1. Experimental Comparison of Performances of Grid
Connected Small Wind Energy Conversion Systems
by
Md. Arifujjaman, M. T. Iqbal and J. E. Quaicoe
REPRINTED FROM
WIND ENGINEERING
VOLUME 34, NO. 6, 2010
MULTI-SCIENCE PUBLISHING COMPANY
5 WATES WAY • BRENTWOOD • ESSEX CM15 9TB • UK
TEL: +44(0)1277 224632 • FAX: +44(0)1277 223453
E-MAIL: mscience@globalnet.co.uk • WEB SITE: www.multi-science.co.uk

2. Experimental Comparison of Performances of Grid
Connected Small Wind Energy Conversion Systems
Md. Arifujjaman, M. T. Iqbal and J. E. Quaicoe
Graduate Student,Associate Professor, Professor Faculty of Engineering and Applied Science,
Memorial University of Newfoundland, St. John’s, NL Canada A1B 3X5
E-mail: mda04@mun.ca
WIND ENGINEERING VOLUME 34, NO. 6, 2010 PP 651–672 651
ABSTRACT
This paper presents an experimental comparative study of performances in terms of
efficiency on two possible grid connected small wind turbine systems. One of the systems is
based on Permanent Magnet Generator (PMG) and the other system is based on Wound
Rotor Induction Generator (WRIG). The Power Conditioning System (PCS) for grid
connection of the PMG-based system requires a rectifier, boost converter and a grid-tie
inverter, while the WRIG-based system employs a rectifier, a switch and an external
resistance in the rotor side with the stator directly connected to the grid. Experimental test
benches for both systems are implemented using Wind Turbine Emulator (WTE) and
Maximum Power Point Tracking (MPPT) control strategy. The procedure for calculate the
PCS losses, energy capture and energy loss is presented. The expected efficiency of the
systems is investigated for wind distributions at eight sites in Newfoundland and Labrador. It
is shown that a WRIG-based system could provide 5% higher efficiency in contrast to a PMG-
based system and could be an optimum alternative in the small wind energy domain.
1. INTRODUCTION
In the small wind energy domain, the existing knowledge of the system of a Permanent
Magnet Generator (PMG)-based small wind turbine system design and performance is quite
rich. In sharp contrast, studies with emphasis on the system of Wound Rotor Induction
Generator (WRIG)-based small wind turbine system are very few. This is also factual because
the market penetration of the PMG-based system is promising than the WRIG-based
system [1]. A main reason for such a depressing picture of the WRIG-based system in contrast
to the PMG-based system is the ability of the PMG-based system to operate in a full variable
speed operation with a more efficient manner. Variable speed operation is an attractive
feature for wind turbines for a number of reasons, including reduced mechanical stress,
increased energy capture and, not least, controllability, which is a primary concern for the
grid connection of wind turbines [2, 3]. Recent advances in power electronics have been a
leading cause in the dramatic advancement of the PMG-based system. However, despite such
rapid growth of the PMG-based system, it is difficult to predict the future for both systems. This
is because ideas borrowed from other fields or other applications could have profound effects
on future penetration. In this context, more investigation is essential on the PMG and

3. WRIG-based systems not only on the basis of performance of power electronics of both
systems but also with respect to the other aspects that significantly affect the global need of a
wind energy conversion system. Indeed, performance of the individual systems is not unique;
the author has, however, not found in the literature any evaluation of the experimental
evaluation of performances of the PMG and WRIG-based system that could significantly
dictate the future research to adapt an optimum system.
Performance comparison with experimental results with a concern on power electronics
losses of different wind energy conversion system is rare to found in the literature. Most of the
attempts usually measure performance of wind energy conversion system based on an
optimized controller [4, 5], power quality, [6–9] or reliability aspect [10–13]. Indeed comparison
of performances by considering the power electronic losses are available, however, the
experimental evaluation is always left unanswered [14–20]. It is well understood that
simulation provides preliminary analysis and prediction of any system performance.
However, in order to validate the results found by the simulation, laboratory experimentation
is essential. Although losses in the power conditioning system are of great interest during
calculation of the efficiency of a system and are strongly dependent on the wind distributions
and can be significant [15].
This paper describes the experimental implementation and determination of the
efficiency for the grid connected PMG and WRIG-based small wind turbine systems. The
Power Conditioning Systems (PCSs) for grid connection of the PMG-based variable system
requires a rectifier, boost converter, and a grid-tie inverter (Fig. 1). On the other hand, the
WRIG-based variable speed system can employ a rectifier, a switch, and an external
resistance in the rotor side, while the stator is directly connected to the grid (Fig. 2). The
power generation and power conditioning systems losses for both systems are investigated
first in this study. Afterwards, the annual energy capture, annual energy loss and efficiency
for the wind speed information of eight test sites in Newfoundland and Labrador, Canada:
Battle Harbour (BH); Cartwright (CW); Little Bay Island (LB); Mary’s Harbour (MH); Nain
(NA), Ramea (RA); St. Brendan’s (SB); and St. John’s (SJ). Finally, a comparison is presented
in terms of the efficiency in order to determine an optimum surrogate system.
This paper is organized as follows: The first section is a literature review and a short
overview of the work presented in this paper. In the second section the development of the
PMG and WRIG-based systems with required PCS is described along with their Maximum
Power Point Tracking (MPPT) control strategy. The third section describes a procedure to
calculate the efficiency of the systems using the wind distributions, while the fourth section
presents the power loss calculation of the components. The test results from laboratory test
benches of the two systems are presented in the fifth section. Finally, the findings of the
investigations are highlighted in the conclusion.
2. DESCRIPTION OF THE TEST BENCH
The PMG and WRIG-based systems considered in this study employ the same wind turbine,
requiring the development of a small Wind Turbine Emulator (WTE) for both systems.
Moreover, the implementation of the test benches requires associated PCSs for both systems
as well as MPPT control strategy.
The overall arrangements of the test bench are shown in Fig. 3 and each main component
is explained in further subsections.
652 EXPERIMENTAL COMPARISON OF PERFORMANCES OF GRID CONNECTED
SMALL WIND ENERGY CONVERSION SYSTEMS

