Reconfiguration of Control Strategies for High Power DFIG Wind Turbine System to Meet Grid Code Requirements
Ling Peng
Department of Electrical Engineering Tsinghua University
100084 Beijing, China [email protected]
Yongdong Li
Member, IEEE
Department of Electrical Engineering Tsinghua University
100084 Beijing, China [email protected]
Bruno Francois
Senior Member, IEEE L2EP, Ecole Centrale de Lille
Cite Scientifique 59651 Villeneuve d’Ascq, France
Abstract -- Nowadays, there is a tendency to increase the capacity of wind turbines, multilevel converter is a realistic alternative in such high power applications. As the wind power penetration continues to increase, grid codes have been revised to require wind farms to operate as conventional power plants.
This paper proposes a reconfiguration scheme of control strategies for doubly fed induction generator (DFIG) based wind turbine system, which dominates the wind market at present due to the significant advantages. Simulation and experimental results have proven excellent performances in both normal and fault conditions and verified that the studied system can completely meet the latest gird code requirements with the proposed reconfiguration methodology.
Index Terms--Doubly fed induction generator (DFIG), multilevel converters, fault ride-through, reconfiguration, wind energy.
I. INTRODUCTION
Wind energy has been noted as the fastest growing renewable power generation technology in the world, with an annual growth rate in excess of 30% and a foreseeable penetration of 12% of global electricity demand by 2020 [1].
Nowadays, there is a tendency to increase the size and capacity of wind turbines, with 5MW as the highest power class of the present installed turbines. For such high power applications, multilevel converters, are preferred due to their significant advantages such as they can increase the output voltage magnitude, reduce the output voltage and current harmonics without increasing the switching frequency or decreasing the output power [2]. In addition, they present better efficiency as the converter losses are reduced. Thus multilevel converters, especially three-level converters, are realistic alternative to conventional converters in high power wind energy applications.
As the wind power penetration continues to increase, the intermittent nature of the wind has created severe challenges for power system operators. So the grid codes have been revised to require wind farms to operate as conventional power plants, with focus on power quality, grid stability, control capabilities and fault ride-through performance.
Normally, the control strategy is designed to extract the maximum power from the wind as well as satisfy reactive power constrain [3]. However, in presence of grid faults,
wind turbines should be controlled to remain connected to the grid and supply adequate reactive power to help the grid voltage recovery [4-6]. Therefore, reconfiguration of the control strategies for the wind turbine system is obligatory according to the operation condition.
This paper aims to implement the reconfiguration scheme on doubly fed induction generator (DFIG) based wind turbines, which dominates the wind market at present. Firstly, the equivalent continuous model of a high power DFIG wind turbine using back-to-back three-level neutral point clamped (NPC) converters is presented with a graphical tool Energetic Macroscopic Representation (EMR). Then a Maximum Control Structure (MCS) is deduced from this EMR through specific inversion rules. Different control strategies are proposed to fulfill modern gird code requirements under various operation conditions. Simulation and experiments have been performed to verify the validity of the proposed methodology in both normal and fault cases.
II. MODELING OF THE SYSTEM
Figure 1 shows the overall wind generation system. The stator of DFIG is directly connected to the grid while the rotor is interfaced through three-level back-to-back power converters. The grid side converter is connected to the grid via three chokes to filter the current harmonics. The system is described using Energetic Macroscopic Representation (EMR), which is a synthetic graphical tool between connected elements. This tool has previously been developed to propose a synthetic and dynamical description of electromechanical conversion systems [7].
A. Modeling of the Turbine
The wind turbine is characterized by its aero dynamical torque, which is given by:
3 1
( , )
p 2
t
T C= λ β ρS⋅ ⋅v
Ω (1) Where Ωtis the angular speed of the turbine, v is the wind velocity, S is the swept area of the turbine and ρ is the air density.
