Journal of Electrical Engineering

www.jee.ro

OPTIMIZATION OF PMSG VARIABLE SPEED WIND ENERGY

CONVERSION SYSTEM CONTROLLER PARAMETERS BY
BIOGEOGRAPHY-BASED OPTIMIZATION
H. M. YASSIN H. H. HANAFY Mohab M. HALLOUDA
[email protected] [email protected] [email protected]
Electrical Power & Machines Department, Faculty of Engineering, Cairo University, Giza 12613, Egypt,

Abstract: This paper presents an optimization procedure Rs The stator resistance of PMSG
of maximum power point tracking method for direct-drive Rgf The resistance of the filter between the grid
permanent-magnet synchronous generator (PMSG) wind side converter and the grid.
generation system. The Field oriented control is used to Te The electromagnetic torque of PMSG
control the frequency converter using the PI controllers to Tm The mechanical torque of the turbine
achieve maximum power point tracking. The setting of the The wind speed (m/s)
parameters of the PI controllers used in the frequency The d- and q-axes stator voltages of PMSG
vds , vqs
converters is very difficult. . An optimization design
vd , vq The d- and q-axes components of the output
procedure for frequency converter using Biogeography-
voltages of the grid side converter
based optimization (BBO) technique is presented in this
Vdc The DC link voltage
paper.
The air density (Kg/m3)
The constrained optimization problem is solved using
Biogeography-based optimization and Genetic Algorithm. The blade pitch angle (deg.)
MATLAB-SIMULINK is used to evaluate the effectiveness The tip-speed ratio
of the proposed BBO technique. In order to verify the f The rotor magnetic flux of permanent magnet
validity and the performance of the proposed controllers, r The electrical angular speed
the simulation results for a VSWT-PMSG using a PI m The rotational speed
controllers optimized by the BBO technique is compared The angular frequency of the grid voltage
with that obtained using Genetic Algorithm. m The electrical angular position of the rotor
g The angular position of the grid voltage
Keywords: Biogeography-based optimization, Variable
speed wind turbine, MPPT, Field oriented control,
Permanent magnet synchronous generators. 1. Introduction
Recently, with the rising demand of electricity,
Nomenclature wind energy conversion system (WECS) has been
The blade swept area (m2) attractive and competitive with conventional fossil
The power conversion coefficient fuel energy resources that it is safe, pollution free,
Es The maximum phase voltage of the grid inexhaustible and free in term of its natural existence.
voltage WECS has become one of the most important
ids , iqs The d- and q-axes stator currents of PMSG renewable sources of alternative energy for the future
id , i q The d- and q-axes components of the output [1-2]. More than 35 GW of new wind power capacity
currents of the grid side converter was installed in 2013. The new global total at the end
J The total moment of inertia of the system of 2013 was 318 GW, representing cumulative
Lds , Lqs The d- and q-axes inductances of PMSG market growth of more than 12.3 %. By the year
Ld , Lq The d- and q-axes components of the 2018, the global wind power capacity is expected to
inductance of the filter be 600 GW. In Egypt, 550 MW of wind power were
Lgf The inductance of the filter between the grid installed. The total wind power installation is
side converter and the grid expected to be 7200 MW by 2020 [3].
Pg The active power output of PMSG There are two types of wind turbine systems
The extracted mechanical power from the wind which are fixed and variable speed wind turbines.
Qs The reactive power output to the grid Variable speed wind turbine (VSWT) is most
R The radius of the blade commonly used with the wind power systems in

