This article appeared in a journal published by Elsevier.
The attached
copy is furnished to the author for internal noncommercial research
and education use, including for instruction at the authors institution
and sharing with colleagues.
Other uses, including reproduction and distribution, or selling or
licensing copies, or posting to personal, institutional or third party
websites are prohibited.
In most cases authors are permitted to post their version of the
article (e.g. in Word or Tex form) to their personal website or
institutional repository. Authors requiring further information
regarding Elseviers archiving and manuscript policies are
encouraged to visit:
http://www.elsevier.com/authorsrights
Author's personal copy
Energy Conversion and Management 77 (2014) 98107
Contents lists available at ScienceDirect
Energy Conversion and Management
journal homepage: www.elsevier.com/locate/enconman
Supervision and control of grid connected PVStorage systems
with the ve level diode clamped inverter
Kamal Himour a,, Kaci Ghedamsi a, El Madjid Berkouk b
a
b
Laboratory of Renewable Energy Mastery, University of Bejaia, Algeria
Control Process Laboratory, National Polytechnic School of Algiers, Algeria
a r t i c l e
i n f o
Article history:
Received 9 June 2013
Accepted 1 September 2013
Keywords:
Photovoltaic generator
MPPT
Battery storage
Five level diode clamped inverter
Space vector modulation
Supervision
a b s t r a c t
This paper aimed to evaluate the use of photovoltaicbattery storage systems to supply electric power in
the distribution grid through a multilevel inverter. The proposed system is composed by four PV generators with MPPT (P&O) control, four battery storage systems connected to each capacitor of the DC link
and a ve level diode clamped inverter connected to the grid by a traditional three phase transformer.
The proposed control has a hierarchical structure with both a grid side control level to regulate the power
and the current injected to the grid and four input side regulation units. The system operator controls the
power production of the four PV generators by sending out reference power signals to each input side
regulation unit, the input side regulation units regulate the voltage of each capacitor of the DC link, regulate the voltage and the state of charge of each battery storage system.
2013 Elsevier Ltd. All rights reserved.
1. Introduction
With the increasing concern about global environmental protection, the need to produce pollutionfree natural energy such
as solar energy has received great interest as an alternative source
of energy for the future since solar energy is clean, pollutionfree
and inexhaustible. In an effort to use the solar energy effectively,
a great deal of research has been done on the grid connected photovoltaic generation systems [18]. In PV systems connected to the
grid, the inverter that converts the output direct voltage of the solar modules to the alternate voltage (AC) is receiving increased
interest in order to generate power to utility.
In order to inject power on demand, certain energy storage devices must be added into the system. These devices must Store PV
energy in excess of electricity demand and subsequently meet
electricity demand in excess of PV energy. The conventional leadacid battery is the most common energy storage device at the present time [9].
Another very important aspect of the systems connected to the
grid is to select a proper power factor according to the grid demands: active or reactive power. The most efcient systems are
those, which allow variation in the active and reactive power injected into the grid, depending on the power grid requirements
[10]. In this scenario, we propose a control strategy for a photovoltaicbattery storage system connected to the grid with a ve level
diode clamped inverter. The proposed structure has many technical
and economic advantages over traditional systems. On one hand,
generally, a two level converter is used to connect production units
to the grid. This structure imposes physical and technical limits for
voltage rise and therefore power. Currently, for the connection of
multiple production units this structure is duplicated as many times
as possible. We need to use a connection point on the network for
each inverter of two levels used. The ve level diode clamped inverter does not have this problem because it has a DC bus segmented by
four series connected capacitors, so it offers many connection possibilities for continuous production units. In addition, a reduction of
harmonic distortion on alternating side (grid) is obtained, also a
reduction of the lter elements and the cost of installation. On the
other hand the proposed structure offers the advantage of balancing
the voltages across each capacitor and controlling the state of
charge of the batteries connected to each capacitor. This structure
also allows injecting to the grid a power smoothed taking into account batteries charge rate limits and stateofcharge constraints.
So, this paper is organized as follows: in Section 2 we presented
the global model of the system: mathematical model of the photovoltaic generator, model of battery bank, model and control of the
ve level diode clamped inverter, Energy management and control
structure. Then, in Section 3 we presented the simulation results
and we terminated by a conclusion of this study in Section 4.
