Journal of Electrical Engineering

www.jee.ro

Control of Photovoltaic Water Pumping System

Chafia Serir Djamila Rekioua
serirch@yahoo.fr dja_rekioua@yahoo.fr
Department of Electrical Engineering, University of Bejaia Department of Electrical Engineering, University of Bejaia
Laboratory LTII, Laboratory LTII,
Bejaia, Algeria Bejaia, Algeria

AbstractDue to the continuous decrease of the solar cells cost, In this paper, we present a PV pumping water which
photovoltaic energy is used in different applications. The most includes photovoltaic array generator, DC/DC converter,
important one is the water pumping system powered by DC/AC converter and induction motor coupled to a centrifugal
photovoltaic generators. These systems can work with or without pump. The FLC controller is applied to ensure a a maximum
storage battery. With the increased use of this application, more operating of the photovoltaic array. And to improve the FLC
attention has been paid to their optimum utilization. Many
controller, we make a comparison with the classical MPPT,
methods have been developed to determine the maximum power
point (MPP). In this paper, to control the DC bus voltage, we the Perturb and Observ (P&O). Obtained results are presented.
apply field oriented control (FOC) strategy to induction motor
(IM) supply by a photovoltaic (PV) system. And to maximize the II. PROPOSED SYSTEM DESCRIPTION
efficiency of the proposed PV pumping system, we use the fuzzy
logic controller (FLC) and the classical Perturb and Observ
The proposed studied system is shown in Fig.1.It consists of a
(P&O). Different tests have been carried to prove the photovoltaic-pumped system, composed of a PV generator,
effectiveness of the proposed control system. DC-AC converter, a field oriented controlled induction motor
and centrifugal pump.
Keywords- Pumping Photovoltaic system, Fiel oriented
contorl, MPPT, Fuzzy logic controlle, Perturb and Observ .
A. Photovoltaic generator model
This model is characterized by a very simple resolution. It
I. INTRODUCTION requires only four parameters namely Isc, Voc, Vmp and Imp. The
Ipv-Vpv characteristic of this model is illustrated as follows:
The photovoltaic array has a unique operating point (MPP)
V pv
that can supply maximum power to the load. The locus of this I pv I sc 1 C 1 exp 1 (1)
point has a non-linear variation with solar irradiation and C 2 .V oc

temperature. Therefore, to maximize the efficiency of the With:
photovoltaic energy system, it is necessary to track the Vm
maximum power point of the PV array. Many methods and 1
V
controllers have been widely developed and implemented to C2 oc
Im
track the maximum power point (MPP) [2-6]. Most control ln( 1 )
I sc
schemes use the Perturb and Observ (P&O) method which is
Im Vm
based on iterative algorithms to track continuously the MPP, C 1 (1 ) exp( )
I sc C 2 .V oc
because it is easy to implement [2,3] but the oscillation
problem is unavoidable. In many references the effectiveness We make validation through the following experimental bench
of a fuzzy logic controller is shown [2-5] compared to the (Fig.3.)
(P&O) method. It improves control robustness and this control
gives robust performance under parameters and load variation.
Several authors present much attention to the study of the
dynamic performance of the photovoltaic pumping systems.
A. Terki and al [7] presented an analysis of the dynamic
performance of a permanent magnet brushless DC motor
controlled through a hysteresis current loop. Betka [8]
presented the performance optimization of an asynchronous
motor associated at a PV generator. Recently, vectorial
command of induction motor pumping system supplied by
photovoltaic generator was studied by Makhlouf and al [9]. In
H. Hadi [10], the photovoltaic pumping system with battery is
proposed to reduce the overheating of the motor temperature
and increase the efficiency. The battery is installed as the
storage of the surplus energy and backup energy. Fig.3. Experimental PV bench,

1
 Journal of Electrical Engineering
www.jee.ro

PV

Vdc I pv
PI Vdref Varef Inverter
P pv +
Ppv Vbref
-
G1

2--3 Vcref
G3
Vpv,ref +
V pv,ref PI
V pv +
Vpv
G2
Fuzzy logic controller
+ Vqref
-

