investment and lower energy
losses [2,3]. The PV panels are generally connected to the maximum power point tracking (MPPT)
controller, which maximizes the energy yield and injects all the power into the utility grid. Many
algorithms have been proposed for this purpose [4–6]; out of all MPPT algorithms, the perturbation
and observation (P&O) method is particularly favored by the market since it is easy to realize and
highly efficient [7]. However, there are technical issues with integrating the conventional electric
grid with decentralized energy generation [8,9]. Therefore, the intermittent nature of PV generators
creates the need for a large amount of ancillary services, which are relatively expensive [10]. Hence,
the annually installed PV capacity has decreased recently [11].
In order to overcome this constraint and increase the use of renewable energy, microgrid
technology has been proposed and becomes popular, especially urban PV implementation [12,13].
A microgrid is an aggregation of energy sources (both traditional and/or renewable), the energy
storage system and the electrical loads, which can be seen as a single unit in the public grid. The PV
power generation can be smoothed by means of energy storage and is ready for self-consumption
by local loads. Hence, the impact of PV power intermittency on the utility grid is greatly reduced
and the economic benefits that could result from this are evident [14–16]. Depending on the nature of
the common bus, the microgrids can be categorized into AC microgrids, DC microgrids, and hybrid
AC-DC microgrids. Recently, researchers have increasingly found that the DC microgrid has potential

Energies 2019, 12, 1233; doi:10.3390/en12071233 www.mdpi.com/journal/energies

Energies 2019, 12, 1233 2 of 12

Energies 2019, 12, x FOR PEER REVIEW 2 of 12

because of its compatibility with DC devices such as PV generators, battery storage and electronic
loads [17–19]. Thus, this work focuses on DC microgrids.
Grid-injected PV generators are often cut off entirely when the grid encounters a stability issue
Grid-injected PV generators are often cut off entirely when the grid encounters a stability issue [20].
[20]. In the microgrid context, partial PV shedding may be needed for different reasons, such as power
In the microgrid context, partial PV shedding may be needed for different reasons, such as power
balancing [21], dealing with the partial shading condition [22], supporting grid voltage by supplying
balancing [21], dealing with the partial shading condition [22], supporting grid voltage by supplying
reactive power [23] or the sudden increase of solar irradiance [24]. This type of shedding operation
reactive power [23] or the sudden increase of solar irradiance [24]. This type of shedding operation can
can be achieved by manipulating the PV output voltage and current. However, the main challenge is
be achieved by manipulating the PV output voltage and current. However, the main challenge is still
still the nonlinear characteristics of a PV generator. Previous research has been carried out on this
the nonlinear characteristics of a PV generator. Previous research has been carried out on this subject,
subject, such as the work of Reddy and Sarkar [24] and Wang et al. [25]. The partial shedding is
such as the work of Reddy and Sarkar [24] and Wang et al. [25]. The partial shedding is generally
generally achieved by increasing the PV voltage, since it leads to a fast response and weak line losses.
achieved by increasing the PV voltage, since it leads to a fast response and weak line losses. However,
However, the physical constraints of the PV converter have not yet been taken into account.
the physical constraints of the PV converter have not yet been taken into account. Moreover, there is a
Moreover, there is a lack of analysis on the energy efficiency during the PV shedding procedure.
lack of analysis on the energy efficiency during the PV shedding procedure.
This study proposes a flexible PV power control mechanism based on the previous work
This study proposes a flexible PV power control mechanism based on the previous work presented
presented in Wang et al. [25]. Thanks to a classical P&O MPPT algorithm and a proportional-integral
in Wang et al. [25]. Thanks to a classical P&O MPPT algorithm and a proportional-integral (PI) power
(PI) power controller, it is supported in both the MPPT mode and the power limiting mode. In
controller, it is supported in both the MPPT mode and the power limiting mode. In addition, the
addition, the seamless switch between the two modes is feasible. The PI power controller allows for
seamless switch between the two modes is feasible. The PI power controller allows for a faster dynamic
a faster dynamic response than the perturbation methods. It was also chosen because of its flexibility;
response than the perturbation methods. It was also chosen because of its flexibility; with an easy
with an easy reconfiguration, PV shedding can be achieved by reducing or raising the PV voltage.
reconfiguration, PV shedding can be achieved by reducing or raising the PV voltage. This choice
This choice is usually linked to the converter constraints. Nevertheless, it has an impact on the
is usually linked to the converter constraints. Nevertheless, it has an impact on the efficiency of
efficiency of the system.
the system.
In light of the statements above, the paper is organized as follows: The principle of PV power
In light of the statements above, the paper is organized as follows: The principle of PV power
limit control is presented in Section 2, the proposed PV control mechanism is presented in Section 3,
limit control is presented in Section 2, the proposed PV control mechanism is presented in Section 3,
the experimental tests are demonstrated and discussed in Section 4 and the concluding remarks are
the experimental tests are demonstrated and discussed in Section 4 and the concluding remarks are
given in Section 5.
given in Section 5.

2. 2.
PVPV Limit
Limit Control
Control Principles
Principles and
and State
State of of
thethe
ArtArt
Figure
Figure 1 shows
1 shows thethe typical
typical power-versus-voltage
power-versus-voltage (P-V)
(P-V) characteristic
characteristic of of
PVPV panels.
panels. TheThe MPPT
MPPT
algorithms
algorithms focus
focus on on finding
finding thethe unique
unique maximum
maximum power
power point
point (MPP),
(MPP), which
which is indeed
is indeed floating
floating due todue
to the variable operating conditions. In the microgrid context, the storage and the
the variable operating conditions. In the microgrid context, the storage and the grid connection can grid connection can
absorb the excessive PV generation to some extent. However, in a microgrid where
absorb the excessive PV generation to some extent. However, in a microgrid where the PV generation the PV generation
is is designed
designed to to
bebethethe major
major power
power supply,
supply, thethe storage
storage andand grid
grid connection
connection sizing
sizing is is limited
limited duedue
to to
economic constrains. Thus, their capability is also limited and PV shedding capability
economic constrains. Thus, their capability is also limited and PV shedding capability is often required is often
in required in order
order to adapt to adapt
to the to the load
load demand demand
[26,27]. [26,27].sometimes
Therefore, Therefore,thesometimes thealgorithm
PV control PV control algorithm
should be
should be switched from MPPT mode to PV limiting mode immediately, reducing
switched from MPPT mode to PV limiting mode immediately, reducing the PV generation to keep the the PV generation
to keep
power the power
balance in the balance
microgrid.in the microgrid.

Figure 1. Power-versus-voltage (P-V) characteristic of a photovoltaic (PV) panel.