1
Journal of Electrical Engineering
www.jee.ro

order to maximize energy captured at various wind In this paper, The Field oriented control is used to
speed. The variable-speed wind turbine systems have control the frequency converter using the PI
a wide speed range of operation and provide 10%- controllers to achieve MPPT of VSWT-PMSG. The
15% higher energy capture from the fixed speed setting of the parameters of the PI controllers, using
types [4] and reduce the load on the drive-train and trial and error method which is commonly used for
tower structure [5]. The wind power generators tuning PI controllers, used in the frequency
connected to the main power grid are generally of converters of VSWT-PMSG generation system is
two types: double fed asynchronous and direct-drive time consumed and the desired accuracy is satisfied
permanent magnet synchronous generators. The high only within a short interval close to the desired
efficiency, power density, wide speed range, operating point. So the objective of this paper is to
reliability and full isolation of the PMSG generator optimize the parameters of PI controllers of the
from the power grid make it preferred in variable machine and grid side converters using BBO, and
speed wind systems [68]. The control method to compare the results of this optimization technique
capture the maximum power from wind turbines in with the results obtained using GA optimization
the variable speed region is called a maximum power technique.
point tracking (MPPT) control. There are four The machine side converter (MSC) and grid side
categories of the MPPT methods in the wind turbine converter (GSC) controller parameters are
system; power signal feedback control, perturbation determined by BBO and GA to optimize the
and observation control, tip-speed ratio control and performance indices. The minimization of the
optimal torque control [9-10]. settling time, maximum overshoot and undershoot of
Due to the decoupling of the generator system the generator speed and the DC link voltage to track
from the grid, the grid support and fault ride through the reference values are considered as the
can be achieved easily [11, 12]. So PMSG wind performance indices to achieve MPPT.
turbine is becoming more favored by the wind power In order to verify the validity and the performance
industry. of the proposed controllers using BBO, the
The conventional PI controllers are very sensitive simulation results for a VSWT-PMSG using a PI
to parameter variations and the nonlinearity of controllers optimized by the BBO based optimization
dynamic systems. The setting of the parameters of technique is compared with that obtained using GA.
the PI controllers used in the frequency converters of The paper is divided as follows: in Section II, the
VSWT-PMSG generation system is exhausted. In modeling of the wind turbine, generator and the
[13], Genetic algorithm (GA) and response surface MPPT are described briefly. In Section III, control of
methodology (RSM) are used to optimally design the the frequency converter used in VSWT-PMSG is
controller used in the frequency converter of a presented. Section IV deals with the Biogeography-
VSWT-PMSG. Generalized reduced gradient (GRG) based optimization (BBO) technique. In Section V,
algorithm is compared with the proposed Gas-RSM the proposed optimization procedure is presented.
algorithm to verify the effectiveness of the designed Section VI presents the simulation results and
parameters using GAs-RSM. An extremum-seeking discussion. Finally, some conclusion remarks are
(ES) optimization algorithm is performed to extract given in section VII.
maximum power from a WECS in [14]. 2. Direct Drive PMSG Wind Turbine System
Modeling of natural process that used to solve The direct-drive PMSG wind generation system is
problems of general optimization is first used to shown in Fig. 1, where the wind turbine is connected
present the application of biogeography for to the PMSG directly.
optimization [15]. The biogeography-based The electrical power generated by the PMSG is
optimization algorithm is similar to other transmitted to a power grid via a variable-frequency
evolutionary algorithms optimization techniques such converter, which consists of a machine-side
as genetic algorithms (GAs), particle swarm converter and a grid-side converter connected back-
optimization (PSO). The fourteen benchmarks to-back via a DC link.
optimization problems are used to compare BBO
algorithm with other evolutionary algorithms 2.1 Modeling of Wind Turbine
optimization techniques to show significant The extracted mechanical power from the wind is
advantages of the BBO algorithm [15]. given by:

2
Journal of Electrical Engineering
www.jee.ro

Machine Side Converter Grid Side Converter

Pw

eA e
VA B
PMSG VB Vc Filter ec
Wind

Grid
Fig. 1: Direct-drive PMSG wind power generation system.

(1) The equation which lies at the base of the PMSG

model is the mechanical equation given as
The power conversion coefficient is a function of (8)
the tip-speed ratio (TSR) and the blade pitch angle
(deg.), in which the TSR is defined as [16]:
0.5 =0

(2)
Cp
max
Cp for given values of and for both fixed and 0.4
=5
variable speed wind turbines is calculated based on
the turbine characteristics in [17] by: 0.3

=10
Cp

(3)
0.2 =15

With =20

0.1
(4)

opt
The constants (C1-C6) are given in Appendix A. 0
0 2 4 6 8 10 12 14
The power conversion coefficient versus tip speed
ratio (Cp ) characteristics for different values of Fig. 2: power conversion coefficient versus tip speed ratio
the pitch angle is illustrated in Fig. 2. The (Cp ) curve for different pitch angle
maximum value of Cp (Cp_max) is achieved for
=0 and opt. The maximum turbine power is found at 1.6
14 m/s

a point of opt and Cp_max. Fig. 3 depicts the

turbine output power as a function of the rotor speed 1.4
13 m/s
(Pwm) for different wind speeds with the blade 1.2

pitch angle =0. The dotted line show the locus of 12 m/s
1
the maximum power point of the turbine which is
used to determine the reference of active power.
P (pu)