4. 2.1. The Developed Wind Turbine Emulator
For this study, a furled wind turbine emulator is developed based on a DC motor and three
main criteria for acceptable performance. Firstly, representation of the furling control and
resulting dynamics; secondly, limiting the initial armature current of the DC motor; and
thirdly, tracking the theoretical optimum shaft speed of the wind turbine rotor by the DC
motor. In view of these, the furling action and resulting dynamics are incorporated in the
emulator with the use of a PC based wind turbine model [15].
WIND ENGINEERING VOLUME 34, NO. 6, 2010 653
Small wind
turbine
Control circuitry
3 phase bridge
rectifier
Boost converter Inverter
Grid
PMG
Idc1 D1 D3 D5
D4 D6 D2
Vdc
Idc Idc2
Vdc2
LD D
SW5
SW1
SW4
SW3
i0
SW2
Gate drive circuit Gate drive circuit
MPPT controllerSystem variable as
necessery for the controller
Figure 1: A PMG-based small wind turbine system.
Small wind
turbine
Slip
ring
Control circuitry
3 phase bridge
rectifier Grid
WRIG
ID1
VDC
IDC LD
VD
Re=
R (1-d)
Switch
SW
D1 D3 D5
D4 D6 D2
Gate drive circuit
MPPT controller
System variable as
necessery for the controller
Figure 2: A WRIG-based small wind turbine system.

5. An initial current limiting controller is designed that ensures slow speed up of the motor,
thus limiting the armature current. In order to follow the optimum shaft speed of the wind
turbine rotor, the reference shaft speed is determined from the DC motor armature current and
compared with the actual shaft speed. The difference of these two speeds introduces an error
and the error is decreased as quickly as possible through a specially designed controller. The
experimental test bench structure of the small WTE developed in this study is shown in Fig. 4.
2.2. Grid Connected PMG-based Small Wind Turbine System
The grid connected PMG-based small wind turbine system is composed of an emulator, a PMG,
a 3-phase bridge rectifier, a Boost Converter (BC) and an inverter. The BC and inverter
together is considered as a single unit (BCI) in this study. The BCI unit is programmed with the
wind turbine model specific power curve [15]. The experimental test bench structure of the
PMG-based system is shown in Fig. 5. An optimum tip-speed-ratio control strategy, as shown in
Fig. 6 is implemented.
2.3. Grid Connected WRIG-based Small Wind Turbine System
The grid connected WRIG-based system is composed of an emulator, a WRIG, a 3-phase
bridge rectifier, and a variable resistance. The model specific wind turbine power curve used
for the WRIG-based system was identical to the PMG-based system, which incorporates the
MPPT control to extract maximum power from the wind turbine. The MPPT control is based
on an optimum tip-speed-ratio control strategy adapted for the WRIG-based system. The
experimental test bench structure of the WRIG-based system is presented in Fig. 7.
The convenient parameter that can be used to modify the power speed characteristics of
the WRIG is the external rotor resistance and the control strategy is shown in Fig. 8.
3. EFFICIENCY CALCULATION
In order to calculate the annual energy capture, annual energy loss and efficiency of the
systems, the relation between the wind speed and the time in one year during which this wind
654 EXPERIMENTAL COMPARISON OF PERFORMANCES OF GRID CONNECTED
SMALL WIND ENERGY CONVERSION SYSTEMS
Wound rotor
induction
generator
(WRIG)
Power
conditioning
system (PCS)
Grid
Grid
Small wind
turbine emulator
Permanent
magnet generator
(PMG)
WRIG-based small wind turbine system
PMG-based small wind turbine system
Power
conditioning
system (PCS)
Figure 3: Basic structure of the PMG and WRIG-based system test bench.

6. speed occurs is needed. There are two approaches to reproduce this characteristic. Firstly,
employ an ideal distribution, which approximates the real wind distribution for a site in a year,
and secondly, use of yearly real wind information for the site under consideration. This study
considered the second approach as the efficiency for any wind energy conversion systems
can be calculated more precisely than the first approach, which employs an ideal wind speed
distribution. In the second approach, as soon as the real wind information is available,
considering a bin of 1 m/s, the time in one year during which the wind speed lies within the
wind speed band is then found. Once the wind speed distribution and the power curve is
known for discrete wind speed [15], the energy capture by a particular wind speed is
determined by a multiplication of the wind speed distribution curve and power curve on a bin
by bin basis. It can be expressed as
WIND ENGINEERING VOLUME 34, NO. 6, 2010 655
PC based wind
turbine model
AC supply
Controlled
voltage Phase
controlled
relay
Bridge
rectifier
Circuit
breaker
Filter
DC motor
armature
DC field
supply
Armature
current
Lab master
I/O board
RPM A/D
D/A
Armature current feedback
RPM feedback
Figure 4: Schematic of the wind turbine emulator test bench.
Small wind
turbine emulator
Permanent magnet
generator (PMG)
Circuit
breaker
3-phase bridge
rectifier
Circuit
breaker
Diversion
load
Power harmonics
analyzer
Maximum power point
tracking (MPPT) control PC
Power harmonics
analyzer
Data collection PC BCI unit
Inverter
Boost
converter
Circuit
breaker
Grid
Idc3
Vdc3
ia1
ia vac
ia1
va1b1
b1
ia
a
b
c
a1
Wind speed
DC
motor
+
−
V
A
Figure 5: Schematic of the PMG-based system test bench.

7. (1)
The energy loss for the wind speed of i m/s, El,i (wi ) can be found by replacing Pg,i (wi)
in (1) with power losses for the PMG and WRIG-based system respectively as described in
E w P w f wg i i g i i dis i, ,( ) ( ) ( )=
656 EXPERIMENTAL COMPARISON OF PERFORMANCES OF GRID CONNECTED
SMALL WIND ENERGY CONVERSION SYSTEMS
w5 m/s
w4 m/s
w3 m/s
w2 m/s
Power,Pwatt
w1 m/s
P7
P8
Vdc1
ω1 ω2 ω6
Vdc2
ω3
Vdc 3
ω4
Vdc4
ω5
Vdc5 Vdc6 Vdc7 Vdc8 Vdc9
Maximum power locus
P6
P5
P4
P3
P2
P1
ω8ω7 ω9
Shaft speed, ω rpm
Rectified DC voltage, Vdc3volt
w5 >w4 > w1
ω ω ω ω
ω
ω ω ω ω ω
Figure 6: Control strategy for the PMG-based system.
Small wind
turbine emulator
Circuit
breaker
Power harmonics
analzyer
Power harmonics
analyzer
Power harmonics
analyzer
Data collection PC
Variable
resistance
3-phase bridge
rectifier
Wound rotor
induction generator
(WRIG)
Circuit
breaker Grid
ib3
VDC3
ia3
a3
b3
c3
c4 b4 a4
Wind speed
IDC3
ia4
ia3
ia4
va3c3
ib3 vb3c3
va4c4
DC
motor
+
−
A
V
Figure 7: Schematic diagram of the WRIG-based system test bench.