2. Global system modeling
Corresponding author.
Email addresses: Himour.kamal@hotmail.fr (K. Himour), kghedamsi@yahoo.fr
(K. Ghedamsi), emberkouk@yahoo.fr (E.M. Berkouk).
01968904/$  see front matter 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.enconman.2013.09.001
Fig. 1 shows the conguration of the gridconnected PV battery
storage system, which consists of four PV generators, four DC/DC
Author's personal copy
99
K. Himour et al. / Energy Conversion and Management 77 (2014) 98107
converters for MPPT, four battery banks connected to the DC link
by four bidirectional DCDC converters and the ve level diode
clamped inverter connected to the grid through a traditional three
phase transformer.
The following sections deal with modeling of the different parts
of the conversion system and validation of the methodology by
simulations.
Iph
Rsh
exp
V I Rs
m KT
q
#
1
V I Rs
Rsh
I Np Iph Np Is exp
Np
V Rs I
Np
Rsh Ns
1
V Rs I
1
m K T=q
Ns
Np
2
2.2. Maximum power point tracking
The PV array must operate electrically at a certain voltage
which corresponds to the maximum power point under the given
operating conditions. To do this, a maximum power point tracking
(MPPT) technique should be applied. Various MPPT techniques like
lookup table methods, perturbation and observation (P&O) methods and computational methods have been proposed in the literature. The perturb and observe (P&O), as the name itself states that
the algorithm is based on the observation of the array output
power and on the perturbation (increment or decrement) of the
power based on increments of the array voltage or current. The
algorithm continuously increments or decrements the reference
current or voltage based on the value of the previous power sample
as shown in Fig. 3. The P&O is the simplest method which senses
the PV array voltage and the cost of implementation is less and
hence easy to implement [12].
where Iph is the photocurrent, Is is saturation current of diode, m is
ideality factor, Rs and Rsh are series and parallel resistance, T is
junction temperature, K is Boltzmann constant, and q is electron
charge.
In general, for a PVG involving an array of Ns cell connected in
series and Np in parallel, its output voltage current relation may be
deduced from the basic cell Eq. (2) as follows:
DC Bus
DC/DC
Converter
MPPT
Udc1
DC/DC
Converter
MPPT
Udc2
Five
level
DC/DC
Converter
MPPT
Udc3
DC/DC
Converter
MPPT
Udc4
R t, Lt
G
R
I
D
DCI
Battery Bank
Converter
Battery Bank
Converter
Utility side regulation unit
Power regulation
Current regulation
Grid synchronization
Battery Bank
Converter
Battery Bank
Converter
Input side
regulation unit 4
P ref4
Input side
regulation unit 3
Input side
regulation unit 2
P ref3
P ref2
Pg ref
Ppv glob
Charge
Fig. 2. Photovoltaic cell equivalent circuit.
The PV generator consists of electrically connected PV modules
and it is modeled by physical oriented equivalent circuits, including one or more diode. The single diode equivalent circuit as shown
in Fig. 2 is the most commonly used model for large PV generators.
Starting from the widely known photovoltaic cell electrical
equivalent circuit, an equivalent model for a more powerful PVG
made of an (Ns Np) array of PV cells, is established [11]:
The current source is the light generated current witch is directly proportional to the solar irradiation. The series and the shunt
resistances represent a voltage loss on the way to the external contacts and the leakage current in the shunt path respectively.
The mathematical model witch relates the output current to the
output voltage is given by the following expression:
"
I
V
Id
2.1. Model of PV generator
I Iph Is
Rs
System operators
Fig. 1. Global PV grid connection system.
Input side
regulation unit 1
P ref1
Author's personal copy
100
K. Himour et al. / Energy Conversion and Management 77 (2014) 98107
Start P&O algorithm
Mesure of V(k) et I(k)
P (k)=I(k)*V(k)
P=P(k)p(k1)
V=V(k)V(k1)
Yes
No
P>0
Yes
Yes
No
No
V>0
V>
V (k+1)= V(k) +V
V (k+1)=V(k)V
V (k+1)=V(k)V
V (k+1)=V(k) +V
Fig. 3. Flowchart of P&O algorithm.