PEDS2009
Fig. 1. High power DFIG based wind turbine system using three-level back-to-back NPC converters The power coefficient Cp(λ β, ) represents the
aerodynamic efficiency of the wind turbine. It depends on the blade design, the tip speed ratio λ and the pitch angle of the blades β. The tip speed ratio is defined as the ratio between the blade tip speed and the wind speed:
R t v
λ= Ω (2) With R is the blade radius.
B. Modeling of the Mechanical Parts
The drive train is composed of the mass corresponding to the large turbine rotor inertia representing the blade, the hub and a small inertia representing the rotor mass of the generator.
With considering the total inertia of the equivalent shaft J, the acceleration can be simply modeled by a first order differential equation:
d g t
J T T f
dtΩ = − − Ω (3) Where T is the torque from the gearbox and f is the g friction coefficient of the equivalent shaft.
The gearbox transfers the aero dynamical power from the slow rotating turbine shaft to the fast rotating generator shaft through the gear radio G, which drives the generator at the mechanical speed Ω. It is mathematically described by the following equations:
t
g em
Ω GΩ
T GT
⎧⎪ =
⎨ =⎪⎩ (4) With T is the electromagnetic torque of the generator. em C. Modeling of the DFIG
In order to represent a more detailed dynamic behavior of the generator current, a fifth-order dynamic model of the DFIG is used in this paper. The model in a synchronously rotating d-q reference frame is given in the form:
d d
d d
sd
sd s sd s sq
sq
sq s sq s sd
v R i t
v R i t
ψ ω ψ
ψ ω ψ
⎧ = − +
⎪⎪⎨
⎪ = − −
⎪⎩
(5)
d d
d d
rd
rd r rd r rq
rq
rq r rq r rd
v R i t
v R i t
ψ ω ψ
ψ ω ψ
⎧ = − +
⎪⎪⎨
⎪ = − −
⎪⎩
(6)
sd s sd rd
sq s sq rq
= L i Mi = L i Mi ψ
ψ
⎧⎪ +
⎨ +
⎪⎩ (7)
rd sd r rd
rq sq r rq
= Mi L i = Mi L i ψ
ψ
⎧⎪ +
⎨ +
⎪⎩ (8) In these equations,R ,s R ,r L and s L are the resistances r
and inductances of the stator and rotor windings, M is the mutual inductance, v v v v i i i i sd, sq, rd, rq, sd, sq, rd, rq,ψ ψs d, sq, ψrdand ψ are the d and q components of the space vectors rq
of the stator and rotor voltages, currents and flux, ωr is the slip angular speed and ωs is the synchronous speed of the generator.
The electromagnetic torque is expressed with stator quantities and the number of pole-pairs p:
Tem=p(ψsd sqi −ψsq sdi ) (9) D. Modeling of the Three-level NPC Converters
The topology of a three-level NPC converter consists of three commutation circuits, which are fed with a capacitive divider, as it is shown in Fig. 2. To produce three voltage levels, two capacitors are connected in series to a dc voltage source udc.
PEDS2009
Fig. 2. Schematic diagram of a three-level NPC converter For the modeling of power electronic converters we will consider that semiconductors are ideal: no power dissipation and instantaneous commutations. Then the clamped commutation circuit is equivalent to a commutation circuit with three ideal switches. And one ideal switch among the three must be at any time switched on.
By considering a continuous conduction mode of semiconductors, an equivalent matrix converter with ideal switches as Fig. 3 can be considered for an easier study.
Fig. 3. Equivalent matrix structure of the three-level converter A switching function fij is defined for each power switch.