0.8 11 m/s

0.6 10 m/s
2.2 Modeling of PMSG
9 m/s
The voltage equations of PMSG are given in d-q 0.4
8 m/s
reference frame as [18-20]:
0.2

(5) 0
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6
Rotor speed (pu)
(6)
Fig. 3: Wind Turbine power characteristic (Pwm) curve
Then, the electromagnetic torque Te is expressed as with maximum power point tracking.
(7)

3
Journal of Electrical Engineering
www.jee.ro

2.3 Grid Model (15)

The input is a three phase voltage supply as
follows: (16)

3. Control of PMSG Wind Turbine

(9)
3.1 Machine Side Converter Control
The objective of MSC control is to control the
The equation in the stationary A B C frame is rotational speed of the PMSG to achieve variable-
speed operation with the MPPT control. Field
oriented control (FOC) is used in the control
methodology as shown in Fig. 5. This type of control
gives high performance. The control strategy requires
(10)
three controllers, two for the currents in the inner
loop and one for the rotor speed in the outer loop.
The torque and flux can be controlled separately by
the q and d-axis components of the generator current,
Where iA, iB and iC are the output currents and VA, respectively. Maximum Torque per Ampere control
VB and VC are the output voltages of the grid side strategy is used to control the d and q-axis currents.
converter. The transformation of three phase The d-axis component is set to zero, to minimize
stationary (ABC) to two phase stationary (-) is the stator current for a given torque, to produce
given by maximum torque per ampere, therefore minimize the
(11) resistive losses [21].
In Fig. 5, the d and q reference voltages and
(12) are given by
(17)
Transformation from the stationary () to the
synchronous (dq) frame, which is rotating with the (18)
angular frequency , can be obtained as
These compensation terms are added to improve
(13) the transient response [21]. They are obtained from
Equations (5) and (6).
(14)
(19)
For the angular position of the grid voltage (g)
which starts rotating with phase a as shown in Fig. 4, (20)
the amplitude of the phase voltages of the grid (eA, Space Vector Pulse Width Modulation (SVPWM)
eB, eC) which equals Es will be placed on the d- axis is the technique used to create the duty cycles of the
( ). The q-axis component voltage will be desired reference voltages.
zero ( ). The output voltages of the grid side
converter in the synchronous d q frame are given 3.2 Grid Side Converter Control
by The main objective of this controller is to regulate
q the DC bus voltage (Vdc) and control the reactive
power delivered to the grid. Fig. 6 shows the
complete scheme of the FOC algorithm applied to the
Vs d
grid side converter. The control strategy requires four
controllers, two for the currents in the inner loop, two
Es for the DC bus voltage and the reactive power in the
g outer loop. The d-axis component of the grid current
is controlled to maintain Vdc constant at a value that
grantee the grid line to line voltage. The reactive
Fig. 4: Space vector diagram of the grid three phase
power control is obtained through the q-axis
voltages.

4
Journal of Electrical Engineering
www.jee.ro

component of the grid current. The aim of the GSC (23)

control is to transfer all the active power produced by
the wind turbine to the grid and also to produce no (24)
reactive power so that unity power factor is obtained.
Grid-voltage phase-locked loop (PLL) is 4. Biogeography Based Optimization
employed to obtain the angular position of the grid At the beginning of the 19th century, Wallace
voltage (g) which is used to synchronize the output [23] and Darwin [24] studied the biogeography. BBO
voltage of GSC with the grid voltage. is an evolutionary algorithm which simulates the
In Fig. 6, the d and q reference voltages and biogeography of nearby islands. It is based on the
are given by mathematical study of the geographical distribution
of biological organisms. The number of species
(21) which are able to live in the island is called habitat
suitability index (HSI). Islands can share information
(22) with each others by emigration and immigration of
These compensation terms are added to improve different species.
the transient response [22]. They are given by
C