2.3. Battery bank model
Lead acid batteries are used to guarantee several hours to a few
days of energy storage. The model representation of the leadacid
battery is shown in Fig. 4. The capacity of the battery is determined
by integrating the main reaction current IMR. To consider the increased gassing losses when charging the battery at high voltage
and temperature, here represented by the losscurrent Igas, constitutes a signicant improvement over alternative battery models
SOC model
Capacity
model
Voltage
model
Gassing current
Loss model
Fig. 4. General structure of battery model.
SOC
for the simulation of hybrid energy systems. The stateofcharge
can be calculated by referring the actual capacity to the rated
capacity of the battery, as expressed by Eq. (5) [13].
The main reaction current of the battery bank can be expressed
as:
IMR t Ibb t Igas t
where IMR is the main battery reaction current (A), Ibb is external
battery current (A), and Igas is battery gassing current (A).
The capacity model of the battery does not limit the charge or
discharge current. Operation of the system with excessive charge
or discharge currents has to be prevented by the selection of
appropriately sized components and the implementation of a suitable control strategy. The actual battery capacity can be determined as:
C b t
IMR tdt C b;i
t0
where Cb is the actual battery capacity (Ah), and Cb,i is Initial battery
capacity (Ah).
The stateofcharge can be calculated by referring the actual
capacity to the rated capacity of the battery:
SOCt
C b t
100%
C 10
The presented voltage model of the battery is based on the
Expanded Kinetic Battery Model, which has been presented in [13].
The voltage model considers that the battery terminal voltage
depends on the following factors:
Author's personal copy
101
K. Himour et al. / Energy Conversion and Management 77 (2014) 98107
Fig. 5. Fivelevel diode clamped inverter.
Table 1
The states of one leg of the ve level DCI.
(b) Discharging (Ibb > 0)
Eb t E0;d AD Xt
Etat
Sx1
Sx2
Sx3
Sx4
Sx5
Sx6
Sx7
Sx8
Vxo
P2
P1
O
N1
N2
1
0
0
0
0
1
1
0
0
0
1
1
1
0
0
1
1
1
1
0
0
1
1
1
1
0
0
1
1
1
0
0
0
1
1
0
0
0
0
1
Udc/2
Udc/4
0
Udc/4
Udc/2
Battery stateofcharge.
Internal battery resistance.
Magnitude and direction of battery current.
where Eb is the internal battery voltage (V), and X is Normalized
maximum charge/discharge capacity (Ah).
The normalized maximum charge/discharge capacity X is given
as:
Xt
(a) Charging (Ibb < 0)
C C Xt
Q max;C
C b t
Q max IMR t
(b) Discharging (Ibb > 0)
Xt
DC XtEFC
(a) Charging (Ibb < 0)
For all calculations shown, different model parameters represent
the characteristic voltage behavior of leadacid batteries when
charging or discharging. The internal battery voltage is calculated as:
Eb t E0;C AC Xt
C d Xt
Dd XtEFd
Q max;d Q max;d C b t
Q max IMR t
The maximum capacity Qmax in dependence of the main reaction
current of the battery is expressed by a third order polynomial
Author's personal copy
102
K. Himour et al. / Energy Conversion and Management 77 (2014) 98107
Fig. 6. Space vector diagram of velevel inverter.
Table 3
Second correction of reference voltage vector.
Hexagon
Component V 0
d
Component V 0
q
V 0
d 1=4
V 0
q
V 0
d 1=8
V 0
q
V 0
d
V 0
d
V 0
d
V 0
d
V 0
q
V 0
q
V 0
q
V 0
q
4
5
6
1=8
1=4
1=8
1=8
p
3=8
p
3=8
p
3=8
p
3=8
(b) Discharging (Ibb > 0)
Q max IMR t D1 IMR t3 D2 IMR t2 D3 IMR t D4
11
Therefore, the battery terminal voltage Vb can be calculated as:
(a) Charging (Ibb < 0)
V b t Eb t R0;C IMR t
Fig. 7. Decomposition of ve level space vector diagram.