8. Section 4. Finally, the annual energy capture and annual energy loss in the year found by the
summing up all the discrete energy loss or mathematically
(2)
(3)
The efficiency of the systems is then calculated as,
(4)
4. POWER LOSS CALCULATION
4.1. Power Loss in PMG-based System
The instantaneous power of a PMG at time t, P(t), is expressed as
(5)
where va, vb and vc are the instantaneous values of PMG output voltages for phase a, b and c
respectively and ia, ib and ic are the instantaneous values of PMG output currents. Under the
assumption of balanced system, at any instant the summation of all instantaneous currents is
zero and can be written as
(6)i i ia b c+ + = 0
P t v i v i v ia a b b c c( ) = + +
η =
−
×
E E
E
g l
g
100%
E E wl l i i
i w
wn
=
=
∑ , ( )
1
E E wg g i i
i w
wn
=
=
∑ , ( )
1
WIND ENGINEERING VOLUME 34, NO. 6, 2010 657
w5 m/s
w4 m/s
d1 d2 d3
w3 m/s
w2 m/s
Power,Pwatt
w1 m/s
P7
P8
Maximum power locus
P6
P5
P4
P3
P2
P1
Shaft speed, ϖ rpm
w5 >w4 > w1
d3 > d2 > d1
ω1 ω2 ω6ω3 ω4 ω5 ω8ω7 ω9ω ω ω ω ω
ω
ω ω ω ω
Figure 8: Control strategy for the WRIG-based system.

9. Substituting the value of ic in (5), the power is expressed as
(7)
As the generator phases are ‘balanced’, then both terms in (7) are equal, hence
(8)
Averaging over one period T of (8) will give the average power. So the output power of the
PMG, is expressed as
(9)
An approximation uses a digital oscilloscope to obtain instantaneous values of Vac = va − vc
and Ia = ia. If there are n samples in one period then by definition of the integral, (9) becomes
(10)
The output DC power from the 3-phase bridge rectifier can be expressed as
(11)
Subtracting (11) from (10), the rectifier power loss is found as
(12)
The PMG-based small wind turbine system has a single-phase output with phases a1, b1 and
a neutral (ground). As a consequence, a similar kind of formulation expressed by (5) to (10) is
derived to compute the power flowing into the grid, , and can be written as
(13)
Equation (13) expresses the power flowing into the grid by the actual voltage and current;
however, the BCI unit itself is able to calculate the power flowing into the grid by its own data
acquisition system and denoted in this research as commercial or expected value. The
commercial power production values from the BCI unit is also stored in the PC and denoted by
. These commercial power output values are recorded in order to observe the
commercial power loss values. This is of particular importance because the commercial and
experimental outcomes from a system will be clearly identified and any misconception about
a system performance will certainly be evaluated.
The BCI unit loss is computed by subtracting (12) from (13) and is given by (14), which
represents the experimental power loss of the BCI unit. In contrast to the experimental power
loss, subtracting the commercial output power from the BCI unit will provide the commercial
power loss of the BCI unit as expressed in (15).
The BCI loss is computed by subtracting (13) from (12) and the commercial power loss of
the BCI unit is expressed by (15). Here commercial power loss refers to the power loss found
from the commercial BCI.
Pout_com,grid
PMG
P
n
V Iout_exp,grid
PMG
a b a
n
≅ ∑
1
1 1 1
Pout_exp,grid
PMG
P P Pt rec
PMG
out gen
PMG
out rec
PMG
, , ,
= −
P V Iout rec
PMG
dc dc,
= 3 3
P
n
V Iout gen
PMG
ac a
n
,
≅ ∑
1
2
P
T
P t dt
T
v v iout gen
PMG
cycle
a c a
cycl
,
( ) ( )= = −∫
1 1
2
ee
dt∫
Pout,gen
PMG
P t v v i v v ia c a b c b( ) ( ) ( )= − = −2 2
P t v v i v v ia c a b c b( ) ( ) ( )= − + −
658 EXPERIMENTAL COMPARISON OF PERFORMANCES OF GRID CONNECTED
SMALL WIND ENERGY CONVERSION SYSTEMS

10. (14)
(15)
The experimental and commercial total power losses for the PMG-based system can be
expressed by (16) and (17) respectively.
(16)
(17)
Afterwards, the efficiency of the individual component is calculated. Equation (18)
expresses the rectifier efficiency, while (19) and (20) represent the experimental and
commercial efficiency of the BCI unit respectively. The composite experimental and
commercial efficiency of the PCS is also calculated and expressed by (21) and (22)
respectively.
(18)
(19)
(20)
(21)
(22)
4.2. Power Loss in WRIG-based System
The three phases, a3, b3 and c3 of the stator of the wound rotor induction generator are
directly connected to the grid. Two power harmonic analyzers are used to determine the total
real power. The power flowing into the grid can be written as
(23)
(24)
The total real average power flowing into the grid is given by
(25)
The rotor of the generator is star connected and balanced. As a result, a similar approach
described by (5) to (10) is used to determine the power losses into the rotor and results in
P P Pout_exp,grid
WRIG
out_ac,grid
WRIG
out_bc,
= + ggrid
WRIG
P
n
V Iout bc grid
WRIG
b c b
n
_ ,
≅ ∑
1
3 3 3
P
n
V Iout ac grid
WRIG
a c a
n
_ ,
≅ ∑
1
3 3 3
η η ηcom,composite rec com,BCI_unit= ×
η η ηexp,composite rec exp,BCI_unit= ×
ηcom,inv_unit
P
P
out_com,grid
PMG
out rec
PMG=
,
ηexp,inv_unit
P
P
out_exp,grid
PMG
out rec
PMG=
,
ηrec
P
P
tout rec
PMG
tout gen
PMG= ,
,
P P Pt com
PMG
t rec
PMG
t com BCI unit
PMG
_ , _ , _
= +
P P Pt_exp
PMG
t,rec
PMG
tt_exp,BCI unit
PMG
= + _
P P Pt com BCI unit
PMG
out com grid
PMG
out re_ , _ _ , ,
= − cc
PMG
P P Pt_exp,BCI unit
PMG
out_exp,grid
PMG
out,re_
= − cc
PMG
WIND ENGINEERING VOLUME 34, NO. 6, 2010 659