C om pensation1
V
dc
'
V ds *
* 0
ids V ds
+ PI ++ dq
- v a* b c

MSC
r Pg* iq* s V q' s V q*s SVPWM
MPPT + PI + PI ++
- -
Pg
id s C om pensation 2 ia s
abc ibs
iqs ics
dq

m m
1/s PMSG

Fig. 5: Machine Side Converter Control

Fig. 6: Grid Side Converter Control

5
Journal of Electrical Engineering
www.jee.ro

Extinction and migration of new species from one 5. Optimization Procedure

island to another is clearly described using The optimal design of the converters controllers
mathematical models of biogeography. of WECS is preferred to ensure the proper operation
The BBOs environment can be modeled as a of the wind generation system under a wide range of
group of islands where each island represents every wind speeds. Hence it is necessary to tune the
possible solution to the optimization problem. parameters of the machine-side and grid-side
Habitat suitability index (HSI) [25], which has a high converters.
similarity with the fitness of evolutionary algorithms As the MPPT of VSWT-PMSG is emphasized, the
(EAs) is high in those geographical areas which are design variables are the proportional and integral
suited as residences for biological species. gains of the generator speed controller (Kpw, Kiw) of
Habitability can be characterized by some the MSC, DC link voltage controller (Kpv, Kiv) and
variables which are called suitability index variables reactive power controller (Kpq, Kiq) of the GSC.
(SIVs).The problem dimension is obtained using the This section introduces the design procedure for
number of SIVs in each solution H. coordinated tuning of MSC and GSC controllers
An island with high HSI is corresponding to a using BBO optimization techniques.
good solution H of optimization problem while low This optimization requires a predefined of
HSI represents a poor one. different parameter of BBO such as migration rates,
Islands that have a large number of species is said mutation, and specifying of mutation rate, population
to have high HSI, while those islands that have a low size and number of generations, the details of the
number of species is said to have a low HSI. BBO design optimization method can be informally
Islands with a high HSI have many species that described as follows [15]:
emigrate to nearby islands, simply by virtue of the 1. Map the problem solutions to SIVs and habitats.
large number of species that they host. Islands with a Then, the BBO parameters can be initialized by
high HIS have low species immigration rate because initializing the maximum species count Smax, the
they are already nearly saturated with species. maximum mutation rate mmax, the maximum
Therefore high HSI Islands are more static in their migration rates E and I, and an elitism parameter.
species distribution than low HSI Islands. By the 2. Initialize a random set of islands, each island
same token, high HSI Islands have a high emigration corresponding to a potential solution to the given
rate; the large number of species on high HSI islands problem.
has many opportunities to emigrate to neighboring 3. For each island, map the HSI to the number of
habitats. Habitats with a low HIS have high species species, the immigration rate, and the emigration
immigration rate because of their sparse populations. rate.
This immigration of new species to low HSI habitats 4. Modify each non-elite island by using
may raise the HSI of the habitat, because the immigration and emigration, then recompute each
suitability of a habitat is proportional to its biological HSI.
diversity. However, if a habitats HSI remains low, 5. The probability of each islands species count is
then the species that reside there will tend to go updated. Then, according to its probability, mutate
extinct, which will further open the way for each non-elite island, and recompute each HSI.
additional immigration. Because of this low HSI 6. Go to step (3) for the next iteration. This loop can
habitats are more dynamic in their species be terminated after reaching an accepted solution or
distribution than high HSI habitats. after a predefined number of generations.
Improving the population is the way to solve Finally, the BBO Technique can be summarized
problems in heuristic algorithms. The BBO migration by the Flow chart shown in Fig. 7.
strategy is similar to the global recombination
approach of evolutionary strategies (ES) [26] in 6. Simulation Results
which many parents can contribute to a single In this section, simulation studies are carried out
offspring, but it differs in at least one important via MATLAB/SIMULINK for 1.5-MW direct-drive
aspect. In ES global recombination is used to create PMSG wind turbine. The PMSG parameters used in
new solutions, while in BBO migration is used to the simulation are given in the Appendix B.
change existing solutions. Mutation is performed for To determine the performance of optimum controller
the whole population in a manner similar to mutation parameters, the system is tested with the wind speed
in GAs (genetic algorithms). changes in steps. The applied wind speed was