Figure 1. Power-versus-voltage (P-V) characteristic of a photovoltaic (PV) panel.
Like the MPPT methods, the instant PV power can be reduced by adjusting the output voltage
Like the MPPT methods, the instant PV power can be reduced by adjusting the output voltage
and current. Some effective methods for PV limit control have been presented in previous research
and current. Some effective methods for PV limit control have been presented in previous research
work. In [24], the authors described a method to control a storage-free PV system with a given power
reference. It is based on the voltage P&O algorithm and the system tends to converge towards a
higher voltage step by step when the power reference is less than the maximum power. Ahmed et al.
 Energies 2019, 12, 1233 3 of 12

work. In [24], the authors described a method to control a storage-free PV system with a given power
reference.
Energies 2019, 12,Itxis
FORbased
PEERon the voltage P&O algorithm and the system tends to converge towards a3 of
REVIEW higher
12
voltage step by step when the power reference is less than the maximum power. Ahmed et al. [28]
presented
[28] presenteda aPV PVcontrol
control method
method basedbased on onaavariable-step-size
variable-step-sizeincremental
incremental conductance
conductance algorithm,
algorithm,
which supports the MPPT operating mode and the reduced
which supports the MPPT operating mode and the reduced power operating mode. When passing power operating mode. When passingto to
thethe reduced
reduced power
power mode,
mode, thethe
PVPV voltage
voltage increases
increases following
following thethe perturbation
perturbation step.
step. A decentralized
A decentralized
power
power management
management system
system waswas designed
designed by Mahmood
by Mahmood et al.et[29]
al. [29]
for afor a hybrid
hybrid PV-battery
PV-battery system.system.
It
It includes
includes a conventional
a conventional MPPTMPPT algorithm
algorithm and aand a control
control loopcan
loop that thatpause
can pause
MPPTMPPT and move
and move the PVthe
PV operating
operating point to point to a voltage
a higher higher voltage
when the when the is
battery battery is fully or
fully charged charged or thecharging
the battery battery power
charging
power is too large. In research by Sangwongwanich et al. [30],
is too large. In research by Sangwongwanich et al. [30], four different constant PV generation four different constant PV generation
strategies
strategies werewere tested
tested in in experimental
experimental tests,
tests, including
including thethe power
power control
control method,
method, thethe current
current limit
limit
method
method and andtwo two P&OP&Oalgorithms
algorithmsconverging
convergingtotothe theleft
left and
and right side of of the
theMPP.
MPP.TheirTheirdynamic
dynamic and
static performance, complexity and stability
and static performance, complexity and stability were compared. were compared.
It should
It should be be noted
noted that
that in these
in these articles,
articles, as well
as well asmost
as in in most literature,
literature, PVPV power
power limiting
limiting is realized
is realized
by by increasing
increasing thethe PVPV voltage.
voltage. This
This is partially
is partially because
because thethe slope
slope onon theright
the rightside
sideofofthe
theMPP
MPPininFigure
Figure 1
1 isis steeper,
steeper, thus itit can canbe beexpected
expectedtoto have
have better dynamic
better dynamic performance.
performance. Moreover,
Moreover, highhigh
voltage would
voltage
lead lead
would to lesstocurrent and thus
less current and fewer ohmicohmic
thus fewer losses.losses.
However, in the aforementioned
However, in the aforementioned work, the operation
work, the
voltage range of the converter was not considered. As each converter
operation voltage range of the converter was not considered. As each converter has an acceptable has an acceptable input voltage
range,
input defined
voltage range, as the operational
defined range in Figure
as the operational range1,inthe lower
Figure andlower
1, the upperand voltage
upperbounds
voltagecorrespond
bounds
to different
correspond to sheddable power. Therefore,
different sheddable it is essential
power. Therefore, it isto considertothese
essential impacts
consider theseinimpacts
real applications.
in real
Moreover, Moreover,
applications. the effect on thethe converter
effect efficiency efficiency
on the converter has not been hasthoroughly explored.explored.
not been thoroughly Even though Eventhe
ohmicthe
though losses
ohmic arelosses
reduced, a high input
are reduced, voltage
a high inputcan lead to
voltage canhigh
leadswitching losses; thus,
to high switching the total
losses; thus,power
the
losses
total powercan increase
losses can in some cases.
increase in This
somecan also impact
cases. This can thealso
economic
impactbenefits of the installed
the economic benefitsPV ofsystem.
the
Hence,PV
installed research
system. work thatresearch
Hence, considers all the
work thatabove pointsallisthe
considers warranted.
above points is warranted.

3. Proposed
3. Proposed PVPV Control
Control Mechanism
Mechanism
The
The purpose
purpose ofofthis
thiswork
workisis to
to conceive
conceive aaflexible
flexiblePV
PVcontrol method
control methodforfor
a DC microgrid,
a DC as shown
microgrid, as
in Figure
shown 2. It should
in Figure be ablebetoable
2. It should operate in both in
to operate MPPT
bothand power
MPPT andlimiting
power modes,
limitingand can switch
modes, and canfrom
one to
switch another
from one totoanother
respondtotorespond
the dynamic
to thepower
dynamicdemand
powerofdemand
the system.
of the system.

Figure
Figure 2. Simplified
2. Simplified presentation
presentation of aofDC
a DC microgrid.
microgrid.

In order to realize simple and robust power control, the proposed PV power limiting control is
In order to realize simple and robust power control, the proposed PV power limiting control is
realized by a PI controller which receives the error between the power reference p∗PV_LI
* M
and the actual
realized by a PI controller which receives the error between the power reference pPV _ LIM and the
PV power p PV . The limited PV voltage reference can be expressed as in Equation (1). The proportional
actual PV power p PV . The limited PV voltage reference can be expressed as in Equation (1). The
proportional gain (Kp) and the integral gain (Ki) determine the dynamic of limited power control and
their values can be tuned in a classical way.
* *
vPV _ LIM = Kp ( pPV − pPV ) ⋅ (1 + Ki ) (1)
 Energies 2019, 12, 1233 4 of 12

gain (Kp) and the integral gain (Ki) determine the dynamic of limited power control and their values
can be tuned in a classical way.

Energies 2019,
Energies 2019, 12,
12, xx FOR
FOR PEER
PEER REVIEW
REVIEW v∗PV_LI M = Kp( p∗PV − p PV )·(1 + Ki ) 44 of
of 12
12(1)

However,
However, this
However,this voltage
thisvoltage reference
voltagereference is only
reference is
is onlyvalid
only validinin
valid inPV
PV
PV power
power
power limiting
limiting
limiting mode.
mode.
mode. During
During
During the MPPT
the the
MPPTMPPT
mode,
**
mode,
mode, the maximum
the maximum
the maximum allowed
allowed
allowed PV
PV power p ∗
PV power
power ppPV
PVis__LIM is greater
is greater
greater
LIM than actual
than
than actual
PV_LI M
actual PV
PV power p PV andppthe
PV power
power and the
PV and
PV
the
difference
betweenbetween
difference
difference them accumulates
between them in the PI in
them accumulates
accumulates controller.
in the
the PI As a result,
PI controller.
controller. Asthis
As calls this
aa result,
result, for
thisan anti-windup
calls
calls for an
for to preventto
an anti-windup
anti-windup tothe
saturation
prevent
prevent the of the integrator.
the saturation
saturation of the
of The saturation
the integrator.
integrator. block makes
The saturation
The saturation suremakes
block
block that this
makes sure
sure voltage reference
that this
that this voltageisreference
voltage within the
reference
is functional
within the range of the
functional system,
range of and
the the windup
system, and is
thecanceled
windup by
is the gain
canceled (Kc).
by The
the entire
gain
is within the functional range of the system, and the windup is canceled by the gain (Kc). The entire power
(Kc). The limiting
entire
control
power
power is depicted
limiting
limiting control
controlinis
Figure
is 3. in
depicted
depicted in Figure
Figure 3.
3.

Figure
Figure 3. The
3.
Figure The PVPV
3. The
PV power limiting
power
power controller.
limiting
limiting controller.
controller.

