Renewable Energy 66 (2014) 371e380

Contents lists available at ScienceDirect

Renewable Energy
journal homepage: www.elsevier.com/locate/renene

Control of a grid-connected direct-drive wind energy conversion

system
Jianhu Yan a, Heyun Lin a, *, Yi Feng a, Z.Q. Zhu b
a
Servo Control Engineering Center of Education Ministry, Southeast University, Nanjing 210096, China
b
Department of Electronic and Electrical Engineering, University of Sheffield, Mappin St., Sheffield S1 3JD, UK

a r t i c l e i n f o a b s t r a c t

Article history: This paper investigates the current control for a grid-connected direct-drive wind energy conversion
Received 10 July 2013 system (DDWECS) with a permanent magnet synchronous generator (PMSG), which utilizes a back-to-
Accepted 27 December 2013 back pulse width modulation (PWM) converter. For the machine-side, the controller adopts a current
Available online 21 January 2014
vector control method based on the rotating reference frame (RRF) and the maximum power extraction
(MPE) is realised through the tip speed ratio (TSR) method. For the grid-side, a novel controller is
Keywords:
proposed for the first time to be successfully used for the DDWECS, which combining a proportional
Current vector control
complex integral (PCI) current inner loop based on stationary reference frame (SRF) for regulating the
Grid-connected direct-drive wind energy
conversion system
grid-side current with a dc voltage outer loop for stabilizing the dc bus voltage and compare with the
Proportional complex integral control proportional resonant (PR) controller. A system simulation model is established by using the Matlab/
Rotating reference frame Simulink to simulate the performance of the DDWECS and a prototype system has been build and tested
Stationary reference frame to verify the validity of the developed control methods for both machine-side and grid-side and the
excellent performance of the DDWECS.
Ó 2014 Elsevier Ltd. All rights reserved.

1. Introduction
With the rapid development of the renewable energy source,
especially in the field of wind power, the variable speed operation is
widely employed in the wind energy conversion system (WECS).
Moreover, the gearless direct-drive permanent magnet synchro-
nous generator (PMSG) is considered to be promising due to its
higher reliability, efficiency and lower acoustic noise [1e3]. Direct-
drive WECS (DDWECS) generally employs a fully controllable
voltage source converter (VSC) with a back-to-back topology to
realise the maximum power extraction (MPE) of wind turbine and
regulate the grid interface for satisfying the grid code [4e7]. The
DDWECS with full VSC-based MOSFET or IGBT converter is
becoming the dominant in low and high power applications,
respectively [8].
From the control aspect, the VSC can be divided into a machine-
side converter and a grid-side converter [9e12]. The machine-side
converter mainly adopts a current vector control with a current
inner loop and a speed outer loop to implement the variable speed
operation for capturing the maximum wind power through the tip
* Corresponding author. Tel.: þ86 2583794169.
E-mail address: hyling@seu.edu.cn (H. Lin).
speed ratio (TSR) method [13]. The grid-side converter can either
apply the voltage-oriented control based on the d
q rotating
reference frame (RRF) or use the proportional-resonant (PR) or
recently proposed proportional complex integral (PCI) current in-
ner control with a dc voltage outer loop control based on the a
b
stationary reference frame (SRF) to satisfy the grid code [14e19].
The advantage of the control method based on SRF is that the phase
lock loop (PLL) can be eliminated in comparison with that based on
RRF. However, the bandwidth of the closed-loop system is larger
than the grid frequency and the loop gain needs to be increased
under other desired frequencies [20,21]. Fortunately, the grid fre-
quency is generally constant so that the disadvantages of the con-
trol method based on the SRF can be neglected to some extent.
Moreover, the PR control method based on the internal mode
principle can offer the zero steady-state tracking error for sinu-
soidal signal and the PCI control method as a simplification and
improvement of the PR control can improve the dynamic response
in the start-up and has been successfully applied in the balanced
system [17,18]. In DDWECS, the machine-side converter generally
employs the flux-oriented current vector control method to regu-
late the speed of PMSG and realise the MPE according to the TSR
method. In addition to considering the cost, PR or PCI control is
available for gird-side converter without additional hardware or
software PLL. Reference [17] points out that the PCI control is more