6
Journal of Electrical Engineering
www.jee.ro

increased from 10 m/s to 12 m/s at 20 seconds and Table1 BBO and GA Characteristics
then it was decreased to 8 m/s at time = 40 second. Total population size 50
These sudden changes in wind speeds are selected to Number of generations 20
perform the effectiveness of the proposed optimized Mutation probability 0.005
controllers. The results of simulations are obtained Fitness function tolerance 1e-6
for reactive power Q = 0 and DC link voltage Vdc =
1500 V to grantee the grid line to line voltage. The parameters of the MSC and GSC controllers
In order to verify the validity and effectiveness of obtained using BBO and GA optimization techniques
the proposed controllers, the simulation results for a are given in Table 2. The proportional and integral
VSWT-PMSG using a PI controllers optimized by gains of the current controllers of the MSC and GSC
the BBO based optimization technique is compared have the same values for the BBO and GA technique.
to GA- optimized controllers. The comparison The maximum power point tracking, the difference
between BBO and GA is done for the same total between the reference and actual generator speed,
population size, number of generations, mutation and the difference between the dc link reference
probability and fitness function tolerance. Table 1 voltage and actual voltage were considered as the
shows the BBO and GA characteristics. objective functions. Maximum overshoot (MPOSr),
maximum peak undershoot (MPUSr), settling time
Describe the problem (Tssr) of the generator speed and the maximum peak
overshoot (MPOSVdc), maximum peak undershoot
(MPUSVdc) and settling time (TssVdc) of DC link
Generate initial solutions voltage are considered as constraints.
Fig. 8 shows the response of generator rotor speed
Set with controller parameters optimized using BBO and
Number of islands GA for step change in wind speed, respectively. It
Number of generations
Number of SIVs per islands can be seen from Fig. 9 that the generator speed
Mutation probability
response with BBO optimized controller parameters
has a smaller overshoot of 3.3 % when compared to
Yes
GA optimized controller parameters which has an
Obtain the best solution
Test No. of overshoot of 6 %. The generator speed response also
generations
settles faster with BBO optimized controller
parameters (4.1 s) than with GA optimized
No parameters (5.7 s).
Arrange the solutions in
ascending order Table 2 Optimal Values of Controller Parameters
Parameters Tuned using BBO Tuned using GA
Assign No. of species (Si) for Kpw 9.63 8.7
each solution based on its order
Kiw 7.19 7.6
Kpv 9.52 9.13
Assign (i, i) for each Kiv 7.81 6.89
solution based on (Si)
Kpq 1.63 1.61
Kiq 9.75 9.43
Yes
Test No. of Apply mutation inverse Fig. 9 and Fig. 10 show the response of DC link
islands proportionally with each sols (Si)
voltage with controller parameters optimized using
BBO and GA for step change in wind speed,
No respectively. It can be seen that The DC link voltage
Check each solution to be response also settles faster with BBO optimized
modified or not based on (i)
controller parameters (4.28 s) than with GA
optimized parameters (5.02 s). It can be noted from
Check each solution to be
modified or not based on (i)
these figures that the DC link voltage response with
BBO optimized controller parameters has a
maximum overshoot of 2.73 % when compared to
Fig. 7: BBO Technique flow chart sequence GA optimized controller parameters which has an

7
Journal of Electrical Engineering
www.jee.ro

1550
overshoot of 2.4 %. The controller performance Actual
indices with BBO and GA optimized parameters for Reference

DC Link Voltage (V)

step increase in wind speed are shown in Table 3. 1500
The less maximum overshoot and settling time
responses of the generator speed and DC link voltage
for BBO optimization technique improve the low 1450

voltage ride through capability of PMSG wind

generation system.
1400

Table3 Performance Indices for Step Change in Wind

Speed 1350
0 10 20 30 40 50 60
Performance indices For BBO For GA
Time (S)
MPOSr (%) 3.3 6
Fig. 10: Response of the DC link voltage for GA
MPUSr (%) 6.56 14.2
technique.
Tssr (s) 4.1 5.7
MPOSVdc (%) 2.73 2.4
The active power responses of the turbine and
MPUSVdc (%) 1.53 1.8
generator for the BBO and GA optimization
TssVdc (s) 4.28 5.02
techniques are shown in Fig. 11 and Fig. 12,
1.1
respectively. In Fig, 11 and 12, it can be noted that
the generator active power tracks the maximum
BBO extracted mechanical power from the turbine to
1
GA
achieve MPPT at the corresponding wind speeds. At
Rotor Speed (pu)

Reference
0.9 the time instant when wind speed steps up, the
electrical torque is first reduced for faster speed
0.8 regulation. As a result, the output active power drops
at the beginning of the generator speeding up period
0.7 to allow the maximum speed rising rate. As the speed
goes up, the output power rises accordingly and
0.6 settles to the steady-state value. On the contrary,
during wind speed step down, the electrical torque is
0.5
0 10 20 30 40 50 60 first increased, helping slow down the generator. The
Time (S) increase of electrical torque leads to higher generator
Fig. 8: Response of the rotor speed of PMSG output power, at the very beginning of the speed
change. As the speed slows down, the active power
1550
decreases as well and settles to the steady-state value.
Actual
Turbine & Generator Power (pu)

Reference Generator
DC Link Voltage (V)

Turbine
1500 1

0.8

1450
0.6

0.4
1400

0.2

1350 0
0 10 20 30 40 50 60 0 10 20 30 40 50 60
Time (S) Time (S)
Fig. 9: Response of the DC link voltage for BBO Fig. 11: Response of the generator and grid active power
technique. for BBO technique.