The
The
The PVPV
PV power
power
power limitlimit
limit is is imposed
is imposed
imposed by bymicrogrid
by the
the the microgrid
microgrid central
central central controller
controller
controller it and
and it
and it can
can vary
can vary atvary
at at any
any time.
any time.
time.
Hence, Hence,
it demands it demands
seamless seamless
switchingswitching
between between
the MPPT the MPPT
control control
and the
Hence, it demands seamless switching between the MPPT control and the PV limited control. The and
PV the PV
limited limited
control.control.
The
The global
global PV
global PV
PV control control
control mechanism mechanism
mechanism is is presented
is presented
presented in in Figure
in Figure
Figure 4. 4. It 4.
It is It is
is worth worth
worth notingnoting
noting that that
that the the
the P&O P&O
P&O MPPT MPPT
MPPT
algorithm
algorithm
algorithm is is chosen
is chosen
chosen in in
in thisthis
this work
work
work forfor
for itsits
its simplicity,
simplicity,
simplicity, but
butbut it can
itit can
can bebe
be replaced
replaced
replaced byby
by any
any
any other
other
other algorithm
algorithm
algorithm based
based
based
onon discrete
discrete steps,
steps, such
such as as the
the incremental
incremental conductance
conductance
on discrete steps, such as the incremental conductance algorithm. algorithm.
algorithm.

Figure
Figure 4. Seamless
4.
Figure Seamless switching
switching
4. Seamless PVPV
PV
switching control
control topology
topology
control in “max”
in
topology “max”
in configuration.
configuration.
“max” configuration.

**∗
While
While
While performing
performing
performing the
thethe MPPToperation,
MPPT
MPPT operation,i.e.,
operation, i.e., pppPV
i.e., PV_LI
PV LIM M
__LIM >> pp> p the
PV ,, PV
PV
, the
the input
input
input of
thethe
of the
of PI PI
PI controller
controller
controller is is
is
positive. By selecting a negative value of Kp, the value of the generated v ∗ is always less
**
positive. By
positive. By selecting
selecting aa negative
negative valuevalue of Kp ,, the
of Kp the value
value of of the generated vvPV
the generated is always
PV_LI
LIM is
PV __LIM always
M less than
less than
than v ∗ . Consequently, v ∗ is sent as the final v ∗ to the associated PV control loop to
** PV_MPPT ** PV_MPPT ** PV
vvPV
PV _ MPPT
command
_ MPPT .. Consequently,
Consequently, vv
the PV converter. PV _ MPPT is
is sent
sent as
as the
the final
final
Conversely, if the power limiting
PV _ MPPT vv PV
PV
to
to the
the associated
associated PV
PV control
control
∗
mode is active and p PV_LI M < p PV , loop
loop to
to
**
v∗PV_LI M the
command
command the PV
will PV converter. be
consequently
converter. Conversely,
greater than
Conversely, if v∗PV_MPPT
if the
the power; limiting
power thus, v∗PV_LI
limiting mode
mode is active
M will
is active
be sent and
and ppPV
to the LIM << ppPV
PV _control
_LIM PV ,,
loop.
* Moreover, in order to prevent the P&O algorithm from
*
vvPV will consequently
LIM will
PV __LIM consequently be be greater
greater thanthan vvPV**
thus,being
MPPT ;; thus,
PV __MPPT
*
vvPV
* disturbed, bethe
will be
LIM will
PV __LIM sent
sent MPPT
to the
to thealgorithm
control loop.
control should
loop.
be paused in power limiting mode. In conclusion, the final PV voltage reference can be expressed as
Moreover, in
Moreover, in order
order to
to prevent
prevent the the P&O
P&O algorithm
algorithm from from being being disturbed,
disturbed, the the MPPT
MPPT algorithm
algorithm should should
shown in Equation (2). This configuration is called the “max” configuration.
be paused
be paused in in power
power limiting
limiting mode.
mode. In In conclusion,
conclusion, the the final
final PV PV voltage
voltage reference
reference can can bebe expressed
expressed as as
shown in Equation (2). This
shown in Equation (2). This configuration configuration is
v∗ =ismax called
called the
(v∗ the “max”“max” configuration.
, v∗ configuration.
) (2)
PV PV_MPPT PV_LI M
** ** **
vvPV
PV == max(
max(vvPV MPPT ,,v
PV __MPPT vPV LIM ))
PV __LIM (2)
(2)
The proposed mechanism can be easily reconfigured to obtain the minimal value between
v∗PV_MPPT
The and v∗PV_LI
The proposed
proposed mechanism can be
M . By altering
mechanism can be theeasily
sign
easily reconfigured
of Kp, one to
reconfigured to obtain
can obtain
force the ∗ minimal
vthe
PV_LI value
M to bevalue
minimal between
less than
betweenVMPP .
** Moreover, the** direction of the relational operator needs to be changed **
vvPV and vvPV
MPPT and
PV __MPPT LIM .. By
PV __LIM By altering
altering the
the sign
sign of Kp ,, one
of Kp one can force vvPV
can force from “greater
to be
LIM to
PV __LIM be less than VVMPP
than”
less than to “less
MPP ..
than” as well.
Moreover, A configuration
the direction
direction of the such as this
the relational
relational is depicted
operator needs in toFigure 5 and is from
be changed
changed called“greater
the “min” configuration.
than” to “less
“less
Moreover, the of operator needs to be from “greater than” to
than” as
than” as well.
well. A
A configuration
configuration such
such as
as this
this is
is depicted
depicted in
in Figure
Figure 55 and
and is
is called
called the
the “min”
“min”
configuration.
configuration.
 Energies
Energies 2019,
2019, 12,
12,12, x FOR PEER REVIEW
x FOR 5 of 12
Energies 2019, 1233 PEER REVIEW 5 of 1212
5 of
Energies 2019, 12, x FOR PEER REVIEW 5 of 12

Figure
Figure 5. 5. Seamless
Seamless switching
switching PVPV control
control topology
topology ofof
thethe “min”
“min” configuration.
configuration.
Figure 5. Seamless
Figure switching
5. Seamless PVPV
switching control topology
control of the
topology “min”
of the configuration.
“min” configuration.
Bothconfigurations
Both configurationscancanachieve
achieveseamless
seamlessswitching
switchingbetweenbetweenthe theMPPT
MPPTand andpower
powerlimiting
limiting
modes.
modes. Both
Both However,
However, themaximum
the
configurations maximum
configurationscancan sheddable
achieve seamless
sheddable
achieve seamlessPV PV powercan
switching
power
switching can bebedifferent,
between different,
between thethe asstated
MPPT
as
MPPT stated
andand
ininthethepreceding
power
power preceding
limiting
limiting
section.
modes.
section.
modes. The
The experimental
However,thethe
experimental
However, validation
maximum
validation
maximum and
sheddablecomparisons
PVPV
and comparisons
sheddable power
power are
cancan
are conducted
be be
conducted in
different,the
as as
in the
different, next section.
stated
next
stated in in
thethe
section. preceding
preceding
section.
section. TheThe experimental
experimental validation
validation andand comparisons
comparisons areare conducted
conducted in in
thethe next
next section.
section.
4. 4. Experimental
Experimental Validation
Validation
4. Experimental
4. Experimental Validation
Validation
The
The experimental
experimental validation
validation was
was carried
carried out
out onon the
the technological
technological platform,
platform, which
which is is located
located onon
thethecampus
campus
TheThe ofofUniversité
Université
experimental
experimental dedeTechnologie
validationTechnologie
validationwaswas dedeCompiègne
carried
carried Compiègne
outout
ononthethe inCompiègne,
Compiègne,
intechnological
technological France,
platform,
France,
platform, asshown
shown
aswhich
which is in inFigure
Figure
located
is located onon
the 6.
6. campus of Université de Technologie de Compiègne in Compiègne, France, as shown in Figure 6.
the campus of Université de Technologie de Compiègne in Compiègne, France, as shown in Figure
6.