11. (26)
where Va4c4 is the line-to-line voltage of the rotor Ia4 is the line current of the rotor
The output power of the 3-phase bridge rectifier can be expressed as
(27)
Subtracting (27) from (26), the rectifier power loss is found as
(28)
The electrical and frictional losses of the slip rings are determined based on the speed, ω of
the generator and expressed by (29) and (30) respectively, while the total slip ring loss is
expressed by (31). The sum of (26) and (31) reflects the total power loss of the WRIG-based
system and is presented by (32).
(29)
(30)
(31)
(32)
5. TEST RESULTS
The grid connected small wind turbine systems described in this study were implemented and
tested in the laboratory environment. First, the emulator test results are presented and
afterwards, the power losses for both systems are described based on the computation
procedure described in section 3. Finally, annual energy capture, annual energy loss and
efficiency are compared based on the wind distributions for Battle Harbour (BH), Cartwright
(CW), Little Bay Island (LB), Mary’s Harbour (MH), Nain (NA), Ramea (RA), St. Brendan’s
(SB), and St. John’s (SJ) of Newfoundland and Labrador. In order to reduce the volume of
results that can be presented for the 8 different sites and 8 different wind speeds for each
system, representative results for each system are presented in the paper.
5.1. Small Wind Turbine Emulator
The wind turbine emulator developed in the laboratory environment was subjected to a
changing wind speed w m/s. A pseudorandom wind speed profile was developed as follows
(Fig. 9a): the input to the emulator was a series of steps, each one of same duration. Although
real wind does not occur with such abrupt slopes, a series of steps is a standard testing signal
which permits a clear interpretation of the system behavior.
The response of the furling angle and expected dynamics are shown in Fig. 9b. During the
wind speed increase, within 10 seconds, the rotor reached a stable state at a furling angle
corresponding to the value used for the furling model [15]. This settling time is required in
order to avoid any excessive overloading of the mechanical part of the wind turbine.
P P P Pt,exp
WRIG
t,rec
WRIG
t,ext
WRIG
t,sliprin
= + + gg
WRIG
P P Pt,slipring
WRIG
elec slipring
WRIG
fric_sl
= +_ iipring
WRIG
P Kfric_slipring
WRIG
= δω
P Kelec_slipring
WRIG
= ωω
P P Pt,rec
WRIG
out_exp,grid
WRIG
t ext
WRIG
= − ,
P V It,ext
WRIG
DC DC= 3 3
P
n
V It rotor
WRIG
a c a
n
,
≅ ∑
1
2 4 4 4
660 EXPERIMENTAL COMPARISON OF PERFORMANCES OF GRID CONNECTED
SMALL WIND ENERGY CONVERSION SYSTEMS

12. The limitation of the armature current is of importance in order to protect the motor
armature. The variation in armature current characteristic corresponding to the change in
wind speed was recorded and averaged to generate a uniform distribution over the entire
wind speed range. It can be seen that the proposed controller limits the amplitude of the
armature current of the DC motor below 4 amperes (Fig. 9c). With the increase in wind speed,
the motor started to draw more current from the main, while no unwanted overshoot of the
current was observed. The experimental results thus prove that the current limitation
functions well, and the controller has a good performance, independent of the speed and
motor dynamics.
In the emulator controller, the controller forced the actual speed of the emulator tracking
the reference rotor speed. The reference speed, actual speed and error between them in rpm
of the emulator are shown in Fig. 9d. The trace shows that after some transients, the speed of
the DC motor always follows the reference speed of the controller. Therefore, the error is
always zero at steady state. It can be observed that the transition time takes less than 10
seconds at the startup of the emulator before the algorithm reaches its steady state.
Examination of the experimental results described above shows that the three criteria that
define acceptable performance of the emulator, namely, achieving the furling control and
expected dynamics; limiting the initial DC motor armature current; and tracking of the
theoretical speed of the rotor by the DC motor were achieved.
5.2. Power Loss in the Systems
5.2.1. Power Loss of the PMG-based System
The power loss determination of the PMG-based system based on the experimental voltage
and current waveform at all wind speeds were performed according to the procedure
described in section IV. Fig. 10a, Fig. 10b and Fig. 10c shows instantaneous line-to-line output
voltage, vac line current, ia and corresponding instantaneous power of the PMG when the PMG
is operating at 25 Hz, i.e., 749 rpm and subjected to a wind speed of 6 m/s. It is observed that the
generated voltage waveform is sinusoidal in nature, with saturated maximum and minimum
values. This is mainly due to the mechanical arrangement of the generator, i.e., number of
poles and the diameter of the generator. There is not enough width for the magnets to cover
3 coils at same time for the sinusoidal waveform to form properly due to the fast rotation of the
stator magnet transversing across the stator winding, resulting in saturation of the maximum
values. Furthermore, in order to quantify the distortion of the voltage and current, the
WIND ENGINEERING VOLUME 34, NO. 6, 2010 661
Figure 9: Variation of the emulator characteristics, a) Wind speed, b) DC motor armature current,
c) Furling angle and expected dynamics, d) Shaft speed.