It represents the ideal commutation orders and takes the values 1 when the switch is closed and 0 when it is open:
o o
{1,2,3} n of the leg {0,1}, with
{1,2,3} n of the switch in the leg
ij
f i
j
∈ ⎧ ∈⎨
⎩ ∈ (10)
This defined switching function is the theoretical (and considered effective) state of the switch. In each commutation circuit, the last switching function is decided by the other ones as:
3 1 2
f j = f j⋅f j, j∈{1,2,3} (11) The modulated voltages may be written as a product of the DC bus voltage uc1, uc2 and modulation functions, which can be defined from switching functions [8]:
11 21
11 12
12 22
21 22
13 23
1 0 1
0 1 1
m m f f
m f f
m m
f f
⎡ ⎤
−
⎡ ⎤ ⎡ ⎤ ⎢ ⎥
=⎢⎣ ⎥ ⎢⎦=⎣ − ⎥ ⎢⎦⎢⎣ ⎥⎥⎦
(12)
Then the modulated voltages and current of the three-level NPC converter can be expressed as:
1 1 2
2 2
m c c
m
m c
u u u
u m mu
u u
+
⎡ ⎤ ⎡ ⎤
=⎢ ⎥= ⎢ ⎥=
⎣ ⎦ ⎣ ⎦ (13)
1 1
2 2
m T = T
m m
i i
i m m i
i i
⎡ ⎤ ⎡ ⎤
=⎢ ⎥= ⎢ ⎥
⎣ ⎦
⎣ ⎦ (14) In this way, the rotor side and grid side converter can be modeled respectively as:
r r
T
rm r r
v m u
i m i
=
⎧⎨ =
⎩ (15)
t g
T
gm g t
v m u
i m i
⎧⎪ =
⎨ =
⎪⎩ (16) With vt is the output voltage of grid side converter and it is the current in the filter, irm and igm are the modulated currents of relevant power converters respectively.
E. Modeling of the Other Electrical Parts
As shown in Fig. 4, the DC bus of back-to-back three- level NPC converters is shared with two capacitors.
Fig. 4. DC bus of the three-level NPC converter
From the Kirchhoff’s current law, the voltage dynamics of the capacitors can be obtained as follows:
1 1 1
2 2 1 2
1 1
2 2
c rm gm
c rm c gm
c c
c c
i i i
i i i i
Cdu i dt Cdu i
dt
= −
⎧⎪
= + −
⎪⎪⎪ =
⎨⎪
⎪ =
⎪⎪⎩
(17)
With C is the capacity value of each capacitor.
And filter currents can be deduced from following differential equations:
t dit t t t g
L v R i v
dt = − − (18) Where Lt is the filter inductance and Rt is the filter
PEDS2009
resistance.
III. RECONFIGURATION OF CONTROL STRATEGIES A Maximum Control Structure (MCS) is obtained from EMR using inversion rules [7]. The EMR and MCS of the whole system are depicted in Fig. 5.
A. General Description
A power monitoring gives a prescribed torque-speed curve that the DFIG has to track [5]. The pitch angle control is implemented to prevent over-speeding of the turbine by setting the speed reference to the maximum value. Vector control based on stator flux orientation has been developed for the control of DFIG. A Space-Vector Hysteresis Current Control (SVHCC) with neutral point voltage balancing is applied on three-level NPC converters [6]. The rotor side converter is designed to control the electromagnetic torque and the stator reactive power respectively. Meanwhile the grid side converter keeps the DC-link voltage constant and guarantees a converter operation with a wished power factor as it is described in [5].
B. Vector Control of the DFIG
Vector control algorithms for the DFIG have already been discussed in a variety of publications [3-6].
With the stator flux orientation, the dynamic model of the DFIG becomes:
+
0
sd s sd
sq s sq s sd
s s sd rd
s sq rq
v R i v R i
L i Mi L i Mi ω ψ ψ
=
⎧⎪ =
⎪⎨ = +
⎪⎪ = +
⎩
(19)
Neglecting the voltage drop across the stator resistance, it becomes:
sd 0
sq s sd
v
v ωψ
⎧⎪ =
⎨ =
⎪⎩ (20) Then the torque expression (9) can be simplified as:
em sd sq sd rq
s
T p i p Mi
ψ ψ L
= = − (21)
By tacking into consideration the chosen reference frame, the reactive power at the stator side can be written as follow:
Qs =v isq sd (22) From (19) and (20), the stator reactive power can be further expressed as:
s s sd s rd
s s
v v M
Q i
L L
= ψ − (23)
In this way, the electromagnetic torque and the stator reactive power can be controlled independently.
