Вы находитесь на странице: 1из 5

Frequency Regulation and AGC in Isolated Systems with DFIG-Based Wind Turbines

Mansour Jalali, Senior Member


Smart Grid and Substation Automation Kinectrics Inc, Toronto, Ontario, Canada Mansour.Jalali@kinectrics.com
Abstract Traditionally, wind turbine generation does not participate in frequency regulation services in power systems. With increasing penetration of wind energy, system operators are demanding more and more participation of wind turbines in the ancillary services especially in frequency regulation provisions. Doubly fed induction generators (DFIGs), with their ability to deliver power with different mechanical speed, and the ability to instantly reduce speed, and release the stored mechanical energy is able to support conventional generators in regulating system frequency. This helps reduce the frequency dip efficiently when the DFIG controller is optimally tuned. This paper examines the role of DFIGs, in frequency regulation and control with different levels of wind penetration into the system. A dynamic model of the power system with conventional generator and DFIG-based wind turbines in single area frequency control is developed, and a parameter optimization method is used to tune the controllers of the DFIG so as to improve the dynamic response of the system.

Kankar Bhattacharya, Senior Member


Department of Electrical and Computer Engineering University of Waterloo, Waterloo, Ontario, Canada kankar@uwaterloo.ca

I. INTRODUCTION The proportion of wind-based generation in the total energy production mix has been growing continuously in many parts of the world. It has been reported that renewable energy will account for as much as 10% of global energy supply by 2020, and increase to as much as 50%, by 2050 [1]. Wind energy is one of most favorable clean energy alternatives to replace fossil-fuel based energy sources. Large-scale penetration of wind sources calls for full integration of wind turbines into the power system and requires detailed studies from various perspectives, including their frequency control and regulation service provisions. Researchers have carried out work on Doubly Fed Induction Generators (DFIG)-based wind turbines to demonstrate their possibility to provide separate active and reactive power control. DFIG-based wind turbines are the preferred choice for network operators in comparison to fixedspeed wind turbines in which active and reactive power control is not independent. In [2], the concept of releasing the kinetic energy of a DFIG-based wind turbine when the frequency of the power system is reduced in order to prevent the reduction of system inertia is presented. The use of additional power set-point in the control system of the wind turbine is suggested. This additional power reference is calculated based on the rate of frequency change and inertia of the network. The additional power released by the DFIG-based wind turbine supports the primary

frequency regulation and thus prevents reduction of system inertia. In [3], a control scheme is proposed that allows DFIGbased wind turbines to participate effectively in system frequency regulation. In this approach, wind turbines operate according to a de-loaded optimum power extraction curve such that the active power provided by each wind turbine increases or decreases during system frequency changes. The control strategy defined for the wind turbine to supply primary frequency regulation capability exploits a combination of control of static converters and pitch control, adjusting the rotor speed and active power according to the de-loaded optimum power extraction curve. Results obtained considering a small isolated power system demonstrate the effectiveness of the approach, the drawback of this method is the loss of revenue accrued because the wind generator is not able to generate its maximum power. Extracting the kinetic energy stored in the mechanical system for variable-speed wind turbines is reported in [4]-[6]. The control system based on this concept, utilizes the ability of DFIG based wind turbines to produce power with variable mechanical speed and extracting the kinetic energy to support the primary frequency regulation, based on the Wind Turbine Inertia Control model (Fig.1). The rate of frequency change and inertia of the network is used to emulate the inertia of the wind turbine. In Fig.1 the controller provides a power set-point (P*) that is based on measured speed (m,meas) and measured electrical power (Pel,meas). The power set-point is an input for the converter control that realizes the torque and power by controlling the generator rotor currents. An additional control signal is added (P*f) which adapts P* as a function of the deviation and the rate of change of grid frequency. The emulated inertia of the additional control signal is proportional to its controller parameters Kdf and Kpf. The support to primary frequency control depends on when this additional loop is activated, as the grid frequency exceeds certain limits, by adding this signal to the torque equation to set the torque demand. As the system frequency drops, the set point torque increases, the rotor slows down and kinetic energy is released. This paper examines the role of variable speed based wind turbines, in particular the DFIG, in frequency regulation and control with different levels of wind penetration into the system. In particular the issue of optimal tuning of the DFIG

978-1-4799-1303-9/13/$31.00 2013 IEEE

control system in isolated area frequency control problems is presented as a means to help improve the performance of the system. The dynamic model of the power system with conventional generator and DFIG-based wind turbines in single area frequency control is developed, and a parameter optimization method is used to tune the controllers of the DFIG so as to improve the dynamic response of the system.

2 mech (4) Ek = Ek0 1 2 1 mech 0 Ek0 is dependent on wind speed which varies between zero and 1 per unit and mech1 cannot be smaller than the minimum mechanical rotational speed of the DFIG-based wind turbine. Furthermore, the instantaneous power extracted from the wind turbine cannot exceed the maximum allowed value, as per manufacture data of the machine. These constraints, as stated in (5)-(8), have to be regarded in the control system.