(b) Discharging (Ibb > 0)
V b t Eb t R0;d IMR t
V 0
d
V d
V d
V d
V d
V d
V d
2
3
4
5
6
V 0
q
1=2
V q
1=4
V q
V bb t Bs V b t
1=4
p
3=4
p
V q 3=4
1=2
V q
1=4
V q
1=4
p
3=4
p
V q 3=4
equation, where the parameters have to be determined by empirical
curve tting from measured data [13]:
(a) Charging (Ibb < 0)
Q max IMR t C 1 IMR t3 C 2 IMR t2 C 3 IMR t C 4
13
The voltage of a string of batteries is given by multiplying the battery voltage with the number of 12 V batteries in series:
Table 2
First correction of reference voltage vector.
Hexagon
12
10
14
where Vbb is the voltage of battery bank (V), and Bs is number of
12 V batteries in series.
2.4. Model and control of the ve level DCI
Multilevel converter gives massive advantages compared with
conventional and very wellknown two level converters like; high
power quality waveforms, low switching losses, high voltage capability, and low electromagnetic compatibility (EMC). At the present
time, the majority of research and development effort seems to
concentrate on the development of three classes of inverters: the
Author's personal copy
K. Himour et al. / Energy Conversion and Management 77 (2014) 98107
diodeclamped multilevel inverter, the multilevel inverter with
cascaded singlephase Hbridge inverters and the multilevel inverter known as the ying capacitor inverter or sometimes as the
imbricate cells multilevel inverter [1416].
2.4.1. Model of the ve level diode clamped inverter
Fig. 5 shows diagram of a velevel diode clamping inverter.
Each leg is composed of four upper and lower switches with
antiparallel diodes. Four series dclink capacitors split the dcbus voltage in half, and eighteen clamping diodes conne the voltages across the switches within the voltages of the capacitors. The
necessary conditions for the switching states for the velevel inverter are that the dclink capacitors should not be shorted, and
the output current should be continuous. As indicated in Table 1,
each leg of the inverter can have ve possible switching states,
P1, P2, O, N1 or N2. When the top four witches Sx1, Sx2, Sx3and Sx4
(x = a, b, c) are turned on, switching state is P2. When the switches
Sx2, Sx3, Sx4and Sx5 are turned on switching state is P1. When the
switches Sx3, Sx4, Sx5and Sx6 are turned on, the switching state is
O. when the switches Sx4, Sx5, Sx6and Sx7 are turned on, the switching state is N1. When the switches Sx5, Sx6, Sx7and Sx8 are turned on,
the switching state is N2 [17].
103
According to the states of the inverter, the output voltage vector
can take several positions in the dq frame. These positions are
indicated on the space vector diagram in Fig. 6.
2.4.2. Simplied space vector modulation
Recurrently three major PWM techniques are applied in multilevel inverters: (1) sinusoidal PWM (SPWM), (2) third harmonic
injection PWM (THPWM), and (3) space vector PWM (SVM) [14].
In SVPWM method the output voltage is approximated by using
the nearest three output vectors that the nodes of the triangle containing the reference vector in the space vector diagram of the inverter. When the reference vector changes from one region to
another, it may induce an output vector abrupt change. In addition
we need to calculate the Switching sequences and switching time
of the states at every change of the reference voltage location.
Thus the computational complexity is greatly increasing with
the increasing number of the reference vectors and it is a main limitation of the application of this typical SVPWM [17].
In this paper, a new method is proposed in which the space vector diagram of velevel inverter is decomposed into six space vector diagrams of three level inverters. In turn, each of these space
vector diagrams of three level inverters is decomposed into six
space vectors diagrams of two level inverters like showed in
Table 4
System parameter values.