13. harmonic content and Total Harmonic Distortion (THD) of the generated PMG voltage and
current were obtained. The results are summarized in Fig. 11a and Fig. 11b. The fundamental
components were omitted in these figures, in order to highlight the harmonic content. From
the figures it is observed that 5th, 7th, 11th, 13th, 17th and 19th harmonics are significant for the
voltage, while the 5th, 7th, 11th, 13th, 17th, 19th and 23rd harmonics are significant for the
current. The total harmonic distortion (THD) was determined to be 5.89% and 41.4% for
the voltage and current respectively, which is quite high. The presence of the high and low-
order harmonics in the voltage and current is obviously not undesirable. High-order
harmonics can interfere with sensitive electronics and communication systems, while low-
order harmonics can cause over heating of the generator and conductors which is not
undesirable for a small wind turbine system. In addition, regulation of the harmonics will
increase the cost of the system to a great extent.
The calculated instantaneous power, (Fig. 10c) corresponding to the PMG output
voltage and current is averaged as 335W, while the rectified DC power, at the output of
the 3-phase bridge rectifier is found as 326 W using (11). As a result, the 3-phase bridge rectifier
power loss, is calculated as 9 W by (12). The grid voltage, va1b1 current, ia1 and power are
shown in Fig. 12a and Fig. 12b. A frequency of 60.5 Hz is observed. The harmonic content for the
voltage and current is presented in Fig. 13a and Fig. 13b respectively. It can be seen that the 5th,
7th, 11th, 17th, and 23rd harmonic content of the voltage is significant, while the 5th, 7th, 11th,
13th, 17th, 23rd, 25th, 29th and 31st is significant for the current. The THD for the voltage and
current is 2.68% and 14.3% respectively. The high THD of the current is because of the
operation of the BCI unit at a low power level. The average power, flowing into the grid
is calculated as 238 W from the instantaneous power curve given in Fig. 12c, which is
quite close to the rated maximum power of 254 W at 6 m/s wind speed. However, the
commercial value of the grid power, from the BCI unit is found to be 242 W,Pout_com,grid
PMG
Pout_exp,grid
PMG
Pt,rec
PMG
Pout,rec
PMG
Pout,gen
PMG
662 EXPERIMENTAL COMPARISON OF PERFORMANCES OF GRID CONNECTED
SMALL WIND ENERGY CONVERSION SYSTEMS
0.0050
−200
0
Voltage(V)
200
(a)
0.01 0.015 0.02 0.025 0.03 0.035 0.04
0.0050
−5
0
Current(A)
5
(b)
0.01 0.015 0.02 0.025 0.03 0.035 0.04
0.0050
−500
0
500
Power(W)
1000
(c)
0.01 0.015 0.02
Time (s)
0.025 0.03 0.035 0.04
Figure 10: Instantaneous value of the PMG output a) Line-to-line voltage vac , b) Line current ia,
c) Instantaneous power.

14. which is slightly higher than the value found by the experimentation. The experimental,
and commercial, power losses of the BCI unit are 89 W and 85 W
respectively. The experimental and commercial values of grid power are less than the
generator power due to the system losses.
The calculation procedures described above based on the experimental voltage and
current waveform is applied to all wind speeds and power losses for each component are
determined. Fig. 14a depicts the power losses of the rectifier when the system is subjected to
Pt_com,BCI_unit
PMG
Pt_exp,BCI_unit
PMG
WIND ENGINEERING VOLUME 34, NO. 6, 2010 663
4
6
2
0
0 5
Voltage
harmoniccomponent(%)
10 15
(a) Harmonic
20 25 30 35
40
20
30
10
0
0 5
Current
harmoniccontent(%)
10 15
(b) Harmonic
20 25 30 35
Figure 11: Harmonic content of the PMG output a) Voltage, b) Current.
0.0020
−500
0
Voltage(V)
500
(a)
0.004 0.006 0.008 0.01 0.012 0.014 0.016
0.0020
−3
0
1.5
−1.5
Current(A)
3
(b)
0.004 0.006 0.008 0.01 0.012 0.014 0.016
0.0020
0
300
Power(W)
600
(c)
0.004 0.006 0.008 0.01 0.012
Time (s)
0.014 0.016
Figure 12: a) Instantaneous grid a) Line-to-line voltage va1b1, b) Line current ia, c) Instantaneous power.

15. 664 EXPERIMENTAL COMPARISON OF PERFORMANCES OF GRID CONNECTED
SMALL WIND ENERGY CONVERSION SYSTEMS
the same wind speed profile that was applied to the emulator. The wind speed of 6 m/s is
included in the presentation of the results for convenience of observing the behavior of
system. It can be seen that the rectifier power loss increases with increasing wind speed and
reaches a maximum of 52 W at a wind speed of 13 m/s. The BCI unit power losses are depicted
in Fig. 14b. A significant difference between the experimental and commercial values is
observed and the circumstances become more severe when the BCI unit operates at high
wind speed. At high wind speed, the BCI unit operates with higher values of voltage and
2
1
0
0 5
Voltage
harmoniccomponent(%)
10 15
(a) Harmonic
20 25 30 35
10
5
0
0 5
Current
harmoniccomponent(%)
10 15
(b) Harmonic
20 25 30 35
Figure 13: Harmonic content of the grid output a) Voltage, b) Current.
6
Exp
Com
5
0
30
Rectifier
powerloss(W)
60
(a)
(b)
(c)
7 8 9 10 11 12 13 14
65
0
200
100
Inverterunit
powerloss(W)
300
7 8 9 10 11 12 13 14
65
0
200
100
Total
powerloss(W)
300
7 8 9
Wind speed (m/s)
10 11 12 13 14
Exp
Com
Figure 14: Characteristic of the PMG-based system a) Rectifier power loss, b) Inverter unit power loss,
c) Total power loss.