Figure
Figure
Figure 6.
PVPV
6. 6. PV generator
generator
generator and
and control
control
and room
room
control of
ofof
room the
the technological
technological
the platform.
platform.
technological platform.
Figure 6. PV generator and control room of the technological platform.
The electricscheme
Theelectric
electric scheme
scheme ofof
of thethe
thePVPVPV subsystem,
subsystem, i.e.,the
i.e.,i.e.,
subsystem, the PV the PVgenerator
generator
generator
PV anditsitspower
and itsand
power powerconverter,
converter,
converter, is is
is presented
presented
in Figure
presented 7in Figure
and some
in Figure
The electric 7 and
of thesome
7 andofsome
scheme of the
parameter
the of
PVthe parameter
values are
parameteri.e.,
subsystem, values
given
values are
in
the are given
Table
PV given 1. in Table
in Table
generator 1.
and1.its power converter, is
presented in Figure 7 and some of the parameter values are given in Table 1.

Figure
Figure 7. 7.
PVPV subsystem
subsystem scheme.
scheme.
Figure
Table 7. PV
1. PV subsystem
subsystem scheme. values.
parameter
Table
Table 1. 1.
PVPV subsystem
subsystem parameter
parameter values.
values.
Parameter
Table 1. PV subsystem parameter values. Value
Parameter
Parameter Value
Value
Number of PV panels SunPower SPR-X21-345 12
Number
Number of
of PVPV panels
panels SunPower SPR-X21-345 1212
Maximum power pointSunPower
Parameter(MPP) voltage SPR-X21-345
for each panel Value
57.3 V
Maximum
Maximum
Number power
of power
PVMPP
panels point
point (MPP)
(MPP)
SunPower
current for each voltage
voltage for
panel for
SPR-X21-345 each
each 57.3
57.3
12 A
6.02 VV
Open circuit
Maximum power panel
voltage
(MPP)for
panel
point each panel
voltage for each 68.2 V
57.3 V
Short
MPP circuit
current
MPP current current
for for
each
for each panel each panel
panel 6.39 A
6.02 AA
6.02
panel
Open circuitLCL filter sizing
voltage L1 = 930 µH, C =68.2
10 µF, L2 = 850 µH
Open
MPPcircuit voltage
current for each
Switching
forfor each
each
panel
frequency
panel
panel 68.2AV V
6.02
14 kHz
Short circuit
Shortcircuit
Open current
circuitvoltage
currentfor for
foreach each panel
eachpanel
panel 6.39
6.39
68.2 VAA
LCL
LCL
Short circuit filter
filter
current sizing
sizing
for each panel L1 = 930 μH, C
L1 = 930 μH, C6.39=
= 10A10 μF,
μF, L2L2 = 850
= 850 μHμH
LCL filter sizing L1 = 930 μH, C = 10 μF, L2 = 850 μH
Energies 2019, 12, x FOR PEER REVIEW 6 of 12

Switching frequency 14 kHz

Energies 2019, 12, 1233 6 of 12
The PV generator is composed of 12 PV panels connected in series, which make up to 4.14 kW
under Standard Test Conditions (STC, solar irradiance of 1000 W/m2 and panel temperature of 25 °C).
The PV generator is composed of 12 PV panels connected in series, which make up to 4.14 kW
Since its MPP voltage is 687.6 V under STC, there is a three-phase interleaved buck converter
under Standard Test Conditions (STC, solar irradiance of 1000 W/m2 and panel temperature of 25 ◦ C).
(Triphase PM15F03P02 15kW) that interfaces the PV generator and the DC bus. The bus voltage is set
Since its MPP voltage is 687.6 V under STC, there is a three-phase interleaved buck converter (Triphase
to 400 V because of the low line losses and the compatibility with the existent infrastructure [31]. The
PM15F03P02 15kW) that interfaces the PV generator and the DC bus. The bus voltage is set to 400 V
power is equally allocated to each phase, and all the phases are equipped with a dedicated LCL filter,
because of the low line losses and the compatibility with the existent infrastructure [31]. The power is
which consists of two inductors and one capacitor. This LCL filter has better filtering performance
equally allocated to each phase, and all the phases are equipped with a dedicated LCL filter, which
than the classical LC filter which consists of a single inductor and a capacitor, and the size of the two
consists of two inductors and one capacitor. This LCL filter has better filtering performance than the
inductors is usually much smaller than that of a single bulk filter, although it calls for a special control
classical LC filter which consists of a single inductor and a capacitor, and the size of the two inductors
algorithm as described in Vanassche [32]. The control algorithm is programmed in
is usually much smaller than that of a single bulk filter, although it calls for a special control algorithm
MATLAB/Simulink and then implemented in the real-time target, which commands all of the
as described in Vanassche [32]. The control algorithm is programmed in MATLAB/Simulink and then
physical converter components.
implemented in the real-time target, which commands all of the physical converter components.
Due to the characteristics of the buck converter, the PV output voltage cannot be less than the
Due to the characteristics of the buck converter, the PV output voltage cannot be less than the DC
DC bus voltage. In addition, the maximum input voltage of this converter is fixed at 760 V with the
bus voltage. In addition, the maximum input voltage of this converter is fixed at 760 V with the aim
aim of being compatible with the aforementioned PV generator. Evidently, this voltage range of 400
of being compatible with the aforementioned PV generator. Evidently, this voltage range of 400 V to
V to 760 V is satisfactory for MPPT control.
760 V is satisfactory for MPPT control.
4.1. Static Characteristic of the PV Panels
4.1. Static Characteristic of the PV Panels
First, the static characteristic of the PV generator is tested. Figure 8 shows the P-V characteristic
First, the static characteristic of the PV generator is tested. Figure 8 shows the P-V characteristic
under certain typical function points, where g indicates the solar irradiance and  represents the
under certain typical function points, where g indicates the solar irradiance and θ represents the PV
PV
cellcell temperature.
temperature.

Figure
Figure8.8.Static
Staticpower-versus-voltage
power-versus-voltagecharacteristic
characteristicofofthe
thePV
PVgenerator.
generator.