8
Journal of Electrical Engineering
www.jee.ro

The reactive power responses of the grid for the Fig. 15 and Fig. 16 show the response of the d-q
BBO and GA optimization techniques are shown in current components of the PMSG with controller
Fig. 13 and Fig. 14, respectively. In Fig, 13 and 14, parameters optimized using BBO and GA,
the reactive power is maintained at zero at the grid respectively. The d-axis current is zero during all the
side for either steady state or transient operation, and simulation time to get the maximum torque/ampere.
unity power factor is kept regardless of the amount of The q-axis current is directly tracking the
real power. electromagnetic torque. Therefore, it is decreasing
when the rotor speed has to increase in order to
Turbine & Generator Power (pu)

Generator
Turbine
obtain a bigger acceleration. The current goes to the
1
value which makes the electromagnetic torque equal
0.8
to the wind turbine torque when the rotor speed
reference is met.
0.6 Fig. 17 and Fig. 18 show the response of the d-q
current components of the grid with controller
0.4 parameters optimized using BBO and GA,
respectively. The grid q-axis current is set all over
0.2
the simulation to zero to get zero reactive power
0
(unity power factor). The d-axis component is
0 10 20 30 40 50 60
directly tracking DC link voltage.
Time (S)
Fig. 12: Response of the generator and grid active power
Generator q-axis Current (pu)

GA technique 1.2 Actual

0.1 Reference
Grid Reactive Power (pu)

0.08 1

0.06
0.8
0.04

0.02
0.6
0

-0.02 0.4

-0.04
0.2
-0.06

-0.08 0
0 10 20 30 40 50 60
-0.1
0 10 20 30 40 50 60
Time (S)
Time (S) (a)
0.1
Fig. 13: Response of the grid reactive power for BBO
Generator d-axis Current (pu)

Actual
technique 0.08
Reference
0.1
0.06
Grid Reactive Power (pu)

0.08
0.04
0.06
0.02
0.04
0
0.02
-0.02
0
-0.04
-0.02
-0.06
-0.04
-0.08
-0.06
-0.1
-0.08 0 10 20 30 40 50 60

-0.1
Time (S)
0 10 20 30 40 50 60
(b)
Time (S)
Fig. 15: Response of the d-q current components of the
Fig. 14: Response of the grid reactive power for GA
PMSG for BBO technique
technique.

9
Journal of Electrical Engineering
www.jee.ro

1
Generator q-axis Current (pu)

Actual Actual
0.9
Reference

Grid d-axis Current (pu)

Reference
1
0.8

0.8 0.7

0.6
0.6
0.5

0.4 0.4

0.3
0.2
0.2

0 0.1
0 10 20 30 40 50 60 0 10 20 30 40 50 60
Time (S) Time (S)
(a) (a)
0.1 0.5
Generator d-axis Current (pu)

0.08
Actual Actual
0.4
Reference Reference

Grid q-axis Current (pu)

0.06 0.3
0.04
0.2
0.02
0.1
0
0
-0.02
-0.1
-0.04

-0.06
-0.2

-0.08 -0.3

-0.1 -0.4
0 10 20 30 40 50 60
Time (S) -0.5
0 10 20 30 40 50 60
(b) Time (S)
Fig. 16: Response of the d-q current components of the (b)
PMSG for GA technique Fig. 18: Response of the d-q current components of the
1
grid for GA technique
0.9 Actual
Grid d-axis Current (pu)

0.8
Reference 7. Conclusion
0.7

0.6 This paper has attempted to present an approach

0.5 to optimally determine the controller parameters
0.4 which control the frequency converter used in
0.3 VSWT-PMSG to achieve MMPT. The Field oriented
0.2 control is used to control the frequency converter
0.1
using the PI controllers to obtain MPPT. Due to the
0
0 10 20 30 40 50 60 nonlinearity of the wind generation system, the
Time (S)
setting of the converters PI controllers parameters is
(a)
0.5
difficult.
0.4 Actual The maximum power point tracking, the
Grid q-axis Current (pu)