0960-1481/$ e see front matter Ó 2014 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.renene.2013.12.037
372 J. Yan et al. / Renewable Energy 66 (2014) 371e380

suitable for the balanced grid-side converter due to its simplifica- ue ¼ pu (2)
tion in the control structure. However, successful applications of
the PCI control in DDWECS have not yet been reported. where p is the pole pair number and u is the mechanical speed of
This paper studies the flux-oriented current vector control for PMSG.
the machine-side and the PCI control based on the SRF for the grid- When the d-axis is aligned along the rotor flux position, the deq
side of DDWECS. For the machine-side, the reference speed is ob- stator flux linkage equation is
tained by the TSR method and the speed is regulated to realise the




MPE of the wind turbine. For the grid-side, the PCI inner loop jsd Lsd isd jf
¼ þ (3)
combined with a dc voltage outer loop is firstly proposed to be used jsq Lsq isq 0
for the DDWECS to stabilise the dc voltage and to regulate the grid
current. The two controllers are designed and analysed in details. where Lsd and Lsq are the d and q components of the stator
Compared with the PR controller in grid-side, the PCI controller can inductance respectively; jf is the flux linkage produced by PM. The
improve the dynamic response effectively. A system simulation electromagnetic torque is
model is established based on the Matlab/Simulink to examine the h   i
performance of the used control methods and an experimental Te ¼ 1:5p jf isq þ Lsd
Lsq isd isq (4)
system is carried out to verify the feasibility and the excellent
performance of the developed DDWECS. If Lsd is equal to Lsq , represented by Ls, as in the surface mounted
PMSG, the electromagnetic torque of the PMSG is simplified as

Te ¼ 1:5pjf isq (5)

2. Configuration of DDWECS
Besides, the mechanical equation of the PMSG is
The DDWECS mainly consists of a wind turbine, a PMSG, a back-
to-back converter, a filter induction and an isolated transformer, as du
J ¼ Ti
Te (6)
shown in Fig. 1. The PMSG connects to the grid through a back-to- dt
back converter, converting the mechanical energy produced by the
wind turbine to the electrical energy [22,23]. The back-to-back where, J is the inertia of the PMSG; Ti is the input torque.
converter contains a machine-side converter and a grid-side con-
verter, which are connected by a dc bus. 3.2. Current vector control
The control system of the DDWECS is divided into a machine-
side control sub-system and a grid-side control sub-system. The The current vector control of the machine-side converter con-
former can utilise the flux-oriented d
q current vector control tains a faster current inner loop and a slower speed outer loop. The
aiming to achieve the MPE from the wind turbine by a speed d-axial current is only affected by vsq. Meanwhile, the q-axial cur-
control loop, which offers the speed reference to the machine-side rent is affected by vsd [1]. Therefore, the current inner loop is
converter by the TSR method. The latter employs the PCI control decoupled, as shown in Fig. 2. The current inner loops of d and q
based on the SRF to maintain the dc voltage constant and to adjust axis jointly adjust the wind turbine speed and regulate the d-axial
the power factor. It should be noted that for high power WECS the current to zero to enhance the efficiency of the PMSG.
variable pitch wind turbine is adopted to prevent the turbine from The decoupled control strategy of d and q current loops is ob-
operating over the rated power at relatively high wind speed tained by rewriting Eq. (1) and Eq. (3) as
[24,25].


di
vsd ¼
Rs isd þ Ls sd
ue Lsq isq (7)
dt
3. Control of machine-side converter


disq
3.1. Mathematic model of PMSG vsq ¼
Rs isq þ Ls þ ue Ls isd þ ue jf (8)
dt
The stator voltage mathematic model of the PMSG in the deq
where the items in the bracket of Eq. (7) and Eq. (8) are the state
RRF is expressed as [26,27],
equations between the voltage and current in the d and q loops




 respectively, while the others are the compensation terms.
vsd isd d jsd
jsq
¼
Rs
þ ue (1)
vsq isq dt jsq jsd
3.3. MPE by TSR method
where Rs is the stator resistance; vsd , vsq , isd , isq , jsd and jsq are the
d and q components of stator voltage, current and flux linkage The output power of the wind turbine can be calculated as [28]
respectively; ue is the electrical angular speed