Duetotothe
Due theaforementioned
aforementionedvoltagevoltageconstraint,
constraint,the thetested
testedPVPVvoltage
voltagevaried
variedfrom
from400400 VVtoto 760
760 V.V.
Thedotted
The dottedlines
linesininthe
theleft
leftpart
partofofthe
thegraph
graphare arejust
justlinear
linearextrapolations
extrapolationsbetween
betweenthe theoriginal
originalpoint
point
andthe
and therespective
respectivemeasured
measuredvalues
valuesatat400
400V.V.ItItcan
canbebeseen
seenthat
thatall
allofofthe
thecurves
curveshavehavethethesame
sameformform
thatwas
that waspresented
presentedininFigure
Figure1.1.The
Thecurves
curvesimplyimplythatthatthetheMPP
MPPvoltage
voltageisisgenerally
generallylesslessthan
than700
700V,V,
whichjustifies
which justifiesthe
thechoice
choiceofofthe
theconverter
converterinput
inputvoltage
voltagerange.
range.
Thestatic
The static performance
performanceofofPV PVpower
power limiting
limitingcontrol can also
control can be observed
also in this figure.
be observed in this PV shedding
figure. PV
can be realized
shedding can beby both decreasing
realized and increasing
by both decreasing the MPP voltage.
and increasing the MPP From the difference
voltage. From thebetween
differencethe
minimal powers on the left and right sides, one can note the different shedding
between the minimal powers on the left and right sides, one can note the different shedding capacities capacities of the two
ofconfigurations. On the one
the two configurations. Onhand,
the one thehand,
power thevalues
poweratvalues760 Vatare
760allVclose to close
are all zero, tomeaning that the
zero, meaning
“max”
that the configuration can shedcan
“max” configuration almost
shed 100%
almost of100%
the PV ofpower.
the PV On the other
power. On thehand,
otherthe power
hand, thevalues
powerat
400 V are at least about 70% of the maximum power, meaning that the “min”
values at 400 V are at least about 70% of the maximum power, meaning that the “min” configuration configuration can only
shed
can 30%
only of PV
shed 30% power.
of PV power.

4.2.Dynamic
4.2. DynamicPerformance
Performanceofofthe
theFlexible
FlexiblePV
PVPower
PowerControl
ControlMechanism
Mechanism
Thedynamic
The dynamicperformance
performanceofofthe
theproposed
proposedflexible
flexiblePV PVcontrol
controlmechanism
mechanismwaswasvalidated
validatedunder
under
realconditions.
real conditions.Its
Itsparameters
parametersare
areshown
shownininTable
Table2.2.ItItshould
shouldbe
benoted
notedthat
thatthe
theP&O
P&OMPPT
MPPTalgorithm
algorithm
works with a frequency of 1 kHz, which is a good tradeoff between the convergence speed and stability.
Meanwhile, the PV power limiting controller works synchronously with the main real-time system, i.e.,
 Energies 2019, 12, x FOR PEER REVIEW 7 of 12

works with a frequency of 1 kHz, which is a good tradeoff between the convergence speed and
stability. Meanwhile, the PV power limiting controller works synchronously with the main real-time
Energies 2019, 12, 1233 7 of 12
system, i.e., 14 kHz, because as a continuous PI controller the high frequency allows for good dynamic
performance. The values of Kp, Ki, and Kc in Table 2 are steady examples obtained by experimental
14tests
kHz,rather
because as aoptimal
than continuous PI controller
values. the high frequency
The fine-tuning allows for good
of these parameters can dynamic
improveperformance.
the dynamic
The values of Kp, Ki, and Kc in Table 2 are steady
performance, but that is not the subject of this work. examples obtained by experimental tests rather than
optimal values. The fine-tuning of these parameters can improve the dynamic performance, but that is
not the subject of this work. Table 2. Parameters for the flexible PV controller.
Parameter
Table 2. Parameters for the flexible PV controller.
Value
MPPT perturbation step size 25 mV
Parameter frequency
MPPT perturbation Value 1 kHz
MPPT perturbation step size Kp = −0.01 in25 case
mV of “max” configuration
MPPT perturbation frequency Kp = +0.01 in case of “min” configuration
1 kHz
PV limited power control parameters
Ki = configuration
Kp = −0.01 in case of “max” 500
Kc =configuration
Kp = +0.01 in case of “min” 100
PV limited power control parameters
Data log sample frequency Ki = 500100 Hz
Kc = 100
Figure 9 shows the
Data log results
sample of a 90 second test of the “min” configuration,
frequency 100 Hz carried out on July 25,
2018, where pPV _ LIM is arbitrarily set to be constant at 2000 W. In Figure 9a, the continuous variance
Figure
of the solar9irradiance
shows theandresults of atemperature
PV cell 90 second test canofbethe “min”
seen. configuration,
Figure 9b shows that carried out on
the MPPT 25 July
algorithm
2018,
is inwhere
control p PV_LI M isthe
before arbitrarily
imposed setPV
to be constant
power limitatis2000 W. In Figure
reached. The PV 9a,generation
the continuous variance
increases as of
the
the solar irradiance
irradiance goes up. andAsPVsoon
cell temperature
as pPV andcan pPVbe_ LIM
seen. Figure
meet, the9bPV
shows
powerthatlimiting
the MPPT algorithm
controller is
takes
incontrol.
control In
before
Figurethe9c,
imposed
a suddenPVdrop
power of limit is reached.
voltage The PV
can be noted. Atgeneration increases
the beginning of thisasprocess,
the irradiance
a sharp
goes up. As soon as p PV and p PV_LI M meet, the PV power limiting controller
overshooting image can be observed in Figure 9b, which is caused by switching between takes control. In Figure
the MPPT9c,
a algorithm
sudden drop of voltage can be noted. At the beginning of this process, a sharp overshooting
and the PI limited power controller. Also, the PV power is well constrained at 2000 W most image
can be observed in Figure
of the time, despite some9b,oscillations.
which is causedOn thebyother
switching
hand,between the MPPT
pPV exceeds the algorithm
limitation and
for athe PI
while
limited power controller. Also, the PV power is well constrained at 2000 W most of the time, despite
between the 34th second and 51st second. During this period, the PV voltage evolution (Figure 9c)
some oscillations. On the other hand, p PV exceeds the limitation for a while between the 34th second
proves that the voltage drops, and the system is saturated at the lower limit of 400 V. In this case, the
and 51st second. During this period, the PV voltage evolution (Figure 9c) proves that the voltage
“min” configuration can shed no more power. In this validation test, the excessive power is absorbed
drops, and the system is saturated at the lower limit of 400 V. In this case, the “min” configuration can
by the other components in the microgrid; thus, in Figure 9c the bus voltage vBUS is kept at 400 V.
shed no more power. In this validation test, the excessive power is absorbed by the other components
However, in a situation where the other components can absorb no more excessive power, this can
in the microgrid; thus, in Figure 9c the bus voltage v BUS is kept at 400 V. However, in a situation where
cause severe accidents.
the other components can absorb no more excessive power, this can cause severe accidents.

(a) (b)
Figure 9. Cont.
Energies 2019, 12, 1233 8 of 12
Energies 2019, 12, x FOR PEER REVIEW 8 of 12