Photovoltaic array [20]
Pmax
150
Vop
34.5
Iop
4.35
Voc
43.5
Icc
4.75
Ns
5
Np
0
W
V
A
V
A
Maximal power
Optimal voltage
Optimal current
Open circuit voltage
Short circuit current
Number of series arrays
Number of parallel arrays
DC bus
Udc
480
DC bus voltage
Filter
Rt
Lt
3
0.05
O
H
Filter resistance
Filter inductance
Grid
Vs
f
380
50
V
Hz
voltage
frequency
Ah
V1
K1
A
Ah
Ah
mO
mO
V
h/A2
h/A
h
Ah
h/A2
h/A
h
Ah
Battery capacity at 10h discharge rate
Voltage coefcient
Temperature coefcient
Normalised gassing current
Maximum charge capacity
Maximum discharge capacity
Internal resistance when charging the battery
Internal resistance when discharging the battery
Limiting internal battery voltage for zero current and fully discharged battery after the initial transient
Parameter reecting the initial linear variation of the internal battery voltage with increasing stateofcharge
Parameter reecting increasing voltage when battery is progressively charged
Parameter reecting sharp increase of voltage when battery is charged to a high SOC
Exponential factor introduced to achieve a closer curve t for voltage behavior when progressively charged
Fully charged internal battery voltage when discharging after the initial transient
Parameter reecting the initial linear variation of the internal battery voltage with decreasing stateofcharge
Parameter reecting decreasing voltage when battery is progressively discharged
Parameter reecting sharp decrease of voltage when battery is discharged to a low SOC
Exponential factor introduced to achieve a closer curve t for voltage behavior when progressively discharged
Parameter to calculate maximum charge capacity
Parameter to calculate maximum charge capacity
Parameter to calculate maximum charge capacity
Parameter to calculate maximum charge capacity
Parameter to calculate maximum discharge capacity
Parameter to calculate maximum discharge capacity
Parameter to calculate maximum discharge capacity
Parameter to calculate maximum discharge capacity
Number of series batteries of 12 V
Battery bank [13]
C10
118
CV
8
CT
0.05
IGO
0.035
Qmax,C
120
Qmax,D
130
Rl,C
75
Rl,D
38
E0,C
11.6
AC
0.01
CC
0.012
DC
130
EFC
0.45
E0,d
12.6
AD
0.007
CD
0.3
DD
165
EFD
1.25
C1
0.0006
C2
0.0543
C3
2.0279
C4
140.29
D1
0.0006
D2
0.0543
D3
2.0279
D4
140.29
Bs
36
Author's personal copy
104
K. Himour et al. / Energy Conversion and Management 77 (2014) 98107
Five
Rt, Lt
G
R
I
D
level
Udc
DCI
Is1,2,3
DC Bus
Vs1,2,3
ABC
Inverter control
SVM
dq
Q ref
P ref
dq
V td ref
Power Control
I td ref
V sd, V sq
Isd, I sq
ABC
V tq ref
I tq ref
Curent Control
Lt. s
PI
Lt. s
PI
Fig. 8. Bloc diagram grid side control.
Vt d_ref
vb d_ref
Cid

Bidirectional DCDC
I bat dc
converter
Battery bank
Vsd
DC
It d_ref
V bat
DC
s.Lt
Isd
I bat
s.Lt
Vt q_ref
vb q_ref
Ciq
+
Compensation
PI
P bat ref
It q_ref
Vsq
Decoupling
D bat
Isq
Fig. 10. Input side control unit structure.
Correcteur
Fig. 7. This modication can reduce considerably the computational time and reduce the algorithm complexity [17].
2.4.2.2. Second correction of reference voltage vector. Having the selected three level inverter and the location of the translated vector,
one hexagon is selected among the six small hexagons that contain
this three level diagram. We make a translation of the reference
~
V 00
s . Table gives the components d and q of the reference voltage
~
V 00
s .
2.4.2.1. First correction of reference voltage vector. Having the location of a given reference voltage vector, one hexagon is selected
among the six small hexagons that contain the ve levels space
vector diagram.
Each hexagon is identied by a number s dened as given by:
2.4.2.3. Determination
of dwelling times. Ones the corrected refer!
ence voltage V 00
s and the corresponding hexagon are determined;
we can apply the conventional two level space vector Modulation
method
to calculate
the #
dwelling times,
"
8
Fig. 9. Bloc diagram of the current control.
8
1
>
>
>
>
>
2
>
>
>
<3
s
>
4
>
>
>
>
>
>5
>
:
6
if
p
3
if
if
if
if
if
6 h 6 p3
6h6
2
p 6 h 6 5p
2
6
5p
6 h 6 76p
6
7p
6 h 6 32p
6
3p
6 h 6 116 p
2
15
After selection of one hexagon, we make a translation of the reference vector V s towards the center of this hexagon (see Tables 24).