Fig. 1. Configuration of DDWCES. Fig. 2. Vector control structure for machine-side converter.
 J. Yan et al. / Renewable Energy 66 (2014) 371e380 373

1
Pw ¼ prCp ðl; bÞR2 v3 (9)
2

where r is the air density; R is the wind turbine blade radius; v is the
wind speed; l is the TSR; b is the pitch angle; Cp is the power co-
efficient, which is a non-linear function of l and b. l is defined as

um R
l ¼ (10)
v

where um is the rotor angular speed. When b is zero, l approaches

the maximum value. Fig. 3(a) shows the Cp
l curve when b is Fig. 4. Block diagram of current inner loop of machine-side controller in d and q axis.
zero. The maximum power coefficient Cp max and the optimal TSR
lopt can be obtained directly from the curve. According to Eq. (9)
and Eq. (10), the maximum output power of wind turbine can Due to adopting the PI control, Kd ðsÞ can be written as
be expressed as
ki
Kd ðsÞ ¼ kp þ (13)
Pw max ¼ 0:5rpR5 Cp max u3m =v3 (11) s
Fig. 3(b) shows the output power of wind turbine as a function where kp is the proportional gain and ki is the integral gain. Further
of rotor angular speed at different wind speeds. It can be seen that the open-loop transfer function without disturbance is
the maximum power only exists at a special rotor speed under a
certain wind speed. The TSR method can be employed to calculate
 
kPWM kp s þ ki
the reference rotor speed to realise the MPE by detecting the wind Gi ðsÞ ¼ (14)
sðTs s þ 1ÞðLs s þ Rs Þ
speed [29,30].
where kPWM is pffiffiffithe pulse width modulation (PWM) gain, which is
3.4. Design of machine-side controller kPWM ¼ Vdc = 3 for space vector PWM (SVPWM), and Ts is the
sampling period.
The block diagram of the d and q inner current loops of the In Eq. (14), the pole point of the open-loop transfer function is
machine-side controller is shown in Fig. 4, which has the identical
Rs =Ls . To obtain the optimal tracking performance of the current
type and also has the identical corresponding compensator. inner loop, the zero point of the PI compensator needs to
Considering the d-axis loop, the compensator Kd ðsÞ generally compensate the pole point of the open-loop transfer function,
adopts the proportional-integral (PI) type, because the reference namely ðki =kp Þ ¼ ðRs =Ls Þ. The open-loop transfer function is thus
commands of isd and isq are both dc signals. Besides esd and esq as rewritten as
the disturbances are the d and q axial induced voltages of PMSG
respectively. kp kPWM
Gi ðsÞ ¼ (15)
sLs ðTs s þ 1Þ
The optimal damping ratio of second-order system is 0.707 [30],
so kp and ki can be respectively calculated by

0:5Ls
kp ¼ (16)
Ts kPWM

0:5Rs
ki ¼ (17)
Ts kPWM
Besides, the close-loop transfer function is

Fig. 3. Characteristics of wind turbine. (a)Power coefficient versus TSR when b is zero.
(b) Output power versus rotor speed at different wind speeds.
kp kPWM
Fi ðsÞ ¼ (18)
Ts Ls s2 þ Ls s þ kp kPWM
According to the parameters of the studied DDWECS, as shown
Table 1 in Appendix, kp and ki are equal to 4 and 978 respectively.
The bode diagrams of the open-loop and close-loop systems are
shown in Fig. 5, which illustrates that the phase margin is
approximately 70 and the bandwidth frequency of the close-loop
system is about 560 Hz. Apparently, The machine-side inner loop
system has good stability and adequate bandwidth to enhance the
tracking capability.
The block diagram of the speed outer loop used to regulate the
speed of PMSG to realise the MPE is given in Fig. 6, in which the
friction of the DDWECS is not considered. Jtotal is the total equiva-
lent inertia of the DDWECS, Tturbine is the input wind turbine
torque.
Ku ðsÞ is the outer loop compensator, which is
374 J. Yan et al. / Renewable Energy 66 (2014) 371e380

k
Ku ðsÞ ¼ kup þ ui (19)
s

where, kup is the proportional gain and kui is the integral gain. As Ts
is relatively small, Ts Ls s2 in Eq. (18) can be neglected. In addition,
the speed sampling period Tus is larger than the current sampling
period so that the open-loop transfer function without disturbance
can be rewritten as
 