Vdcref
ia
Ib
Pref 3--2 Ic

ref
Teref Lr Iqsref
IM
PI
p.M. ref
Iqs
Pump
ref
1 Idsref Ids
M

M I qsref r s s
Tr ref
p
r

Fig.1. Proposed FOC-FLC of PV Pumping System

Figure 4 shows the current/voltage characteristics obtained by

simulation and compared with the experimental values
corresponding to a 110 Wc Siemens panel (Table 1.). From
these characteristics, the non-linear nature of the PV array is
apparent. Therefore, an MPPT algorithm must be incorporated
to force the system to always operate at the maximum power
point (MPP).
Table.1.
Parameter of the PV panel SIEMENS SM110-24
PPV 110W
Impp 3.15A
Vmpp 35V
Isc 3.45A
Voc 43.5V
sc 1.4mA/C
oc -152mV/C
Pmpp 110W
B. MPPT CONTROLLERS
B.1.Perturb& Observ method
This is the most widely used method [2-12]. A feedback loop
and few measures are needed. The panel voltage is
deliberately perturbed then the power is compared to the Optimal zone
Fig.4.Comparison of experimental results with simulation ones
power obtained before to disturbance.

2
 Journal of Electrical Engineering
www.jee.ro

Specifically, if the power panel is increased due to the Where P is the electrical power input of the motor-pump, h is
disturbance, the following disturbance will be made in the the total head and a(h), b(h), c(h), d(h) are the coefficients
same direction. And if the power decreases, the new corresponding to the working total head.
perturbation is made in the opposite direction. The advantages With: ai, bi, and di constants which depend on the type of sub-
of this method can be summarized as follows: knowledge of solar pumping system.
the characteristics of the photovoltaic generator is not The calculation of the instantaneous flow in terms of power is
required, it is relatively simple. Nevertheless, in steady state, calculated using Newton-Raphson method. Thus at the kth
the operating point oscillates around the MPP, which causes iteration, the flow Q is given by the following equation:
energy losses. The MPPT is necessary to draw the maximum For d Pa (Q)> 0:
amount of power from the PV module [8-11]. F(Q k 1 ) (4)
Q Q
k
k 1
F ' (Q k 1 )
B.2. Fuzzy logic controller
Fuzzy logic controller is introducing to determine the With:
3 2
operating point corresponding to maximum power for F(Qk 1 ) a Q k 1 b Q k 1 c Qk 1
d Pa (Qk 1 ) (12)
different insolation levels and temperature. In this case, inputs F(Qk-1) is the derivative of the function F(Qk-1)
of the fuzzy logic controller are power variation (Ppv) and We use an induction motor which is modeled using voltage
voltage variation (Vpv). The output is reference voltage and flux equations referred in a general frame:
variation (Vpv,ref). In order to converge towards the optimal d Sd
point, rules are relatively simple to establish. These rules V Sd R S I Sd
dt (5
depend on the variations of power Ppv and voltage Vpv. In d Sq
accordance with Table.3, if the power (Ppv) increased, the V Sq R I

S Sq
dt
operating point should be increased as well. However, if the Where: (Isd Isq), (Vsd, Vsq) and sd, sq) are the (d,q)
power (Ppv) decreased, the voltage (Vpv,ref ) should do the components of the stator current, voltage and flux, Rs is the
same. stator resistance.
Table.2 d Rd d
Fuzzy rule table [27]. 0 V Rd R R I Rd dt dt Rq (14)
d Rq d
Ppv BN MN SN Z SP MP BP 0 VRq R R I Rq Rd
Vpv dt dt
BN BP BP MP Z MN BN BN Where: IRd, IRq are (d,q) rotor current, Rd, Rq are (d,q) rotor
MN BP MP SP Z SN MN BN flux, Rr is the rotor resistance.
SN MP SP SP Z SN SN MN We obtain the follow mathematical model:
di ds
R P r L 2m L m .R r P r Lm
Z BN MN SN Z SP MP BP s
dt Ls Ls Lr Ls L r Ls
SP MN SN SN Z SP SP MP di qs
P L2 R P r Lm
i
L m .R r ds
r m
s i (6)
MP BN MN SN Z SP MP BP dt 1 Ls L r Ls Ls L s L r qs
L m .R s
di dr P r Lm Rr i
BP BN BN MN Z MP BP BP P. r dr
dt Ls L r Lr Lr i qr
di qr
P r Lm L m .R s R
P. r r
From these linguistic rules, the MPPT algorithm contain dt Lr Ls Lr Lr
measurement of variation of photovoltaic power Ppv and