>
>
>
T1 4
>
>
>
<
p
jV 00
s j T s sin 3 a
sin p
3
!
jV 00
s j T s
sina
>
>
T2 4
>
sin p3
>
>
>
:
T3 Ts T1 T2
16
2.4.2.4. Conversion and! sequence of the switching states. The reference voltage vector V 00
s is approximated using the nearest three
states which are nodes of the triangle containing the vector identied as X, Y and Z. the optimum sequence of these states is selected
so as to minimize the total number of switching transitions [17].
Author's personal copy
105
K. Himour et al. / Energy Conversion and Management 77 (2014) 98107
Energy and power informations
Yes
Yes
No
SOC= SOC max
Excess power
inject to grid
No
Yes
SB charge by
excess power
SOC= SOC min
SB charge by
PV power
No
inject to grid PV
power and deficient
power from SB
Fig. 11. Operation of storage battery (SB).
1.4
Irradiance 1
400
Irradiance (sun)
1.2
Battery banks power (W)
Irradiance 2
Irradiance 3
Irradiance 4
1
0.8
0.6
0.4
0
0.2
0.4
0.6
0.8
1.2
1.4
1.6
1.8
Power of battery bank 1
Power of battery bank 2
200
Power of battery bank 3
Power of battery bank 4
200
400
0
0.2
0.4
0.6
0.8
1.2
1.4
1.6
1.8
1.6
1.8
Time (s)
Time (s)
Fig. 12. Prole irradiance of each PV generator.
Fig. 14. Battery banks power (W).
3500
State of charge (%)
Photovoltaic power (W)
29.5
3000
2500
Global PV power
Power of GPV 1
Power of GPV 2
Power of GPV 3
Power of GPV 4
2000
1500
1000
500
29
28.5
SOC of battery bank 1
SOC of battery bank 2
28
SOC of battery bank 3
SOC of battery bank 4
27.5
0
0.2
0.4
0.6
0.8
1.2
1.4
1.6
1.8
0.2
0.4
0.6
0.8
1.2
1.4
Time (s)
Time (s)
Fig. 15. State of charge of battery banks.
Fig. 13. PV generators power (W).
2.5. Energy management and control structure
2.5.1. Grid side regulation
Fig. 8 shows the structure of the grid side regulation [1819].
The system operator controls the power production of the four
PV generators by sending out reference power signals to each input
side regulation unit. In this paper a simply dispatching function
distributes power reference to inside regulation units based on a
Proportional distribution of the available active power.
2.5.1.1. Power control. The active and reactive power (Pg, Qg) can be
both expressed by using Park components of supply voltage (Vtd, Vtq) and line current (Itd, Itq) as follows:
prefi
ppv
ppv
pg
ref
17
Pg V td Itd V tq Itq
Q g V td Itd V tq Itq
18
global
The control structure is composed of two main blocs:
Reference currents Itd ref ; Itq ref which allows setting the desired
reference active and reactive powers Pgref ; Q gref , as follows:
Author's personal copy
K. Himour et al. / Energy Conversion and Management 77 (2014) 98107
122
Mag (% of Fundamental)
Battery banks voltages (V)
106
120
Voltage of battery bank 1
Voltage of battery bank 2
118
Voltage of battery bank 3
Voltage of battery bank 4
116
0.2
0.4
0.6
0.8
1.2
1.4
1.6
1.8
Fundamental (50Hz) = 1093 , THD= 15.61%
4
3
2
1
0
10
15
20
25
30
35
40
Harmonic order
Time (s)
Fig. 19. THD analysis of output voltage of 5 level inverter.
120.05
Grid current (A)
DC bus voltages (V)
Fig. 16. Voltage of battery banks (V).
120
Udc 1
119.95
0.2
0.4
Udc 2
0.6
0.8
Udc 3
1.2
1.4
Udc 4
1.6
1.8
5
0
5
Time (s)
0.2
0.4
0.6
0.8
1.2
1.4
1.6
Time (s)
Fig. 17. DC link capacitors voltages (V).
Output voltage (V)
Fig. 20. Phase 1 grid currant (A).
2000
grid voltage (Fig. 9). The electric equations of the lter (Rt, Lt) connected to the grid are given below:
1000
V tq Rt Itq LT
0.01
0.02
0.03
0.04
0.05
0.06
Time (s)
Fig. 18. Phase 1 output voltage of 5 level inverter (V).