C. Normal Operation Mode
A power monitoring gives a prescribed torque-speed curve that the DFIG has to track. After the starting, a control strategy is designed to extract the maximum power from the wind by setting a torque reference in the Maximum Power Point Tracking (MPPT) region.
From (21), the electromagnetic torque is proportional to the q-axis component of the rotor current, so a torque control can be achieved by setting the reference of this current as:
_ _
_
rq ref 1 s em ref
sd ref
i L T
pψ M
= − (24)
Where the stator flux reference can be obtained from (20):
_ _
sd ref_ sq ref s ref
s s
v v
ψ = ω = ω (25)
Fig. 5. EMR and MCS of the DFIG wind turbine system
PEDS2009
And from (23), the generated reactive power can be regulated by the d-axis component of the rotor current:
_
_ _
_
rd ref sd ref s s ref
s ref
i L Q
M v M
=ψ − (26)
D. Fault Operation Mode
Due to the relative small rating of the power converters compared to the system, wind turbine system based on DFIG is very sensitive to grid disturbances. The abrupt drop of the grid voltage will cause over-current in the rotor windings. A space-vector hysteresis current controller is adopted to control the rotor side NPC converter, which guarantees the rotor current in a safe range.
Depending on the latest grid code, wind turbines are required to provide active power in proportion to the retained voltage during voltage dips. Thus the reference of q-axis rotor current should remain unchanged in the case of grid faults, which can still be calculated by (24).
If the fault lasts only for a short time, the d-axis rotor current can also be controlled to remain unchanged.
However, if it lasts for a long time, for example, more than 150 ms, the wind turbine is required to supply a maximum reactive current to the grid without exceeding generator limit.
Thus the reactive component of the rotor current should be set to:
2 2
_ _ max
ird ref = ir −irq
(27) Moreover, the wind turbine should resume active power exportation as quickly as the technology allows and in any event within 1 second of the grid voltage recovering to the normal operating range.
IV. SIMULATION STUDY
In order to study the operation performance of the system, simulation has been carried out using Matlab Simulink™, by considering a 2MW DFIG. A three-phase voltage dip of 50%
depth and 500 ms duration at the stator terminal will be considered.
Fig. 6 shows the simulated results. In normal condition, the generator is controlled to operate in unity power factor mode. After the voltage dip occurs at 3s, the rotor current maintains almost unchanged in 200 ms thanks to the proposed space-vector hysteresis current control strategy, as it is shown in Fig. 6. Then as the dip lasts for more than 200 ms, the reconfiguration scheme improves the reactive component of rotor current to provide ancillary reactive power in order to assist the grid voltage recovery, as we can see in Fig. 6. It can be noticed from Fig. 6 that the generator can also supply little reactive power before the reconfiguration scheme activated. This is because the stator flux drops during the voltage dip. From (25) and (26), although the reactive power reference is set to be zero, the actual value is still positive due to the voltage dip. From Fig.
6, the generator can provide adequate active power in
proportion to the retained voltage during the dip, and it resumes normal operation shortly after the clearance of the fault.
2.8 3 3.2 3.4 3.6
-1 -0.5 0 0.5 1
Grid voltage (p.u.)
2.8 3 3.2 3.4 3.6
-1 -0.5 0 0.5 1
Rotor current (p.u.)
2.8 3 3.2 3.4 3.6
0 0.2 0.4 0.6 0.8 1
Active power (p.u.)
2.8 3 3.2 3.4 3.6
0 0.2 0.4 0.6 0.8 1
Reactive power (p.u.)