1
0


variation of photovoltaic voltage Vpv proposes a variation of Ls
1
0
the voltage reference Vpv,ref according to eq.2. 1

Ls v ds

L m 0 v qs
Ppv Ppv ( k ) Ppv ( k 1) L s .L r
(2) L
Vpv Vpv (k ) Vpv ( k 1) 0 m
V L s .L r
pv ref ( k ) Vpv (k 1) Vpv ref ( k ) With: is the leakage coefficient
Where:Ppv(k) and Vpv(k) are the power and voltage of the -The mechanical equation is given as:
photovoltaic generator at sampled times (k), and Vpv,ref (k) the
d r (7)
instant of reference voltage Tem TLoad J .
dt
C. Modeling subsystem pumping With: r is the AC motor velocity angular, J the inertia of the
Many different varieties of pumps are used with PV-pumping AC motor.
system. In our case, we use the model expresses the water The electromagnetic torque can be written as:
flow output (Q) directly as a function of the electrical power Tem P.(sd .is q sq .isd ) (8)
input (P) to the motor-pump, for different total heads. A
polynomial fit of the third order expresses the relationship III. FIELD ORIENTED CONTROL
between the flow rate and power input, as described by the
In our work, we choose the orientation of rotor flux such as:
following equation [6, 7]:
P(Q, h) a(h)Q 3 b(h)Q 2 c(h)Q d (h) (3) rd = r and rq = 0. This means that the flux r is

3
 Journal of Electrical Engineering
www.jee.ro

aligned permanently along the d-axis. Finally, as the chosen Ppvmax

frame implies rq = 0, the expression of the electromagnetic % 100
Pmax
torque becomes: (19)
M (9)
Te p . . r . i sq With: Vpvmax and Ppvmax are voltage and power obtained with
Lr
The rotor flux as a function of the current isd and the rotor time MPPT.
constant Tr=Lr/Rr is given by the following expression: Table.3.
M . isd Error calculation for the two method s under different
r (10) conditions
1 Tr . s Tests conditions MPPT % % p v
Where: s represents the derivative operator.
High irradiance P&O 0.630 1.220
The knowledge of s , by using the internal angular relation E=850 W/m2 FLC 0.190 0.090
s r p . and the mechanical speed of the machine T=35C
Medium irradiance P&O 1.740 1.100
is measured continuously; the speed of the rotor field is
E=500 W/m2 FLC 1.060 0.060
estimated by the following expression:
T=28C
M .isq Low irradiance P&O 0.570 0.820
r (11)
E=350 W/m2 0.280 0.330
Tr . r FLC
T=17C
Then, s can be written in the following way: Table.4.
M .i sq Efficiency of two methods under STC test conditions
s p. (12)
Tr . r STC Test condition MPPT %
The output power Ppv from PV will be fed to induction motor
PO 99.409
pump relating with torque and speed can be shown in the
E=1000 W/m2 FLC 99.822
following equation
T=25C
Tem K . 2 (13)
The mechanical output power of induction motor pump is:
P mec K . 3 V. APPLICATION
(14)
where K is the Pump constant. We make a sizing of the various components of the studied
From energy theory (Pi,, = Po,, = P, ) the frequency angle is: system which consists of a water tank of 70 m3 to satisfy the
3 domestic needs of a family. The dynamic level head is about
Ppv
(15) 12m and the nominal flow rate is of 21m3/h. Table5.
K summarizes the obtained results of the sizing system.
The mechanical torque of induction motor pump can be Table5.
written as: Sizing pumping system
3 Symbols Expressions Results
T em K .P 2 pv (16)
Phyd Phyd .g.HMT.Qn 680 W