1:5pjf kup s þ kui
Gu ðsÞ ¼ (20)
Jtotal s2 ðTus s þ 1Þ
The PI compensator of the speed outer loop can be designed
according to the type-2 system [30] as

1:5pjf kui hþ1

¼ (21)
Jtotal 2h2 Tu2s

where h is the intermediate frequency wideband

kup =kui
h ¼ (22)
Tus
Generally, h is equal to 5 [31]. kup and kui can thus be calculated
by

Jtotal
kui ¼ (23)
12:5Tu2 s pjf

kup ¼ 5Tus kui (24)

kup and kui are equal to 0.36 and 3.57 respectively in the studied
DDWECS. The bode diagrams of the open-loop and close-loop
systems are shown in Fig. 7, which illustrate that the phase
margin is approximately 45 and the bandwidth frequency of the
close-loop system is about 8 Hz. It can be seen that the machine-
side outer loop system has good stability and narrow bandwidth
to decrease the disturbance.

4. Control of grid-side converter

4.1. Model of grid-side converter based on SRF

The schematic of the grid-side converter is shown in Fig. 8,

where the three-phase voltage source is described as the voltage at
the point of common coupling (PCC). In the SRF, the voltage
equation is [32,33]
       
vga d i ga vga v
¼ Lg þ Rg þ ga 1 (25)
vgb dt igb vgb vgb1

Fig. 5. Bode diagrams of the open-loop and close-loop systems of machine-side cur-
rent loop. (a) Bode diagram of open-loop system. (b) Bode diagram of close-loop
system.
where iga , igb , vga , vgb , vga1 and vgb1 are the a and b axial components
of grid current, the converter output voltage and the PCC voltage
respectively; Lg and Rg are the grid-side inductance and resistance
respectively.
The real and reactive power delivered to the PCC is:

P ¼ 1:5 vga1 iga þ vgb1 igb (26)

Q ¼ 1:5
vga1 igb þ vgb1 iga (27)

Thus, iga and igb can be controlled through the regulation of the
Fig. 6. Block diagram of speed outer loop of machine-side controller. real and reactive power. The reference command of iga and igb is
 J. Yan et al. / Renewable Energy 66 (2014) 371e380 375

Fig. 8. Schematic of grid-side converter.

2
Cdc dVdc
Pin ¼ þ Pout (30)
2 dt

where Cdc is the dc capacity; Vdc is the dc voltage; Pin is the input
power and Pout is the output power, which is

2Lg dP 2Lg dQ
Pout ¼ P þ P þ Q (31)
3v2ga1 dt 3v2ga1 dt

where vga1 is the PCC phase voltage. When the reactive power is
zero to operate under unit power factor, Eq. (30) can be rewritten as

!
2
dVdc 2 2 2Lg dP
¼ P
Pþ P (32)
dt Cdc in Cdc 3v2ga1 dt

4.2. Design of PCI inner loop controller

The current inner loop block diagram in the grid-side is given in

Fig. 9, where KI ðsÞ is the compensator; k is the PWM gain; Rg and Lg
are the grid-side resistance and inductance respectively.
The close-loop transfer function without disturbance can be
expressed as

kKI ðsÞ
FI ðsÞ ¼ (33)
Lg s þ Rg þ kKI ðsÞ
The error of the current signal is

Lg s þ Rg
EI ðsÞ ¼ (34)
Lg s þ Rg þ kKI ðsÞ
It is clear that from (34) the error will approach to zero when it
satisfies the condition:

jKI ðsÞjs¼jug ¼ N (35)

where ug is the angular frequency of the grid-side current and

Fig. 7. Bode diagrams of the open and close-loop systems of machine-side speed loop. voltage. Hence, a PCI compensator with the infinite gain is adopted,
(a) Bode diagram of open-loop system. (b) Bode diagram of close-loop system.
which is expressed as [17]
!
2 vga1 vgb1
iga ref ¼ Pþ 2 Q (28)
3 v2ga1 þ v2 vga1 þ vgb1
2
gb1

!
2 vgb1 vga1
igb ref ¼ P
2 Q (29)
3 v2ga1 þ v2 vga1 þ vgb1
2
gb1

The power equation without power loss consideration is Fig. 9. Block diagram of current inner loop of grid-side controller in a axis.
376 J. Yan et al. / Renewable Energy 66 (2014) 371e380

Fig. 11. Block diagram of dc voltage outer loop controller.