Reference difference between the reference and actual generator

0.3

0.2 speed, and the difference between the reference and

0.1 actual dc link voltage were considered as the
0 objective functions. The minimization of the settling
-0.1 time, maximum overshoot and undershoot of the
-0.2 generator speed and the DC link voltage to track the
-0.3 reference values are considered as the performance
-0.4 indices to achieve MPPT.
-0.5
0 10 20 30 40 50 60 The constrained optimization problem is solved
Time (S) using BBO. The effectiveness of the designed
(b) parameters using BBO is then compared with that
Fig. 17: Response of the d-q current components of the obtained using genetic algorithm (GA).
grid for BBO technique

10
Journal of Electrical Engineering
www.jee.ro

The validity of the proposed methodology is pairs of PMSG; and r is the amplitude of the flux
confirmed through the simulation results. The induced by the permanent magnets of the rotor.
effectiveness of the proposed methodology is verified
for step change in wind speed. The overshoots, Base power Pb (MVA) 1.5
undershoots and settling time of DC link voltage Base voltage Vb (V)
response and generator speed response have reduced Base frequency fb (Hz) 11.5
with optimal parameters obtained using BBO and the Pole pairs of PMSG p 40
proposed BBO technique achieves the MPPT.
Nominal WT mechanical power (pu) 1.1
The detail of the optimum design procedure is
discussed, which can be applied to other Nominal WT speed (pu) 1.2
inverter/converter topology used widely in variable WT inertia constant (pu) 4.8
speed wind energy conversion systems. PMSG inertia constant (pu) 0.5
Finally, it is concluded that BBO optimization Shaft stiffness (pu) 2
technology might be a good choice for optimum Rated generator torque (pu) 1
design of controller parameters. Rated generator power (pu) 1
Rated generator line voltage 1
APPENDIX A
Wind Turbine Parameters Rated generator speed (pu) 1
Parameter Value Generator inductance in the d frame (pu) 0.7
C1 0.5176 Generator inductance in the q frame (pu) 0.7
C2 116 Generator stator resistance (pu) 0.01
C3 0.4
Flux of the permanent magnets (pu) 0.9
C4 5
C5 21 DC-link capacitance (pu) 1
C6 0.0068 Line inductance (pu) 0.1
0 Rate wind speed (m/s) 12

APPENDIX B
References
Parameters of the PMSG-Based WECS
1. Abdullah MA, Yatim AHM, Tan CW.: A study of
Parameters of the WECS in the simulations are maximum power point tracking algorithms for wind
converted to a pu system and the real values can be energy system. In: Proceedings of the 1st IEEE
derived by multiplying each pu value and the base conference on clean energy and technology (CET),
value. The system base values are defined as follows 2011, pp. 321326.
[27]: 2. Saidur R, Islam MR, Rahim NA, Solangi KH.: A
review on global wind energy policy. In: Renewable
and Sustainable Energy Reviews, vol. 14, 2010, pp.
17441762.
3. Global Wind Report Annual Market Update 2013.
Global Wind Energy Council (GWEC) [Online].
Available: http://www.gwec.net.
4. Q. Wang and L. Chang: An intelligent maximum
power extraction algorithm for inverter-based
variable speed wind turbine systems. In: IEEE
Transactions Power Electron., vol. 19, Sept. 2004, pp.
1242-1249.
Where the variables with subscript b represent the 5. Y. Zhao, C. Wei, Z. Zhang, and W. Qiao: A review on
base values; P, V, I, f are the power, voltage, current, position/speed sensorless control for permanent-
and electrical frequency, respectively; Z,L,C are the magnet synchronous machine-based wind energy
impedance, inductance, and capacitance, conversion systems. In: IEEE J. Emerging and
respectively; , m are the electrical and mechanical Selected Topics in Power Electron., vol. 1, no. 4, Dec.
2013, pp. 203-216.
angular frequencies, respectively; T is the torque; J,K
6. Molina MG, dos Santos EC, Pacas M.: Advanced
are the inertia and stiffness, respectively; p is the pole power conditioning system for grid integration of