IV. NUMERICAL SIMULATION Pmec Pmec

Phyd
1240W
ipump
In order to prove the robustness of the proposed MPPT using Pelec Pelec
P
mem 1470W
FLC, we compare it with conventional MPPT using P&O imec
Pelec
algorithm in terms of tracking of the PPM at different tests Pd Pd
inv
1550W

conditions (High, medium and low irradiances). d d

V
3.33 hours
We make error calculation of the power and voltage at steady Qn

state for the two MPPT methods (Table 3.). We use the Ec E c d .Pd 5.16kWh / day

following expressions: Pg Pg
Ec
1.94 kW
p 1
pertes
Ppvmax . madele .lectrique Ppvmax N pv Ppv max N pv 18pannels
p % 100 Ns
N pv
Pmax t
Ppvmax .madele .lectrique I Ns 6pannels
N paralel m
(17) N paralel I opt N paralel 3

Vpvmax .madele .lectrique Vpvmax

v % 100 We obtain the following scheme (fig.5) with the different
V pvmax . madele. lectrique results:
(18)
The efficiency can be calculated by:
U GPV N p .series * I pv 6*35 210V

4
 Journal of Electrical Engineering
www.jee.ro

I GPV N braches * I pv 3*3.14 9.42 A Table 7.

Comparison results with and without MPPT
PPV UGPV * IGPV 210*9.42 1.978kW . Strategies Response Pump flow Error pump Pump
time(s) QMPPT (m3/s) flow Efficiency
Ipv
Q % %
-3
FLC 0.0178 5.613 10 3.22 53.29
P&O 0.0552 5.298 10-3 8.65 50.30
Pg Direct 0.273 4.600 10-3 20.69 43.67
DC/A
18 pannels

Pelec Pmec coupling

Vpv IG A vector control based on FLC of the induction motor,
with optimization is used. Simulation results are carried out
Pump
DC/DC under variation of environmental conditions to verify the
ability of the photovoltaic pumping system to give desired
Tank
water flow in accordance to the user needs.
Phyd
Ipv MPPT Controller The reference speed is calculated from a reference power
(FLC and P&O) which is function on the water flow. The reference power is
Vpv obtained from the available maximum photovoltaic power and
Fig. 5.Diagram power for the studied system the batteries which compensate the power deficit to provide a
continuous delivery of energy to the motor pump. Simulation
The induction machine parameters are given in Table 6 results using the vector control strategy are given. The flux
TABLE6. and Vdcref reference values are applied ( d ref = 0.7Wb;
Induction machine parameters Vdcref =465V). The control strategy is tested through the
Nominal power PN 1.5 (KW) variations of insolation. The solar radiation varies up from 500
Nominal current Isn 5.2/3 (A) to 1000W/m2 (fig.7.).
Nominal voltage 220/380 (V) 160

Frequency f 50 (Hz) 140

Number of poles pairs p 2 120

Rated speed Nn N=1460 (tr/mn) 100

vitesse (N.m)

80

We make error calculation of the flow pump at steady state for 60

the two MPPT methods (Table 5.). We use the following 40

e s ti m e
r f r e n ce

expressions: 20

qv ( dsir ) Qmppt 0
0 0 .2 0 .4 0.6 0 .8 1
t (s)
1 .2 1.4 1 .6 1 .8 2

Q % 100 Fig.7 Reference ref and estimated est speed

qv ( dsir )
(20) 0 .8

0 .7

And we calculate the pumped efficiency by: 0 .6

P 0 .5

% h 100
flux (Web)

0 .4 r f r e nc e

Pmec 0 .3
es ti m

(21)
We can remark, that MPPT strategies improve the pumping 0 .2

system compared to a direct coupling. The FLC gives us a fast 0 .1

response compared to P&O method which requires much time 0

0 0 .2 0 .4 0 .6 0.8 1
t (s )
1.2 1 .4 1.6 1 .8 2

to track the Maximum Power Point (MPP). Fig.8 Rotor flux rd for(1000W/m and 500W/m2) 2
1200
10

1100
8

1000
6

900 4
Eclairement (W/m 2)

courant statorique (A)

800 2

700 0

h=6m -2
600
h=12m
h=9m
-4
500

-6
400

-8
300
1 1.5 2 2. 5 3 3.5 4 4.5 5 5.5 6
Dbit (m3 /s)
-3 -10
x 10 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2
t (s)

Fig.6 Water Flow for different heads Fig.9 Stator currents for (1000W/m2 and 500W/m2)