Therefore, the compensator can also approach the zero steady-state

error when the angular frequency is ug .
In order to obtain a fast response, the bandwidth frequency in
the studied DDWECS is chosen as 700 Hz. The expression of
bandwidth is

20lgjFI ðsÞjs¼jub ¼
3 (37)

where ub is the bandwidth frequency. Thus, kIp can be calculated by

assuming that only the proportional term is available
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
 2 ffi
Rg þ Lg ub þ 2R2g
kIp ¼ (38)
k
Further kIi can be obtained when ub properly increases to
700 Hz. kIp and kIi are 10 and 974 respectively in the studied
DDWECS. The bode diagrams of the open-loop and close-loop
systems are shown in Fig. 10, which illustrates that the phase
margin is approximately 90 and the bandwidth frequency of the
close-loop system is about 724 Hz. Apparently, the grid-side cur-
rent inner loop system has good stability and broad bandwidth to
increase the response time.

4.3. Design of dc voltage outer loop controller

According to Eq. (32) without the disturbance of input power

2 is
Pin , the transfer function from P to Vdc

Fig. 10. Bode diagrams of the open and close-loop systems of grid-side current loop.
(a) Bode diagram of open-loop system. (b) Bode diagram of close-loop system.

kIi
KI ðsÞ ¼ kIp þ (36)
s
jug

where kIp is the proportional gain; kIi is the integral gain. The PCI
compensator is similar to the traditional PI compensator in form. Fig. 12. Bode diagrams of the open-loop systems of grid-side dc voltage loop.
 J. Yan et al. / Renewable Energy 66 (2014) 371e380 377

Fig. 13. Machine-side and grid-side control results of DDWECS. (a) Speed of PMSG. (b) Power coefficient of wind turbine. (c) q-axial current of PMSG. (d) DC voltage. (e) Grid current
and voltage when wind speed up. (f) Grid current and voltage when wind speed down.