Energies 2019, 12, x FOR PEER REVIEW 8 of 12

(c)
(c)
Figure
Figure 9.
9. Experimental
Experimentalvalidation
validationof of the
the “min”
“min” configuration:
configuration: (a)(a) solar
solar irradiance
irradiance and
and PV
PV panel
panel
Figure 9. evolution;
temperature
temperature Experimental(b)validation
evolution; (b) real PV
real of the generation
PV power
power “min” configuration:
generation and PV
and (a) solar
PV power
power irradiance(c)
limitation;
limitation; and
(c) PV
PV
PV panel voltage
output
output voltage
andtemperature
and current evolution;
current with
with (b) real PV power generation and PV power limitation; (c) PV output voltage
bus voltage.
bus voltage.
and current with bus voltage.
A second
A second experimental
experimentaltest test was
was carried
carriedout out onon 24
24 July
July 2018
2018 for
for the
the “max”
“max” configuration
configuration where where
A second experimental test was carried out on 24 July 2018 for the “max” configuration where
the
the PVPV power limit was also set to 2000 W. By comparing Figure 10a,b, one can observe the PV
the power
PV power limit was
limit wasalso
also set
set to 2000 W.
to 2000 W.By Bycomparing
comparing Figure
Figure 10a,b,10a,b,
one canoneobserve
can observe
the PV the PV
generation
generation
generation
following
following
following
the
the solar
thesolar
irradiance
solar irradiance
evolution
evolution
irradiance evolution
during
during
during
the
the the
MPPT
MPPT
MPPT operating
operating
operating
mode.
mode.mode.
The power
The power
The power
limiting mode is activated when the solar irradiance exceeds approximately 500 W/m 2 . Accordingly,
limiting mode
limiting modeis activated
is activatedwhen
whenthe solarirradiance
the solar irradiance exceeds
exceeds approximately
approximately 500 .W/m
500 W/m 2 2 . Accordingly,
Accordingly,
in Figure
in Figure 10c,10c,
10c,
in Figure thethe
the sudden
sudden increase
suddenincrease
increase of of vvvPVPVcan
of canbebe
can beobserved
observed
observed at at the
at
the the moment
moment
moment of of
of
the the activation
the activation
activation of of the
of
the PV the PV
PV
PV
2
limiting mode.
limiting
limiting mode.mode. Moreover,vvPV
Moreover,
Moreover, vPV reaches
reaches 730 730VVwhen
whenthethe solar
solar irradiance
irradiance goes goes
up toupaboutto about
900 W/m 900
2, W/m2 ,
PV reaches 730 V when the solar irradiance goes up to about 900 W/m ,
meaning
meaningthat thethe “max”
“max”configuration can still increase
the the voltage to shed even more PV power if
meaning thatthat
the “max” configuration can
configuration canstill
stillincrease
increase voltage
the to shed
voltage even more
to shed even PV power
more PV ifpower if
needed. There
needed. Therewere
were some
someoscillations within±20
oscillations within ±20 W,W, which
which can can mostly
mostly be observed
be observed in power in limiting
power limiting
needed. There were some oscillations within ±20 W, which can mostly be observed in power limiting
mode.mode. These
These oscillationsare
oscillations are relatively
relatively smallsmall(about
(about 1%)1%)
compared
compared to thetoactual power,power,
the actual and therefore
and therefore
mode. These oscillations are relatively small (about 1%) compared to the actual power, and therefore
can can be considered
be considered acceptable.
acceptable.
can be considered acceptable.

(a) (b)

(a) (b)

(c)

Figure 10. Experimental validation of the “max” configuration: (a) solar irradiance and PV panel
temperature evolution; (b) real PV power generation and PV power limitation; (c) PV output voltage
and current with bus voltage.
(c)
Energies 2019, 12,
Figure 10.xExperimental
FOR PEER REVIEW
validation of the “max” configuration: (a) solar irradiance and PV panel9 of 12
temperature evolution; (b) real PV power generation and PV power limitation; (c) PV output voltage
Figure
and 10. Experimental
current validation of the “max” configuration: (a) solar irradiance and PV panel
with bus voltage.
temperature evolution; (b) real PV power generation and PV power limitation; (c) PV output voltage
and
Two
Energies current
other
2019, with
testsbus
12, 1233 voltage.
were carried out on 19 and 20 March 2019 for 2 h, as shown in Figure 119and of 12
Figure 12. The PV power limits were respectively 2500 W and 2000 W. By comparing PV power and
solarTwo other tests
irradiance, were
it can be carried outthe
seen that oneffectiveness
19 and 20 March 2019
of both for 2 h, as shown
configurations in Figure
was proved for11daily
and
FigureTwo12. other
The tests
PV were
power carried
limits out
were on 19 and
respectively 20 March
2500 W 2019
and for
2000 2 h,
W. as
By shown in
comparing
operation. Moreover, the seamless switching between the MPPT and power limiting modes worked Figures
PV 11
powerand 12.
and
The PV power
solarwithout
well irradiance, limits were
it canissues,
stability respectively
be seen that
even 2500
the the
when W and
effectiveness 2000
irradiance of W. By comparing
both configurations
changed suddenly. PV power and
was proved for dailysolar
irradiance,Moreover,
operation. it can be seen
the that the effectiveness
seamless of both the
switching between configurations
MPPT and powerwas proved for modes
limiting daily operation.
worked
Moreover,
well without thestability
seamless switching
issues, even between
when thethe MPPT and
irradiance power suddenly.
changed limiting modes worked well without
stability issues, even when the irradiance changed suddenly.

Figure 11. Two-hour experimental test of the “min” configuration.

Figure 11. Two-hour experimental test of the “min” configuration.
Figure 11. Two-hour experimental test of the “min” configuration.

Figure 12. Two-hour experimental test of the “max” configuration.

Figure 12. Two-hour experimental test of the “max” configuration.
In the above tests, the dynamic performance of both configurations were validated under real
In the aboveBoth Figure
tests, 12.
theable Two-hour
dynamic experimentaloftest
performance of the
both “max” configuration.
configurations
circumstances. are to switch between the MPPT mode and were validated
the power undermode,
limiting real
circumstances. Both are able to switch between the MPPT mode and
according to the given power limit. However, the “min” configuration has limited shedding capacitythe power limiting mode,
In thetoabove
according tests,power
the previous
given the dynamic performance
limit.InHowever, of both
the “min” configurations
configuration has were
limitedvalidated
shedding under real
capacity
as stated in the section. the microgrid context, failure to reach the given PV generation
circumstances.
as stated Both are able to switch between the MPPT mode and the power limiting mode,
limit mayinbreak
the previous
the power section.
balanceIn of
thethemicrogrid
microgrid context,
and the failure to reach
PV systems the given
should PV generation
generally be cut off
according
limit may to the given
break the power
power limit. However,
balance of the the “min”
microgrid and configuration
the PV systemshasshould
limitedgenerally
sheddingbecapacity
cut not
off
entirely from the microgrid to prevent serious safety issues. Obviously this is a last resort and
as stated
entirely in the previous section. In the microgrid context, failure to reach the given PV generation
the bestfrom
optiontheasmicrogrid
it is not onlyto prevent
a waste serious safety solar
of available issues.energy,
Obviously this is
but may a last
also resort
cause and not
stability the
issues
limitoption
best may breakas it the
is power
not only balance
a waste ofavailable
of the microgrid
solar and the but
energy, PV may
systems
also should
cause generally
stability be cut
issues off
since
since the other sources in the microgrid may not have enough power to supply the loads without
entirely
the from the microgrid to prevent serious safety issues. Obviously this is a last resort and not the
the other sources
PV sources. in simple
A the microgrid
solution mayis tonotmake
have enough
the capacitypower to supply
of the energythe loads
storage without
and thegrid
public PV
best option
sources. as it is not only a waste of available solar energy, but may also cause stability issues since
injectionA simple solution
oversized, but thatiscallsto make the capacity
for greater investment of the energy storage
in equipment. and solution
A better public gridis toinjection
conceive
the other sources
oversized, but thatin calls
the microgrid
for greatermay not have in
investment enough power A
equipment. to better
supplysolution
the loads is without
to the PV
conceive
an intelligent microgrid control, which is able to limit the PV shedding power according to the an
PV
sources. A microgrid
intelligent simple solution control, is to makeis the
which ablecapacity
to limit ofthe
thePV energy storage
shedding powerand according
public gridtoinjection
the PV
generation forecast and load power prediction. In this way, the excessive power can be absorbed by the
oversized, but
generation that and
forecast callsload
for power
greaterprediction.
investmentIninthis equipment.
way,tothe A better solution
excessive is be
to absorbed
conceive by an
energy storage and public grid connection without having make any ofpower can
the components oversized.
intelligent microgrid control, which is able to limit the PV shedding
the energy storage and public grid connection without having to make any of the components power according to the PV
generation
4.3. Energy forecast
oversized. Efficiencyand load power prediction. In this way, the excessive power can be absorbed by
Analyses
the energy storage and public grid connection without having to make any of the components
In the previous experimental results, it was proved that the “min” configuration causes the current
oversized.
to rise and the “max” configuration causes the voltage to rise in power limiting mode. A high current
leads to high ohmic losses and a high voltage leads to high switching losses [33]. Hence, experimental
tests were carried out to measure the integral energy efficiency of the PV converter at different voltages.
 current to rise and the “max” configuration causes the voltage to rise in power limiting mode. A high
current leads to high ohmic losses and a high voltage leads to high switching losses [33]. Hence,
experimental tests were carried out to measure the integral energy efficiency of the PV converter at
different voltages.
EnergiesFor this
2019, 12, investigation
1233 the PV generator was set to work with some specific voltages. The10input
of 12
power of the PV converter is pPV  vPV  iPV and the output power pPV _ OUT is the product of the DC
bus For
voltage and the output
this investigation thecurrent of the converter
PV generator was set to iwork . Thus,
PV _ OUTwith thespecific
some converter efficiency
voltages. The can be
input
power of the PVvBUS  iPV _ OUT is p PV = v PV × i PV and the output power p PV_OUT is the product of the
converter
defined as   . From the efficiency curves shown in Figure 13, it can be noted that all of
DC bus voltage and vPVthe iPVoutput current of the converter i PV_OUT . Thus, the converter efficiency can
v BUS ·i PV_OUT
be
thedefined as η =
efficiencies increase when
v PV ·i PV . From
pPV the efficiency
increases, andcurves
most shown in Figure
importantly that 13,
theitconverter
can be noted
at the that all
same
of the efficiencies
power increase
is more efficient whenwhen thep PV increases,
voltage and At
is lower. most importantly
medium powerthat the
(2000 converter
W), at the same
the efficiency at 450
power is more efficient when the voltage is lower. At medium power (2000 W),
V is 7% higher than that at 740 V, and this difference becomes more significant when the power the efficiency at 450 Vis
islower.
7% higher than that
This means that atin740 V, andtothis
regard thedifference becomes
buck converter, themore
“min”significant when which
configuration, the power is lower.
converges at
This means that in regard to the buck converter, the
the low voltage, is favorable in the view of energy efficiency. “min” configuration, which converges at the low
voltage, is favorable in the view of energy efficiency.