16. WIND ENGINEERING VOLUME 34, NO. 6, 2010 665
current and consequently power losses increase and so does the difference between
experimental and commercial loss. The sum of the losses of the rectifier and BCI unit provides
the total power losses (commercial and experimental) of the PCS as illustrated in Fig. 14c,
while Fig. 5.28d incorporates power values at several stages of the system. The PMG output
power has the highest value as expected. The theoretical maximum power of the system is
presented (15a) and can be seen that the experimental and commercial power nearly follows
the maximum power values for each wind speed and the corresponding rotational speed is
presented in Fig. 15b. It can be seen that for each wind speed, the system is able to produce
maximum power, while maintaining the optimum speed and ensuring the MPPT control
strategy. It should be mentioned that the experimental and commercial power values are not
exactly the same as the theoretical maximum power values; however, they behave logically.
Fig. 16a shows the efficiency characteristics of the rectifier and BCI unit in the case where the
wind speed is changed from low to high under the condition of maximum power transfer to
the grid. The measured efficiency of the rectifier ηrec is at least 97%, while the experimental
efficiency, ηexp,BCI_unit and commercial efficiency, ηcom,BCI_unit for the BCI unit is 86% and 91%
respectively. A lower experimental value is observed which is expected as the power losses of
the BCI unit from the experimental is higher than the commercial. The composite efficiency is
the product of rectifier and BCI unit efficiencies and illustrated in Fig. 16b. Both experimental,
ηexp,composite and commercial, ηcom,composite efficiency is more than 80% at high wind speed
situation. Indeed the experimental efficiency is lower in value than the commercial efficiency,
and the results suggest that the manufacturer’s stated efficiency may not be reliable. The
generalized claim of the manufacturers rarely mentions the operating condition of a system.
From the composite efficiency, it is observed that the efficiency drops at low wind speed, i.e.,
light load; a typical challenge in variable speed wind generation system.
5.2.2. Power Loss of the WRIG-based System
Figures 17a and 17b show the observed instantaneous waveforms of the line-to-line stator
voltages va3c 3, vb3c3 and line current ia3,ib3 respectively when the system is operating at
Experimental
Commercial
65
0
1000
500
Power(W)
1500
(a)
(b)
7 8 9 10 11 12 13 14
65
500
1500
1000
Speed(rpm)
2000
7 8 9 10
Wind speed (m/s)
11 12 13 14
Maximum
Experimental
Optimum
Figure 15: Variation of the PMG-based system a) Power, b) Speed.

17. 666 EXPERIMENTAL COMPARISON OF PERFORMANCES OF GRID CONNECTED
SMALL WIND ENERGY CONVERSION SYSTEMS
60.5 Hz in steady state. From the measurement result of the stator voltage and current, the
instantaneous values of the power are calculated and presented in Fig. 17c. The sum of
the average value of the generated stator electric power, is 243 W, which is quite
close to the rated maximum power of 254 W at 6 m/s wind speed. In addition, the stator
voltage and current is essentially sinusoidal, which is a desirable feature because harmonics
Pout_exp,grid
WRIG
90
100
80
70
60
65
Componentefficiency(%)
7 8 9 10 11 12 13 14
90
100
80
70
60
65
Compositeefficiency(%)
7 8 9 10 11
Wind speed (m/s)
12 13 14
Rectifier
BCI commercial
BCI experimental
Commercial
Experimental
(a)
(b)
Figure 16: Variation of the efficiency a) Component, b) Composite.
500
250
0
−250
−500
0 0.0045
Statorvoltage(V)
0.009
(a)
0.0135
a3c3
a3c3
b3c3
b3c3
b3
a3
0.018
5
2.5
0
−2.5
−5
0 0.0045
Statorcurrent(A)
0.009
(b)
0.0135 0.018
1000
500
0
−500
0 0.0045
Statorpower(W)
0.009
Time (s)
(c)
0.0135 0.018
Figure 17: Instantaneous value of stator a) Line-to-line voltage va3b3, (b) Line-to-line voltage vb3c3,
c) Line current ia3, d) Line current ib3, e) Line power , f) Line power .Pout_b c ,grid
WRIG
3 3
Pout_a c ,grid
WRIG
3 3

18. in the supply may produce adverse effects on the supply system. The experimental
instantaneous waveforms for the rotor voltage and current are presented in Fig. 18a and
Fig. 18b respectively when the generator is operating at 17.5 Hz, i.e., 527 rpm at super
synchronous speed. The rotor voltage is sinusoidal and is rich in harmonics, which is easily
notable from the harmonic of rotor voltage as presented in Fig. 19a. It can be seen that the high
as well as low-order harmonic contents are present; however, the amplitude as well as the THD
(11.9%) are low. In contrast to the voltage, the effect of commutation overlap is evident from
the waveform of the rotor current and is not sinusoidal. The associated harmonic content is
presented in Fig. 19b, and it can be seen that the 5th and 7th harmonics have the dominant
effect on the rotor current and the THD was found to be 23.7%. The THD for both voltage and
current is quite high. The power graph (Fig. 18c) produces an average instantaneous power
loss, into the rotor circuit which was amended to the corresponding slip as 40 W. The
power loss in the external rotor resistance was found to be 38 W and after calculation, the
rectifier power loss, was found to be 2W. The slip-ring electrical, and frictional,
losses were calculated based on the speed of the generator and were found to be 5 W
and 12 W respectively. The same procedure was repeated for the same wind profile applied to
the emulator. The results of the losses of the rectifier are presented in Fig. 20a. The external
rotor resistance losses are presented in Fig. 20b, while the electrical, frictional and total
losses of the slip ring are presented in Fig. 20c. The summation of all the losses is the total
power losses of the system and is depicted in Fig. 20d. Power losses increase with increase in
wind speed, which is an expected behavior from the system. In order to verify the MPPT
control strategy, the speed of the system (optimum and experimental) is recorded and
presented in Fig. 21a, followed by the corresponding power injection to the grid (maximum
and experimental) by the stator, in Fig. 21b. It is evident from this figure that the
experimental grid power and speed maintains a high agreement with the theoretical
maximum power and speed for each wind speed, which consequently ensures the MPPT
control strategy.
Pfric_slipring
WRIG
Pelec_slipring
WRIG
Pt,rec
WRIG
Pt,ext
WRIG
Pt,rotor
WRIG
WIND ENGINEERING VOLUME 34, NO. 6, 2010 667
0.01
Rotor
voltage(V)
0.02 0.03 0.04 0.05 0.060
0
−50
50
(a)
0.01
Rotor
current(A)
0.02 0.03 0.04 0.05 0.060
0
−5
5
(b)
0.01
Rotor
power(W)
0.02
Time (s)
0.03 0.04 0.05 0.060
0
100
−100
200
(c)
Figure 18: Instantaneous rotor output a) Line-to-line voltage va4b4, b) Line current ia4.