2 kv s þ 1
KV ðsÞ ¼

Cdc s
where kv ¼ ð2Lg P0 =3v2ga1 Þ is a constant number.
The block diagram of the dc voltage outer loop is given in Fig. 11,
which contains a dc voltage compensator HV ðsÞ, a power controller
and a control plant KV ðsÞ. From Eq. (39), HV ðsÞ needs to be multi-
plied by
1 to compensate the negative sign in Eq. (39). As a result,
the open loop transfer function can be expressed as
GV ðsÞ ¼
HV ðsÞGP ðsÞKV ðsÞ
In order to eliminate the phase delay of the PCI current inner
loop, the cutoff frequency uc is taken as a very low value. At this
time, jGP ðsÞjs¼juc z1. Equation Eq. (40) combining Eq. (39) can thus
be rewritten as
kv s þ 1
GV ðsÞ ¼ HV ðsÞ
s side operation under unity power factor with a stable dc voltage by
Compensator HV ðsÞ needs to satisfy three conditions: (i) GV ðsÞ is
at uc with the slope of
20 dB/dec; (ii) the larger attenuation slope
in high-frequency region is required to suppress the high frequency
noise; (iii) the phase margin of GV ðsÞ is adequately large [34]. So,
(39) HV ðsÞ adopts a second-order regulator, as
Cdc kv1 s þ kv2
HV ðsÞ ¼ (42)
2 sðs þ kv3 Þ
where kv1 , kv2 and kv3 are taken as 200, 174 and 20, respectively.
The bode diagrams of the open-loop systems are shown in Fig. 12,
which illustrates that the grid-side voltage outer loop system has
good stability because the phase margin is approximately 62 and
the cutoff frequency of the open-loop system is about 9.2 Hz less
(40) than the bandwidth of the power control loop.
5. Simulation model
To verify the proposed control method and control strategy and
compare with the PR controller in gird-side, a DDWECS system
simulation model based on the Matlab/Simulink is established to
(41) realise the MPE in the machine-side converter and regulate the grid-
the PCI and the PR control method respectively. The steady-state and
dynamic curves of power coefficient of wind turbine Cp , machine-
side q-axial current, machine speed, dc voltage and grid current
378 J. Yan et al. / Renewable Energy 66 (2014) 371e380
Fig. 14. Experimental platform and results. (a) Wind turbine and PMSG. (b) Control system. (c) Measured wind speed. (d) Measured speed of PMSG. (e) d and q axial current in
machine-side. (f) Measured dc voltage. (g) Grid voltage and current.
and voltage are given in Fig. 13 respectively. Before t ¼ 2.5 s, the wind
speed is 6.2 m/s. The PMSG is controlled through the q-axial current
under isd ¼ 0 to realise the MPE with the maximum Cp 0.42, and the
output power is 82 W. At t ¼ 2.5 s, the wind speed increases from
6.2 m/s to the rated value 9.5 m/s. The reference speed of the PMSG is
changed according to the TSR control method. From Fig. 13(s), it can
be seen that the PMSG responds very fast and tracks the target value
within 1 s and the MPE is implemented as shown in Fig. 13(b) by
regulating the machine-side q-axial current. From Fig. 13(c), the q-
axial current tracks the reference value fast and smoothly, which
shows the good performance when the wind speed up. Besides, for
the grid-side, considering the PMSG performance and the startup
wind speed, the dc voltage set at 60 V is appropriate; and the
controller result is shown in Fig. 13(d). The grid-side currents
controlled by the PCI and the PR method are displayed in Fig. 13(e).
The grid-side current controlled by the PCI method can reach the
steady-state within one cycle, which illustrates that the PCI method
has better dynamic performance than the PR method. At t ¼ 4 s, the
wind speed decreases from 9.5 m/s to 7.8 m/s, the speed of the PMSG,
the q-axial current, the dc voltage and the grid-side currents [see
Fig. 13(f)] are also successfully controlled to obtain the MPE. The
simulation results illustrate that the machine-side converter with the
current vector control method based on the RRF and the MPE
through the TSR method is effective and the parameters of the
controllers are correct. The grid-side converter with the PCI current
inner loop with the dc voltage outer loop under the SRF is practical
and has better dynamic performance.
6. Experiments
To verify the proposed control method and the performance of
the grid-connected DDWECS, a small prototype system is con-
structed, as shown in Fig. 14(a) and (b), and some experiments are
carried out. The wind turbine with 1.5 m diameter adopts the
horizontal type. The control system is implemented with a
TMS320F2812 DSP [35], and the switching frequency is 5 kHz.
Fig. 14(c)e(g) shows the measured results of the wind speed, the
speed of the PMSG, the d and q-axial currents of the machine, the dc
voltage and the grid-side current and voltage. Compared with the
simulation model, the natural wind speed is variable and gusty
swiftly. The requirements of the speed response and the tracking
performance of the system are relatively higher to realise the MPE by
quickly tracking the optimal speed of the PMSG under the variable
wind speed shown in Fig. 14(c). Fig. 14(d) shows that the speed of
 J. Yan et al. / Renewable Energy 66 (2014) 371e380 379