Figure 13. Energy efficiency comparison of the power converter under different values of v PV .
Figure 13. Energy efficiency comparison of the power converter under different values of vPV .
From the above experimental results, it can be seen that PV power shedding is feasible both by
Fromand
increasing the decreasing
above experimental results,Both
the PV voltage. it can be seen that of
configurations PVthepower shedding
proposed is feasible
PV control both by
mechanism
increasing
can and decreasing
switch seamlessly between the PV
thevoltage.
MPPT and Boththeconfigurations
power limiting ofmode,
the proposed
but theyPV control
have mechanism
different power
can switch seamlessly between the MPPT and the
shedding capacities and thus result in different levels of efficiency. power limiting mode, but they have different
power shedding capacities and thus result in different levels of efficiency.
5. Conclusions
5. Conclusions
With the aim of sustaining PV source integration in a microgrid, this work presents a simple
With the
and flexible PVaim of sustaining
power PV source integration
control mechanism which supports in a microgrid, this work
both classical MPPT presents
and PVageneration
simple and
flexible PV
shedding. Thepower
research control
includes mechanism
the physicalwhich supports
constraints andboth
energyclassical MPPT
efficiency andpower
of the PV generation
converter,
shedding. The research includes the physical constraints and energy
in terms of PV operation in a DC microgrid. Two configurations, which increase or decrease the efficiency of the power
PV
converter,
voltage whileinlimiting
terms of thePVPVoperation
generation, inwere
a DC microgrid.
proposed. Two configurations,
Experimental which out
tests were carried increase
on a PVor
decrease the PV voltage while limiting the PV generation, were proposed.
system with a buck converter and both configurations were validated. Results show that the “min” Experimental tests were
carried out oncan
configuration a PVshedsystem
only 30% withof a PV
buck converter
power whileandthe both
“max” configurations
configurationwere validated.
can shed nearlyResults
100%.
show thatthe
However, theformer
“min”has configuration
better energy canefficiency
shed only 30%
than theoflatter.
PV power while the “max” configuration
can shed nearly
This fact 100%. However,
suggests the former
that the choice between has the
better
twoenergy efficiency should
configurations than thebe latter.
well considered.
This fact
The “max” suggests that
configuration canthe bechoice
applied between
in casesthe two the
where configurations
power balance should be well
capacity considered.
of the microgridThe is
“max” configuration
relatively weak. The “min” can be applied in cases
configuration is more where the when
suitable powereconomic
balance capacity
costs are of thesensitive
more microgrid butis
relatively
an weak.
appropriate The “min”
microgrid configuration
control is needed is to more suitable
coordinate the when economic costs
load consumption andarePVmore sensitive
generation in
but an
order appropriate
to maintain microgridincontrol
PV shedding a feasibleis range.
needed to coordinate the load consumption and PV
generation in orderantointelligent
In conclusion, maintain microgrid
PV shedding in acan
control feasible range. which integrates the proposed PV
be developed
control mechanism and optimal power dispatching. In this developed
In conclusion, an intelligent microgrid control can be way, PV power which integrates
shedding canthe
beproposed
planned
PV control
within mechanism
a feasible range and and theoptimal
economic power dispatching.
benefits In this way, PV power shedding can be
can be maximized.
planned within a feasible range and the economic benefits can be maximized.
Author Contributions: All the authors contributed equally to this research work.
Funding: This research was funded within the national research project MOBEL_CITY, selected by ADEME, in the
context of the call for projects APRED 2017 (convention number #1766C0006).
Conflicts of Interest: The authors declare no conflict of interest. The funders had no role in the design of the
study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to
publish the results.
Energies 2019, 12, 1233 11 of 12