19. 5.3. Performance Characteristic of the Systems
One of the basic objectives of this study is to compare the performance characteristic over a
period of a year, i.e., overall efficiency for the two systems in a number of wind conditions. The
wind distribution for the considered sites is presented in Fig. 22 to Fig. 23. Afterwards, the
annual energy capture of the PMG and WRIG-based system is calculated as presented in
section 4. It should be mentioned that due to the introduction of commercial power loss of the
PCS, the annual energy capture, annual energy loss and efficiency will be different for the
PMG-based system, while the WRIG-based system is not subjected to such situation.
668 EXPERIMENTAL COMPARISON OF PERFORMANCES OF GRID CONNECTED
SMALL WIND ENERGY CONVERSION SYSTEMS
6
8
4
2
0
0 5
Rotorvoltage
harmoniccomponent(%)
10 15
(a) Harmonic
20 25 30 35
20
10
0
0 5
Rotorcurrent
harmoniccomponent(%)
10 15
(b) Harmonic
20 25 30 35
Figure 19: Harmonic content of the rotor output a) Voltage, b) Current.
65
0
25
Rectifier
loss(W)
50
7 8 9 10 11 12 13 14
14
14
14
(a)
65
0
200
Resistance
loss(W)
400
7 8 9 10 11 12 13
(b)
6
Frictional Electrical
5
0
20
Slipring
loss(W)
40
7 8 9 10 11 12 13
(c)
65
0
250
Total
loss(W)
500
7 8 9 10
Wind speed (m/s)
11 12 13
(d)
Figure 20: Characteristic of the losses of the WRIG-based system a) Rectifier, b) External rotor
resistance, c) Slip ring, d) Total.

20. Table 1 and Table 2 show the comparison of the efficiency for the PMG and WRIG-based
system respectively from 7 m/s to 13 m/s wind speeds for all eight sites. The tables include the
annual energy capture as well as the annual energy loss as a percentage of the annual energy
capture for both systems. It is evident that the commercial energy production of the PMG-
based system always shows a higher value than the experimental, which is because low
power values were observed in all experiments. Furthermore, the maximum power is
WIND ENGINEERING VOLUME 34, NO. 6, 2010 669
65 7 8 9 10 11 12 13 14
1000Power(W)
1500
(a)
Maximum
Experimental
500
0
65 7 8 9 10
Wind speed (m/s)
11 12 13 14
Speed(rpm)
2800
(b)
Optimum
Experimental
2400
2500
2600
2700
2300
Figure 21: Variation of the WRIG-based system a) Power, b) Speed.
6
0
600
No.ofhours
1200
7 8 9 10 11 12 13
(a)
6
0
400
No.ofhours
800
7 8 9 10 11 12 13
(b)
6
0
750
No.ofhours
1500
7 8 9 10 11 12 13
(c)
6
0
700
No.ofhours
1400
7 8 9 10
Wind speed (m/s)
11 12 13
(d)
Figure 22: Wind speed distribution for a) Battle Harbour (BH), b) Cartwright (CW), c) Little Bay Island
(LB), d) Mary’s Harbour (MH).

21. supposed to be the same for both systems; however, a slight variation in values was observed
and reflected on the annual energy capture values. This deviation is not unusual as it is very
difficult to inject the same power to the grid for both systems. Nevertheless, it should not affect
the calculation as more (less) power injection to the grid will produce more (less) power losses
on the required power electronics for each system. The annual energy loss for the WRIG-
based system is lower for all sites than the calculated annual energy loss of the PMG-based
system. But the commercial annual energy loss of the PMG-based system is lower than the
WRIG-based system at Battle Harbour and Ramea, while it remains almost the same for
Mary’s Harbour and Nain. However, it should be noted from the total power losses figures that
above 7 m/s, the total power loss of the WRIG-based system starts to increase, while it remains
670 EXPERIMENTAL COMPARISON OF PERFORMANCES OF GRID CONNECTED
SMALL WIND ENERGY CONVERSION SYSTEMS
6
0
200
No.ofhours
400
7 8 9 10 11 12 13
(a)
6
0
400
No.ofhours
800
7 8 9 10 11 12 13
(b)
6
0
400
No.ofhours
800
7 8 9 10 11 12 13
(c)
6
0
700
No.ofhours
1400
7 8 9 10
Wind speed (m/s)
11 12 13
(d)
Figure 23: Wind speed distribution for a) Nain (NA), b) Ramea (RA), c) St. Brendan’s (SB),
d) St. John’s (SJ).
Table 1: Performance characteristics of the PMG-based SWT system
Annual energy loss as
a percentage of the
annual energy capture______________________________
Annual energy Diode bridge Inverter unit Efficiency, η
capture [Wh] rectifier [%] [%] [%]_________________ ______________ ______________ ______________
Region Com Exp Com Exp Com Exp Com Exp
BH 1311280 1260084 4.43 4.61 19.03 23.87 76.52 71.50
CW 635363 613573 4.62 4.79 22.37 26.71 73 68.49
LB 1556458 1499049 4.66 4.84 20.49 25.11 74.83 70.04
MH 1247259 1200958 4.50 4.67 20.22 24.86 75.26 70.45
NA 377700 363818 4.57 4.74 20.59 25.19 74.83 70.06
RA 1081526 1039260 4.60 4.78 18.81 23.64 76.58 71.56
SB 831915 801319 4.60 4.78 20.58 25.19 74.80 70.02
SJ 1165376 1124061 4.50 4.66 21.54 26.01 73.94 69.31