PMSG can track the tendency of the optimal reference speed very References
well. Fig. 14(e) displays the d and q components of the stator current
under isd ¼ 0 control, which shows that isd is approximately equal [1] Li S, Haskew TA, Xu L. Conventional and novel control designs for direct driven
PMSG wind turbines. Electr Power Syst Res 2010;80:328e38.
to zero, and isq is proportional to the generator torque. [2] Abbes M, Belhadj J, Bennani ABA. Design and control of a direct drive wind
Fig. 14(f) shows the dc voltage waveform, which is properly turbine equipped with multilevel converters. Renew Energy 2010;35:936e45.
maintained around the setting reference voltage 60 V with minimal [3] Li S, Haskew TA, Swatloski RP, Gathings W. Optimal and direct-current vector
control of direct-driven PMSG wind turbines. IEEE Trans Power Electr
fluctuation through the grid-side controller. The grid-side current 2012;27(5):2325e37.
and voltage are given in Fig. 14(g), in which the grid-side power [4] Koutroulis E, Kalaitzakis K. Design of a maximum power tracking system for
factor is about 0.976 and the THD of the grid current is approxi- wind-energy-conversion applications. IEEE Trans Ind Electr 2006;53(2):486e
94.
mately 2.32%. The results illustrate that the grid-side PCI control [5] Kazmi SMR, Goto H, Guo H, Ichinokura O. A novel algorithm for fast and
method can implement the unity power factor operation and efficient speed-sensorless maximum power point tracking in wind energy
effectively eliminate the harmonics. conversion systems. IEEE Trans Ind Electr 2011;58(1):29e36.
[6] Wang Z, Chang L. A DC voltage monitoring and control method for three-
phase grid-connected wind turbine inverters. IEEE Trans Power Electr
2008;23(3):1118e25.
7. Conclusions [7] Singh M, Khadkikar V, Chandra A. Grid synchronisation with harmonics and
reactive power compensation capability of a permanent magnet synchronous
This paper investigates the machine-side and the grid-side generator-based variable speed wind energy conversion system. IET Power
Electr 2011;4(1):122e30.
converter control methods for DDWECS. The parameters and the
[8] Grauers A. Efficiency of three wind energy generator systems. IEEE Trans
performances of the controllers are calculated and analysed in Energy Convers 1996;11(3):650e7.
details. In gird-side, the comparison between the PCI control [9] Hu J, He Y. Modeling and control of grid-connected voltage-sourced con-
verters under generalized unbalanced operation conditions. IEEE Trans Energy
method and the PR method are analysed and simulated. The
Convers 2008;23(3):903e13.
comparison results show that the PCI controller has faster response [10] Hazra S, Sensarma PS. Self-excitation and control of an induction generator in
time than the PR control method. A simulation model and an a stand-alone wind energy conversion system. IET Renew Power Gener
experimental system are built and constructed respectively to 2010;4(4):383e93.
[11] Khajehoddin SA, Karimi-Ghartemani M, Jain PK, Bakhshai A. A control design
verify the designs and the performance of the controller. The results approach for three-phase grid-connected renewable energy resources. IEEE
illustrate that the current vector control for the machine-side can Trans Sustain Energy 2011;2(4):423e32.
adjust the stator current effectively to change the speed of PMSG to [12] Mahmood H, Jiang J. Modeling and control system design of a grid connected
VSC considering the effect of the interface transformer type. IEEE Trans Smart
realise the MPE, which has good dynamic and steady-state per- Grid 2012;3(1):122e34.
formance; and the PCI inner loop with dc voltage outer control for [13] Hong Y, Lu S, Chiou C. MPPT for PM wind generator using gradient approxi-
the grid-side is able to regulate the dc voltage and grid current with mation. Energy Convers Manag 2009;50(1):82e9.
[14] Espi Huerta JM, Castello-Moreno J, Fischer JR, Garcia-Gil R. A synchronous
better dynamic response. The developed controller is feasible for reference frame robust predictive current control for three-phase grid-con-
DDWECS and has excellent performance. nected inverters. IEEE Trans Ind Electr 2010;57(3):954e62.
[15] Yang C, Smedley K. Three-phase boost-type grid-connected inverter. IEEE