References
1. Comello, S.; Reichelstein, S.; Sahoo, A. The road ahead for solar PV power. Renew. Sustain. Energy Rev. 2018,
92, 744–756. [CrossRef]
2. Sechilariu, M.; Locment, F.; Wang, B. Photovoltaic Electricity for Sustainable Building. Efficiency and Energy
Cost Reduction for Isolated DC Microgrid. Energies 2015, 8, 7945–7967. [CrossRef]
3. El-Baz, W.; Tzscheutschler, P.; Wagner, U. Day-ahead probabilistic PV generation forecast for buildings
energy management systems. Sol. Energy 2018, 171, 478–490. [CrossRef]
4. Subudhi, B.; Pradhan, R. A Comparative Study on Maximum Power Point Tracking Techniques for
Photovoltaic Power Systems. IEEE Trans. Sustain. Energy 2013, 4, 89–98. [CrossRef]
5. El Khateb, A.; Rahim, N.A.; Selvaraj, J.; Uddin, M.N. Fuzzy-Logic-Controller-Based SEPIC Converter for
Maximum Power Point Tracking. IEEE Trans. Ind. Appl. 2014, 50, 2349–2358. [CrossRef]
6. Shadmand, M.B.; Balog, R.S.; Abu-Rub, H. Model Predictive Control of PV Sources in a Smart DC Distribution
System: Maximum Power Point Tracking and Droop Control. IEEE Trans. Energy Convers. 2014, 29, 913–921.
[CrossRef]
7. Houssamo, I.; Locment, F.; Sechilariu, M. Maximum power tracking for photovoltaic power system:
Development and experimental comparison of two algorithms. Renew. Energy 2010, 35, 2381–2387. [CrossRef]
8. Anzalchi, A.; Sarwat, A. Overview of technical specifications for grid-connected photovoltaic systems.
Energy Convers. Manag. 2017, 152, 312–327. [CrossRef]
9. Jo, J.H.; Aldeman, M.R.; Loomis, D.G. Optimum penetration of regional utility-scale renewable energy
systems. Renew. Energy 2018, 118, 328–334. [CrossRef]
10. Baranes, E.; Jacqmin, J.; Poudou, J.-C. Non-renewable and intermittent renewable energy sources: Friends
and foes? Energy Policy 2017, 111, 58–67. [CrossRef]
11. Honrubia-Escribano, A.; Ramirez, F.J.; Gómez-Lázaro, E.; Garcia-Villaverde, P.M.; Ruiz-Ortega, M.J.;
Parra-Requena, G. Influence of solar technology in the economic performance of PV power plants in
Europe. A comprehensive analysis. Renew. Sustain. Energy Rev. 2018, 82, 488–501. [CrossRef]
12. Rodriguez-Diaz, E.; Vasquez, J.C.; Guerrero, J.M. Intelligent DC Homes in Future Sustainable Energy Systems:
When efficiency and intelligence work together. IEEE Consum. Electron. Mag. 2016, 5, 74–80. [CrossRef]
13. Sechilariu, M.; Wang, B.; Locment, F. Building-integrated microgrid: Advanced local energy management
for forthcoming smart power grid communication. Energy Build. 2013, 59, 236–243. [CrossRef]
14. Sechilariu, M.; Wang, B.C.; Locment, F. Supervision control for optimal energy cost management in DC
microgrid: Design and simulation. Int. J. Electr. Power Energy Syst. 2014, 58, 140–149. [CrossRef]
15. Castillo-Cagigal, M.; Caamaño-Martín, E.; Matallanas, E.; Masa-Bote, D.; Gutiérrez, A.; Monasterio-Huelin, F.;
Jiménez-Leube, J. PV self-consumption optimization with storage and Active DSM for the residential sector.
Sol. Energy 2011, 85, 2338–2348. [CrossRef]
16. Yu, H.J.J. A prospective economic assessment of residential PV self-consumption with batteries and its
systemic effects: The French case in 2030. Energy Policy 2018, 113, 673–687. [CrossRef]
17. Lotfi, H.; Khodaei, A. AC Versus DC Microgrid Planning. IEEE Trans. Smart Grid 2017, 8, 296–304. [CrossRef]
18. Boscaino, V.; Guerrero, J.M.; Ciornei, I.; Meng, L.; Riva Sanseverino, E.; Zizzo, G. Online optimization of a
multi-conversion-level DC home microgrid for system efficiency enhancement. Sustain. Cities Soc. 2017, 35,
417–429. [CrossRef]
19. Wu, H.; Sechilariu, M.; Locment, F. Influence of Dynamic Efficiency in the DC Microgrid Power Balance.
Energies 2017, 10, 1563. [CrossRef]
20. Yan, R.; Saha, T.K.; Modi, N.; Masood, N.-A.; Mosadeghy, M. The combined effects of high penetration of
wind and PV on power system frequency response. Appl. Energy 2015, 145, 320–330. [CrossRef]
21. Sechilariu, M.; Wang, B.; Locment, F. Building Integrated Photovoltaic System With Energy Storage and
Smart Grid Communication. IEEE Trans. Ind. Electron. 2013, 60, 1607–1618. [CrossRef]
22. Guichi, A.; Talha, A.; Berkouk, E.M.; Mekhilef, S.; Gassab, S. A new method for intermediate power point
tracking for PV generator under partially shaded conditions in hybrid system. Sol. Energy 2018, 170, 974–987.
[CrossRef]
23. Franco, F.D.; Vu, T.V.; Gonsulin, D.; Vahedi, H.; Edrington, C.S. Enhanced performance of PV power control
using model predictive control. Sol. Energy 2017, 158, 679–686. [CrossRef]
Energies 2019, 12, 1233 12 of 12

24. Reddy, M.K.K.; Sarkar, V. LPPT control of a photovoltaic system against sudden drop of irradiance. In
Proceedings of the 6th International Conference on Computer Applications in Electrical Engineering-Recent
Advances (CERA), Roorkee, India, 5–7 October 2017; pp. 562–567.
25. Wang, B.; Houssamo, I.; Sechilariu, M.; Locment, F. A simple PV constrained production control strategy.
In Proceedings of the 21st IEEE International Symposium on Industrial Electronics, Hangzhou, China, 28–31
May 2012; pp. 969–974.
26. Li, Y.; Xu, Z.; Xiong, L.; Song, G.; Zhang, J.; Qi, D.; Yang, H. A cascading power sharing control for microgrid
embedded with wind and solar generation. Renew. Energy 2019, 132, 846–860. [CrossRef]
27. Choudar, A.; Boukhetala, D.; Barkat, S.; Brucker, J.-M. A local energy management of a hybrid PV-storage
based distributed generation for microgrids. Energy Convers. Manag. 2015, 90, 21–33. [CrossRef]
28. Ahmed, A.; Ran, L.; Moon, S.; Park, J.-H. A Fast PV Power Tracking Control Algorithm with Reduced Power
Mode. IEEE Trans. Energy Convers. 2013, 28, 565–575. [CrossRef]
29. Mahmood, H.; Michaelson, D.; Jiang, J. Decentralized Power Management of a PV/Battery Hybrid Unit in a
Droop-Controlled Islanded Microgrid. IEEE Trans. Power Electron. 2015, 30, 7215–7229. [CrossRef]
30. Sangwongwanich, A.; Yang, Y.; Blaabjerg, F.; Wang, H. Benchmarking of Constant Power Generation
Strategies for Single-Phase Grid-Connected Photovoltaic Systems. IEEE Trans. Ind. Appl. 2018, 54, 447–457.
[CrossRef]
31. Wu, H.; Sechilariu, M.; Locment, F. Impact of power converters’ efficiency on building-integrated
microgrid. In Proceedings of the 17th European Conference on Power Electronics and Applications (EPE’15
ECCE-Europe), Geneva, Switzerland, 8–10 September 2015; pp. 1–10.
32. Vanassche, P.; Triphase, N.V. Shaping Dynamic Response Using Virtual Circuit Control. Available
online: https://triphase.com/img/support/37/TriphaseHPC-VirtualCircuitControl.pdf (accessed on 17
March 2019).
33. Lana, A.; Mattsson, A.; Nuutinen, P.; Peltoniemi, P.; Kaipia, T.; Kosonen, A.; Aarniovuori, L.; Partanen, J.
On Low-Voltage DC Network Customer-End Inverter Energy Efficiency. IEEE Trans. Smart Grid 2014, 5,
2709–2717. [CrossRef]

© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).