>
>
>
: Itq
dItq
dt
xs LT Itq V gq
20
1000
2000
8
>
>
>
< Itd
V td Rt Itd Lt dIdttd xs Lt Itd V gd
ref
P gref b
V td Q gref b
V tq
bV 2 bV 2tq
td
ref
19
P gref b
V tq Q gref b
V td
bV 2 bV 2tq
td
The unity power factor is obtained simply by setting the reactive
power reference null. We can also generate or absorb (Qgref < 0 or
Qgref > 0).
2.5.1.2. Current control. The vector current control in Park reference
frame is carried out by using the synchronized reference with the
2.5.1.3. Multilevel control. The ve level diode clamped inverter is
controlled by the simplied space vector modulation like presented in Section 2.4.2.
2.5.2. Input side regulation units
The input side regulation units have two main objectives: control of battery bank storage system as shown in Fig. 10 and regulation of the DC link capacitor with a PI corrector who gives the
reference current to inject into the DC link capacitor. The battery
is integrated to the DCLink via buckboost DC/DC. The diagram
block of the battery converter control is illustrated in Fig. 10.
The battery power reference is generated by the supervisory
controller. The PI regulator adjust the current ibat in order to follow the reference ibatref, with an objective of charging or discharging the battery according to the need.
Fig. 11 shows ow chart of operation of storage battery system
(see Figs. 1216).
Mag (% of Fundamental)
Fundamental (50Hz) = 4.309 , THD= 1.98%
1.4
1.2
1
0.8
0.6
0.4
0.2
0
500
1000
1500
Frequency (Hz)
Fig. 21. THD analysis of phase 1 grid current.
2000
Author's personal copy
Active power (W)
K. Himour et al. / Energy Conversion and Management 77 (2014) 98107
107
4000
4. Conclusion
3000
This paper proposed the study and the control of photovoltaicBattery storage grid connected system, the use of a ve level DCI
with his simplied space vector modulation as a grid interface
gives a good results in term of THD and power quality, also, the
aim was in this work to inject to the grid a xed power whatever
solar irradiance and temperature condition. The results obtained
from this performance analysis conrm that the control strategy
adopted achieves the specied performance objectives.
2000
Grid active power
Reference power
1000
0
0.2
0.4
0.6
0.8
1.2
1.4
1.6
1.8
Time (s)
Acknowledgments
Reactive power (VAR)
Fig. 22. Grid active power (W).
I would like to express my sincere gratitude to may tutors, Prof
Dr. Ghedamsi Kaci and Prof. Berkouk Elmadjid without whom none
of this would ever have happened. Their professional supervision,
insightful mentoring, encouragement and vast help steered the
undertaken research into right direction.
2
0
References
2
Grid reactive power
4
Reference reactive power
6
0.2
0.4
0.6
0.8
1.2
1.4
1.6
1.8
Time (s)
Fig. 23. Grid reactive power (var).
Grid voltage and current
600
Grid voltage
Grid current*40
400
200
0
200
400
0.8
0.85
0.9
0.95
1.05
1.1
1.15
1.20
Time (s)
Fig. 24. Zoom of grid currant (A) and grid voltage (V).
3. Results simulation
In this section, the photovoltaic grid connection system is simulated using SIMULINKMATLAB. PV generators have different irradiance proles and temperature. Each PV generator is composed of
ve series connected arrays.
Each PV generator has a prole of irradiance and temperature different from other as shown in Fig. 12 and the photovoltaic power
output is the sum of the four PV generators powers (Fig. 13). Results
simulation show the contribution of storage ingrid connected photovoltaic systems. Withbattery storage units, we can balance the
DC bus as shown in Fig. 17 and the strategy adopted allows controlling the level of storage of each storage unit when the photovoltaic
productions are different and compensate the uctuations of the
photovoltaic poweras illustrated in Figs. 1416. Also, the results
show that the connection of multiple generators using a single multilevel inverter can inject more power to grid as shown in Figs. 22
and 23. The output current of the multilevel inverter has a good quality (Fig. 21). The results demonstrate the effectiveness of the inverter
control by the simplied space vector modulation as shown in
Fig. 18 and Fig. 19. Fig. 20 shows that when the reference power
changes from 2 to 3 KW the current injected by the inverter changes.