22. low below 7 m/s. Therefore, from cut-in wind speed to 7 m/s wind speed, the total power losses
of the WRIG-based system are less than the PMG-based system. The efficiency of the systems
shows that the WRIG-based system maintains a 5% higher efficiency for the sites in contrast to
the PMG-based system provided that the efficiency values are considered based on the
experimental outcomes. As a result it is concluded that the WRIG-based system could be an
optimum surrogate in the small wind energy conversion area.
6. CONCLUSIONS
This paper has presented a comparison study on grid connected small PMG and WRIG based
wind energy conversion systems. A complete implementation of the systems is described and
the required maximum power point control strategy is defined for each system. It is found that
both of the systems are able to sustain the maximum power point control strategy, thus
ensuring the variable speed operation. It is found that experimentally a WRIG-based system
could provide 5% higher efficiency than a PMG-based system for eight sites in Newfoundland
and Labrador, Canada and consequently, can be considered to be a better option for small
wind energy conversion system.
ACKNOWLEDGMENTS
The Author would like to thank the National Science and Engineering Research Council
(NSERC) Canada for providing support for this research.
REFERENCES
1. Hansen, A.D. and Hansen, L.H. “Wind turbine concept market penetration over 10 years
(1995–2004),” Wind Energy, 10, 2006, 81–97.
2. Carlson, O., Hylander, J.K., and Thorborg, “Survey of variable speed operation of wind
turbines,” Proceedings in the European Union Wind Energy Conference (EUWEC),
1996, 406–409.
3. Zinger, D.S., and Muljadi, E., “Annualized wind energy improvement using variable
speeds,” IEEE Transaction on Industrial Applications., 33(6), 1997, 1444–1447.
4. Tapia, A., Tapia, G., Ostolaza, J. X., and J. R. Sáenz, “Modeling and control of a wind
turbine driven doubly fed induction generator,” IEEE Transaction. on Energy
Conversion, 18 (2), 2003, 194–204.
WIND ENGINEERING VOLUME 34, NO. 6, 2010 671
Table 2: Performance characteristics of the WRIG-based SWT system
Annual energy loss as a percentage of the
annual energy capture___________________________________________
Slip ring Slip ring Diode bridge Switch and
Annual energy electrical frictional rectifier external rotor Efficiency,
Region capture [Wh] [%] [%] [%] resistance [%] η [%]
BH 1331215 1.29 2.91 1.80 18.61 75.36
CW 645811 1.49 3.35 1.38 17.90 75.86
LB 1581484 1.39 3.13 1.58 18.26 75.61
MH 1266039 1.37 3.08 1.67 18.40 75.45
NA 383953 1.39 3.13 1.60 18.32 75.53
RA 1100044 1.30 2.91 1.78 18.62 75.36
SB 844841 1.40 3.14 1.58 18.25 75.62
SJ 1183171 1.45 3.27 1.46 18.12 75.68

23. 5. Cárdenas, R., and Pena, R., “Sensorless vector control of induction machines for
variable-speed wind energy applications,” IEEE Transaction. on Energy Conversion, 19
(1), 2000, 196–205.
6. Carrillo, C. , Feijóo, A. E., Cidrás, J., and González, J., “Power Fluctuations in an Isolated
Wind Plant,” IEEE Transaction on Energy Conversion, vol. 19 (1), 2004, 217–221.
7. Cárdenas, R., Pen~a, R., Asher, G., and Clare, J., “Power Smoothing in Wind Generation
Systems Using a Sensorless Vector Controlled Induction Machine Driving a Flywheel,”
IEEE Transaction on Energy Conversion, 19 (1), 2004, 206–216.
8. Kinjo, T., Senjyu, T., Urasaki, and Fujita. H, “Output levelling of renewable energy by
electric double-layer capacitor applied for energy storage system,” IEEE Transaction on
Energy Conversion, 21 (1), 2006, 221–227.
9. Senjyu, T., Sakamoto, R., Urasaki, N., Funabashi, T., Fujita, H., and Sekine, H., “Output
Power Leveling of Wind Turbine Generator for All Operating Regions by Pitch Angle
Control,” IEEE Transaction on Energy Conversion, 21(2), 2006, 467–475.
10. Voorspools, K.R., and D’ haeseleer, W.D., “Critical evaluation of methods for wind-power
Appraisal,” Renewable and Sustainable Energy Reviews, 11(1), 2007, 78–97.
11. Karki, R., Hu, P., and Billinton, R. “A Simplified Wind Power Generation Model for
Reliability Evaluation,” IEEE Transaction on Energy Conversion, 21 (2), 2006, 533–540.
12. Chowdhury, A.A., “Reliability Models for Large Wind Farms in Generation System
Planning,”Proceedings on Power Engineering Society General Meeting, 2, 2005, 1926– 1933.
13. Voorspools, K.R., and D’haeseleer, W.D., “An analytical formula for the capacity credit
of wind power,” Renewable Energy, 31 (1), 2006, 45–54.
14. Grauers, A., “Efficiency of three wind energy generator systems,” IEEE Transaction on
Energy Conversion, 11(3), 1996, 650–657.
15. Arifujjaman, Md., Iqbal, M.T., and Quaicoe, J.E., “Performance Comparison of Grid
Connected Small Wind Energy Conversion Systems,” Wind Engineering, 33(1), 2009,
1–18, 2009.
16. Ohyama, K., Arinaga, S., and Yamashita, Y. “Modeling and simulation of variable speed
wind generator system using boost converter of permanent magnet synchronous
generator, “Proceedings of the European Power Electronic and Applications
Conference, 2007, 1–9.
17. Orlando, N.A., Liserre, M., Monopoli, V.G., Mastromauro, R.A., and Dell’Aquila, A.
“Comparison of power converter topologies for permanent magnet small wind turbine
system,” Proceedings of the International Symposium on Industrial Electronics (ISIE),
2008, 2359–2364.
18. Hoffmann, R. and Mutschler, P., “The influence of control strategies on the energy
capture of wind turbines,” Proceedings of the Industry Applications Conference, 2000,
886–893.
19. Polinder, H. , Van der Pijl, F.F.A., De Vilder, G.J. , and Tavner, P.J. “Comparison of direct-
drive and geared generator concepts for wind turbines,” IEEE Transaction on Energy
Conversion, 21(3), 2006, 725–733.
20. A. Inoue, A., Hasan Ali, M., Takahashi, R., Murata, T., Tamura, J., Kimura, M. Futami, M.,
Ichinose, M., and Kazumasa, I., “A calculation method of the total efficiency of wind
generators,” Proceedings on International Conference on Power Electronics and Drives
Systems (PEDS), 2005, 1595–1600.
672 EXPERIMENTAL COMPARISON OF PERFORMANCES OF GRID CONNECTED
SMALL WIND ENERGY CONVERSION SYSTEMS