Trans Power Electr 2008;23(5):2301e9.
Acknowledgements [16] Castello Moreno J, Espi Huerta JM, Garcia Gil R, Alejandro Gonzalez S. A robust
predictive current control for three-phase grid-connected inverters. IEEE
Trans Ind Electr 2009;56(6):1993e2004.
This work was supported by the National Natural Science [17] Guo XQ, Wu WY. Improved current regulation of three-phase grid-connected
Foundation of China (51077012), the Specialized Research Fund for voltage-source inverters for distributed generation systems. IET Renew Power
Gener 2010;4(2):101e15.
the Doctoral Program of Higher Education (20090092110033), the
[18] Hwang JG, Lehn PW, Winkelnkemper M. A generalized class of stationary
Department of Science and Technology Project of Jiangsu Province frame-current controllers for grid-connected AC-DC converters. IEEE Trans
of China (BY2011151), and the Senior Expert Special Foundation in 6 Power Deliv 2010;25(4):2742e51.
Fields of Jiangsu Province of China (2011-ZBZZ-016). [19] Rodriguez P, Luna Alvaro, Munoz-Aguilar RS, Etxeberria-Otadui I,
Teodorescu R, Blaabjerg F, et al. A stationary reference frame grid synchro-
nization system for three-phase grid-connected power converters under
adverse grid conditions. IEEE Trans Power Electr 2012;27(1):99e112.
Appendix [20] Rowan TM, Kerkman RJ. A new synchronous current regulator and an analysis of
current-regulated PWM inverters. IEEE Trans Ind Appl 1986;IA-22(4):678e90.
[21] Rolim LGB, da Costa DR, Aredes M. Analysis and software implementation of a
Table 1 robust synchronizing PLL circuit based on pq theory. IEEE Trans Ind Electr
Parameter of the DDWECS. 2006;53(6):1919e26.
[22] Munteanu I, Bratcu AI, Cutululis N, Ceanga E. Optimal control of wind energy
Parameter Value Units systems. London: Springer; 2008.
[23] Kusiak A, Verma A, Wei X. Wind turbine frontier from SCADA. Wind Syst Mag
Psys (rated output power of DDWECS) 200 W 2012;3(9):36e9.
Rs (PMSG stator resistance) 0.9 U [24] El Mokadem M, Courtecuisse V, Saudemont C, Robyns B, Deuse J. Fuzzy logic
Ls (PMSG stator inductance) 3.7 mH supervisor-based primary frequency control experiments of a variable-speed
jf (PM flux) 0.28 Wb wind generator. IEEE Trans Power Syst 2009;24(1):407e17.
p (pole pairs) 4 [25] Iyasere E, Salah M, Dawson D, Wagner J, Tatlicioglu E. Optimum seeking-based
Vdc (dc voltage) 60 V non-linear controller to maximise energy capture in a variable speed wind
Cdc (dc capacitor) 3300 mF turbine. IET Control Theory Appl 2012;6(4):526e32.
Ts (sampling period) 0.2 ms [26] Zhang S, Tseng K, Vilathgamuwa DM, Nguyen TD, Wang X. Design of a robust
grid interface system for PMSG-based wind turbine generators. IEEE Trans Ind
Rg (grid-side resistance) 0.5 U
Electr 2011;58(1):316e28.
Lg (grid-side inductance) 5 mH
[27] Valenciaga F, Puleston PF. High-order sliding control for a wind energy con-
Ug (grid voltage) 220 V
version system based on a permanent magnet synchronous generator. IEEE
fg (grid frequency) 50 Hz Trans Energy Convers 2008;23(3):860e7.
N (ratio of isolated transformer) 15:1 [28] Stiebler M. Wind energy systems for electric power generation. Berlin:
R (radius of wind turbine) 0.75 m Springer; 2008.
Jtotal (inertia of DDWECS) 0.02 Kg m2 [29] Miller A, Muljadi E, Zinger DS. A variable speed wind turbine power control.
lopt (optimal TSR) 5.3 IEEE Trans Energy Convers 1997;12(2):451e7.
Cp max (maximum Cp ) 0.42 [30] De Broe AM, Drouilhet S, Gevorgian V. A peak power tracker for small wind
vn (rated wind speed) 9.5 m/s turbines in battery charging applications. IEEE Trans Energy Convers
nn (rated rotational speed) 640 rpm 1999;14(4):1630e5.
380 J. Yan et al. / Renewable Energy 66 (2014) 371e380

[31] Golnaraghi F, Kuo BC. Automatic control systems. 9th ed. USA: Wiley; 2010. [34] Yazdani A, Iravani R. Voltage-sourced converters in power systems. USA:
[32] Yang S, Fang ZP, Qian Z. A robust control scheme for grid-connected voltage- Wiley; 2010.
source inverters. IEEE Trans Ind Electr 2011;58(1):202e12. [35] Texas Instruments Incorporated. TMS320F2812 digital signal processors
[33] Chinchilla M, Arnaltes S, Burgos JC. Control of permanent-magnet generators manual; 2004.
applied to variable-speed wind-energy systems connected to the grid. IEEE
Trans Energy Convers 2006;21(1):130e5.