Fig. 24 show a zoom of grid current and voltage.
[1] Hassaine L, Olas E. Simulation of gridinterface connecting photovoltaic power
systems. In: Proceedings of the world renewable energy congress IX orence,
Italy; 2006.
[2] Hamrouni N, Jraidi M, Chrif A. New control strategy for 2 stage grid connected
photovoltaic power system. Renew Energy 2008;33:221221.
[3] ReyBou Alexis B, GarcaValverde Rafael, et al. An integrative approach to the
design methodology for 3phase power conditioners in Photovoltaic GridConnected systems. Energy Convers Manage 2012;56:8095.
[4] Wu TF, Nien HS, Shen CL, Chen TM. A singlephase inverter system for PV
power injection and active power ltering with nonlinear inductor
consideration. Trans Ind Appl 2005;41(4):107583.
[5] Dali M, Belhadj J, Roboam X. Hybrid solarwind system with battery storage
operating in grid connected and standalone mode: control and energy
managementexperimental investigation. Energy 2010;35(6):258795;
Tsang KM, Chan WL. Threelevel gridconnected photovoltaic inverter with
maximum power point tracking. Energy Convers Manage 2013;65:2217.
[6] Suroso, Noguchi Toshihiko. A new threelevel currentsource PWM inverter
and its application for grid connected power conditioner. Energy Convers
Manage 2010;51:14919.
[7] Altin Necmi, Ozdemir Saban. Threephase threelevel grid interactive inverter
with fuzzy logic based maximum power point tracking controller. Energy
Convers Manage 2013;69:1726.
[8] Wu JinnChang, Wu KuenDer, Jou HurngLiahng, Wu ZongHan, Chang ShengKai. Novel power electronic interface for gridconnected fuel cell power
generation system. Energy Convers Manage 2013;71:22734.
[9] Chakrabarti Mohammed Harun, Roberts Edward Pelham Lindeld, Bae
Chulheung, Saleem Muhammad. Ruthenium based redox ow battery for
solar energy storage. Energy Convers Manage 2011;52:25018.
[10] Pouresmaeil E, MontesinosMiracle D, GomisBellmunt O, BergasJan J. A
multiobjective control strategy for grid connection of DG (distributed
generation) resources. Energy 2010;35(12):502230.
[11] Di Fazio Anna Rita, Russo Mario. Photovoltaic generator modelling to improve
numerical robustness of EMT simulation. Electric Power Syst Res
2012;83:13643.
[12] Esram T, Chapman PL. Comparison of photovoltaic array maximum power
point tracking techniques. IEEE Trans Energy Conver 2007;22:43949.
[13] Wichert Benjamin. Control of photovoltaic diesel hybrid energy systems. Ph.D.
thesis of the Curtin University of Technology; 2000.
[14] Lega Alberto. Multilevel converters: dual twolevel inverter scheme. Ph.D.
thesis. University of Bologna; March 2007.
[15] Lai JS, Peng FZ. Multilevel converters : a new breed of power converters. IEEE
Trans Ind Appl 1996;32(3):50917.
[16] Colak I, Kabalci E, Bayindir R. Review of multilevel voltage source inverter
topologies and control schemes. Energy Convers Manage 2011;52:111428.
[17] Lalili D, Berkouk EM, Boudjema F, Lourci N, Taleb T, Petzold J. Simplied space
vector PWM algorithm for threelevel inverter with neutral point potential
control. Medeteranean J Meas Control 2007;3(1):309.
[18] Blaabjerg F, Teodorescu R, Liserre M, Timbus A. Overview of control and grid
synchronization for distributed power generation systems. IEEE Trans Ind
Electron 2006;53(5):1398409.
[19] Bouchafaa F, Beriber D, Boucherit MS. Modeling and control of a gird
connected PV generation system. In: 18th Mediterranean conference on
control & automation; 2010. p. 31520.
[20] BP Solar BP SX150150W Multicrystalline Photovoltaic Module Datasheet;
2001.
Viel mehr als nur Dokumente.
Entdecken, was Scribd alles zu bieten hat, inklusive Bücher und Hörbücher von großen Verlagen.
Jederzeit kündbar.