Times (s)
Fig. 6. Simulation results with reconfiguration scheme under a voltage dip of 50% which lasts for 0.5 s
V. EXPERIMENTAL VERIFICATION
An experiment platform has been built up to validate the PEDS2009
proposed control scheme. The schematic diagram of the DFIG control system is shown in Fig. 7.
Fig. 7. Schematic diagram of the DFIG control system
A 4.5kW DFIG is driven by a permanent magnet synchronous motor (PMSG) which simulates the wind turbine.
(a) 1: Grid voltage (80 V/div); 3: Rotor current (4 A/div)
(b) 1: Grid voltage (80 V/div); 2: Active power (0.5 kW/div);
3: Reactive power (0.5 kVar/div)
Fig. 8. Experimental results with reconfiguration scheme under a voltage dip of 50% which lasts for 0.5 s
The PMSG is controlled by a variable speed driver that provides speed regulation. The stator of DFIG is connected to the grid through three-phase choke which represents the connection cable between the generator and the PCC. And the generator rotor is feeding by a three-level converter. The control system is based on the dSPACE DS1104 DSP board.
The experiment results, which prove excellent performance of the system as well as good consistency with the simulated results, are presented in Fig. 8. As we can see, with the proposed reconfiguration strategy, the system can ride- through the grid fault and resume normal operation within 1 second after the grid voltage recovers to rated value.
VI. CONCLUSION
In this paper, a reconfiguration of control strategies for DFIG based wind turbine system is proposed to meet the latest grid code requirements. With the proposed methodology, the system can supply adequate active and reactive power to satisfy the power system constrain.
Moreover, the fault ride-through capability of the DFIG system can be greatly enhanced with providing ancillary reactive power under long duration voltage dips. And it can resume normal operation immediately after the clearance of the fault. As the grid side converter can also supply reactive power during the voltage dip, future work will be done to add appropriate control strategies of grid side converter in the reconfiguration system.
REFERENCES
[1] C. Millais and S. Teske, "Wind force 12: A blueprint to achieve 12% of the world’s electricity from wind power by 2020," Brussels: European Wind Energy Association, Tech. Rep., 2005.
[2] J. Rodriguez, J. S. Lai, and F. Z. Peng, “Multilevel Inverters: A Survey of Topologies, Controls, and Applications,” IEEE Trans. Ind. Appl., vol. 49, no. 4, pp. 724–736, Aug. 2002.
[3] R. Pena, J. C. Clare and G. M. Asher, “Doubly fed induction generator using back-to-back PWM converters and its application to variable speed wind-energy generation,” IEE Proc. Electr. Power Appl., vol.
143, no. 3, pp. 231–241, May 1996.
[4] J. Lopez, P. Sanchis, X. Roboam and L. Marroyo, “Dynamic Behavior of the Doubly Fed Induction Generator During Three-Phase Voltage Dips,” IEEE Trans. Energy Conv., vol. 22, no. 3, pp. 709–717, Sept.
2007.
[5] L. Peng, Y. Li and B. Francois, “Modeling and Control of the DFIG Wind Turbine System using Causal Ordering Graph during Voltage Dips,” in Proc. ICEMS Conf., Wuhan, China, Oct. 2008, pp. 2412–
2417.
[6] L. Peng, B. Francois and Y. Li, “Low Voltage Ride-Through of High Power DFIG Wind Turbine using Three-level NPC Converters,” in Proc. IECON, Porto, Portugal, 2009.
[7] A. Bouscayrol, B. Davat, B. de Fornel, B. Francois, J. P. Hautier, F.
Meibody-Tabar and M. Pietrzak-David, “Multi-machine multi- converter system: application for the electromechanical conversion,”
EPJ Appl Phys 2000, 10(2), pp. 131-47.
[8] B. Francois, E. Semail, “Modeling and control of a three-phase neutral point clamped inverter by means of a direct space vector control of line to line voltages,” Proceedings of EPE-PEMC, Croatia, September 9–11, 2002.
PEDS2009