5
 Journal of Electrical Engineering
www.jee.ro

10

8
The FLC measures instantaneously PV voltage and current
Is d

7
Is q
variations and determines quickly the optimal increment
6
required to have the operating voltage for tracking the MPP
courant (A)

5
even when the operating environmental conditions change
4

3
rapidly. Fig. 10 shows the DC voltage waveform in the output
2
of the DC converter. We note that the DC bus voltage kept
1 constant and follows its reference Vdcref whatever insolation
0
0 0.2 0.4 0 .6 0.8 1
t (s)
1.2 1 .4 1.6 1 .8 2 variations.
Fig.10 Currents waveforms Id,iq
18

16
VI. CONCLUSION
14
In this paper, we have applied two MPPT methods (P&O, and
12

10
FLC) to a photovoltaic pumping system with field oriented
control (FOC). An application is made to satisfy water needs
couple (N.m)

6 of a family. The simulation results show that the control with

4
FLC method is more efficient in terms of stability, precision
2

0
and speed to reach the maximum power point
-2
0 0.2 0 .4 0 .6 0.8 1 1 .2 1.4 1.6 1 .8 2
t (s)

Fig.11 Electromagnetic torque VII. REFERENCES

x 10
-3
[1] Chenni, R. Zarour L., Bouzid A. and Kerbache T.,
6

Comparative study of photovoltaic pumping systems using a

5 permanent magnet synchronous motor (PMSM) and an
asynchronous motor (ASM) , Rev. Energ. Ren. vol. 9 , pp.17
4
28, 2006,
debit (Web)

3 [2] Lalouni S, Idjdarene K, Rekioua D, Vector Control based on

fuzzy logic controller of Induction Motor for Photovoltaic
2

Pumping System, in Proc. IREC 2012, Tunis, 20-22, 2012,

1 pp:1-6
[3] Salas V., Olias E., Barrado A., Lazaro A, Review of the
0
0 0 .2 0.4 0.6 0 .8 1
t (s)
1 .2 1 .4 1 .6 1 .8 2
Maximum Power Point Tracking Algorithms for Stand-Alone
Fig.12 Water Flow Photovoltaic Systems, Solar Energy Materials & Solar Cells,
vol. 90, no.11, pp.1555-1578, 2006.
500 [4] Lalouni S, Rekioua D, Rekioua T, Matagne E, Fuzzy logic
450
control of stand-alone photovoltaic system with battery storage,
400
Journal of Power Sources, vol.193, no.2, pp. 899907, 2009.
350
[5] Rekioua D., Bensmail S., Bettar N., Development of hybrid
300
photovoltaic-fuel cell system for stand-alone application, (2014)
V (V)

250
International Journal of Hydrogen Energy, 39 (3) , pp. 1604-1611.
dc

200 avec optimisation

Reference
rfrenc e Vdcref
sans optimis ation
[6] Patcharaprakiti N., Premrudeepreechacharn S., Sriuthaisiriwong
150

100
With FLC Y, Maximum power point tracking using adaptive fuzzy logic
50
Without MPPT control for grid connected photovoltaic system. Renewable
0
Energy, vol. 30, no.11, 2005, pp.1771-1788
0.5 1 1.5 2 2.5 3

460
t (s)
[7] Terki A., Moussi A., Betka A., Terki N., An improved
efficiency of fuzzy logic control of PMBLDC for PV pumping
450
system, Applied Mathematical Modelling, Vol.36, no.3, pp.
934944, 2012.
440 [8] Betka A. and Moussi A., Performance Optimization of a
Photovoltaic Induction Motor Pumping System, Renewable
430 Energy, vol. 29, pp. 2167 2181, 2004.
[9] Makhlouf M., Messai F., Benalla H., vectorial command of
induction motor pumping system supplied by a photovoltaic
Fig.13. DC bus voltage generator, journal of electrical engineering, vol. 62, no.1, pp:
310, 2011.
[10] Rekioua D, Matagne E., Optimisation of Photovoltaic Power
Systems: Modelization, Simulation and Control, Springer 2012.
[11] Rekioua D., Achour A.Y., Rekiouaa T., Tracking power
photovoltaic system with sliding mode control strategy, (2013)
Energy Procedia, 36 , pp. 219-230.