11
Journal of Electrical Engineering
www.jee.ro

direct-driven PMSG wind turbines. In: IEEE Energy 16. Y. Xia, K. H. Ahmed, and B. W. Williams: Wind
Conversion Congress and Exposition (ECCE), 2010, turbine power coefficient analysis of a new maximum
pp. 33663373. power point tracking technique. In: IEEE
7. Li S, Haskew TA, Xu L.: Conventional and novel Transactions on Industrial Electronics, vol. 60, March
control designs for direct driven PMSG wind turbines. 2013, pp. 1122-1132.
In: Electric Power Systems Research, vol. 80, 2010, 17. Siegfried Heier.: Grid Integration of Wind Energy
pp. 328338. Conversion Systems. John Wiley & Sons Ltd, ISBN 0-
8. Muyeen SM, Takahashi R, Murata T, Tamura J.: A 471-97143-X, 1998.
variable speed wind turbine control strategy to meet 18. T. H. Nguyen, S.-H. Jang, H.-G. Park, and D.-C. Lee.:
wind farm grid code requirements. In: IEEE Sensorless control of PM synchronous generators for
Transactions on Power Systems, vol. 25, 2010, pp. micro wind turbines. In: Proceedings of IEEE 2nd Int.
331340. Power Energy Conf., December 2008, pp. 936941.
9. G. Hua and Y. Geng.: A novel control strategy of 19. J. Dai, D. Xu, B.Wu, and N. R. Zargari.: Unified DC-
MPPT tracking dynamics of wind turbine into link current control for low-voltage ride-through in
account. In: Proc. IEEE Power Electron. Spec. Conf., current-source-converter-based wind energy
Jun. 2006, pp. 1-6. conversion systems. In: IEEE Trans. Power Electron.,
vol. 26, Jan. 2011, pp. 288297.
10. K.-H. Kim, D.-C. Lee, and J.-M. Kim.: Fast tracking
20. G. Foo and M. F. Rahman.: Sensorless direct torque
control for maximum output power in wind turbine
and flux-controlled IPM synchronous motor drive at
systems. In: Proceedings of AUPEC, December 2010,
very low speed without signal injection. In: IEEE
pp. 1-5.
Transactions Power Electron., vol. 57, Jan. 2010, pp.
11. H. Polinder, S.W. H. de Haan, M. R. Dubois, and J. 395403.
Slootweg.: Basic operation principles and electrical 21. Busca, C., et al., et al.: Vector control of PMSG for
conversion systems of wind turbines. In: Proceedings wind turbine applications. In: s.l.: Dept. of Energy
of the Nordic Workshop Power Ind. Electron. , June Technol., Aalborg Univ., Industrial Electronics
2004, pp. 1416, Norway. (ISIE), Aalborg, Denmark, 2010, pp. 3871 - 3876.
12. G. Michalke, A. D. Hansen, and T. Hartkopf.: Control 22. YANG Yong, RUAN Yi, SHEN Huan-qing, and
strategy of a variable speed wind turbine with TANG Yan-yan.: Grid-connected inverter for wind
multipole permanent magnet synchronous generator. power generation system. In: J Shanghai Univ, 2009,
In: Proceedings of the Eur.Wind Energy Conf. Exhib. pp. 5156.
, May 2007, pp. 710, Milan, Italy. 23. A. Wallace.: The Geographical Distribution of
13. Hasanien H. M, Muyeen S. M.: Design Optimization Animals. (Two Volumes). Boston, MA: Adamant
of Controller Parameters used in Variable Speed Media Corporation, 2005.
Wind Energy Conversion System by Genetic 24. C. Darwin.: The Origin of Species. New York:
Algorithms. In: IEEE Transactions on Sustainable Gramercy, 1995.
Energy, vol.3, April 2012, pp.200- 208. 25. T. Wesche, G. Goertler, and W. Hubert.: Modified
14. A. Ghaffari, M. Krsti, and S. Seshagiri.: Power habitat suitability index model for brown trout in
optimization and control in wind energy conversion southeastern Wyoming. In: North Amer. J. Fisheries
systems using extremum seeking. In: Proceedings of Manage., vol. 7, 1987, pp. 232237.
the American Control Conference (ACC), 2013, pp. 26. T. Back.: Evolutionary Algorithms in Theory and
2241-2246. Practice. In: Oxford University Press, 1996.
15. D. Simon. Biogeography-based optimization. In: 27. Geng, H.; Yang, G.; Xu, D.; Wu, B.: Unified power
IEEE Transactions on Evolutionary Computation, vol. control for PMSG-based WECS operating under
12, December 2008, pp. 702-713. different grid conditions. In: IEEE Transactions
Energy Convers., vol. 26, 2011, pp. 822830.

12