Ek0 = f (Wind Speed)


p
*
df dt

(5) (6) (7) (8)

p *f

0 Ek0 1 pu = f (Wind Speed )

p* df
dt

d dt

p *f

mechmin mech1
Ek 0 + Ek Ek max

p* f

III. DYNAMIC MODELS FOR FREQUENCY REGULATION AND AGC WITH DFIG BASED WIND TURBINES The dynamic model of a power system comprising a conventional prime-mover, a non-reheat type turbine and a DFIG-based wind turbine is represented by the smallperturbation transfer function model in Fig.3. This model simulates the primary frequency regulation following a disturbance and includes conventional system parameters such as load damping factor (D), regulation droop (R), inertia H, governor time constants Th and Tt of the conventional system equivalent unit (governor and turbine). The system dynamic behavior depends on the choice of network parameters and the wind turbine speed controllers Kwp and Kwi. If several generators are connected to the system, the equivalent regulation droop can be determined as follows: 1 1 1 1 = + + ... + (9) R R1 R2 Rn In Fig. 3, Ph is the incremental hydraulic governor valve position change, is the incremental speed of wind turbine, f is the incremental frequency change, X1 is frequency increment after measurement by transducer, X2 is the frequency change after washout filter and X3 is the incremental change in DFIG integral speed control. The incremental model is obtained by linearizing the system around a nominal operating point. The dynamic model in state-space form is obtained from the transfer-function model in Fig.3, as follows: dX (10) = AX + p dt
T X = P X2 X3 P h P g f X1 NC

Fig.1. Principle of DFIG Inertial Emulation Control [7] II. DFIG-BASED WIND TURBINE The basic blocks for a DFIG based wind turbine are shown in Fig.2. DFIG machines with slip-ring rotors are generally designed with three ac windings in the stator and rotor. The stator is normally connected directly to the grid and the rotor is supplied via a frequency converter.

Fig.2. WECS with DFIG based wind turbine The angular velocity of stator rotating field is,

s
ps

= mech

r
pr

(1)

In (1), ps and pr denote the number of stator and rotor poles respectively, s is system frequency, mech is angular velocity of mechanical rotation and r is the rotor current frequency. Depending on direction of the supply frequency, this machine can operate in over- or under-synchronous speed. The kinetic energy released from the wind turbine shaft when the speed is reduced, Ek, can be formulated from the relations below: 1 2 Ek0 = Jmech (2) 0 2 1 2 1 2 Ek0 Ek1 = Jmech Jmech (3) 0 1 2 2

X is the state-vector, p is the perturbation and A, are state and perturbation matrices respectively. The small perturbation transfer function model for AGC studies is shown in Fig.4. The model is similar to that described in Fig.3 except that a secondary control is added at the governor reference to regulate the set-point. The state-vector X is given by:

X = P h P g P ref f

X1 X2 X3 PNC

J = f 2 (t0 + kt )
k =1

(12)

1 R

PD
Ph 1 1 + sTh

Pref +

1 1 + sTt

Pg +

+
PNC
1 1+ sTa

KP 1 + sTP

1 1+ sTr
Tw 1 + sT w
1 R

X 1

The optimal Kwp and Kwi are obtained by searching for the minimum value of J. At first, Kwi is fixed at a certain value and Kwp is varied over a wide range. It is observed that J decreases as Kwp is increased, attains a minimum, JMin, and increases again. The values of Kwp and Kwi that yields the minimum of JMin (J*Min), are the optimal controller parameters. The impact of wind penetration in total energy supply mix is examined by defining a Penetration Index (W):

PNC

max
1 2 sH e +

K wp

max

X 2

+
PNC ,ref

min

K wi s

X 3

min

W =

Total wind generation 100 Total generation from all sources

Fig 1. Primary frequency regulation block diagram

Four cases of DFIG penetration are considered, W = 5%, 10%, 20% and 50%. Figure 5 shows the plot of J versus Kwp for several values of Kwi for W = 20%. It is also observed from Table-1 that the optimal Kwp and Kwi are slightly affected by W and the performance is expected to worsen with a higher W since J*Min increases with penetration level.
TABLE-I: OPTIMAL DFIG CONTROLLER PARAMETRS Kwi Kwp J*Min W 5% 0.05 1.00 0.0151 10% 0.05 1.00 0.0161 20% 0.1 1.23 0.1209 50% 0.1 1.98 0.1235

K AGC s

1 R

PD

Pref +

1 1 + sTh

Ph

1 1 + sT t

Pg +
+

Kp 1 + sT p

f
1 1 + sTr
X 1

PNC
1 1 + sTa

PNC

K wp

1 2sH e

max

sTw 1+ sTw
1 R

K wi s

+
PNC,ref

min

X 3

Fig. 4. Dynamic model for AGC studies with DFIG-based wind turbine IV. OPTIMAL TUNING OF DFIG CONTROLLERS It is necessary to determine the optimal settings of Kwp and Kwi so as to improve the wind generators participation in frequency control. An Integral of Squared Error (ISE) technique is used and a step perturbation of 0.02 per unit is considered. The dynamic performance of the system, for a given set of control parameters, is measured as the squared of the frequency deviation errors for tuning the controller parameters, as stated below. Fig. 5. Optimal DFIG controller parameters for W = 20% V. ANALYSIS AND RESULTS A. Primary Frequency Regulation Simulations are carried out considering the dynamic model of Fig.3 for a 0.02 p.u. step load perturbation with and without DFIG-WT, to examine its contribution in primary frequency regulation. It should be noted that, DFIG-WT refers to Fig.1 where the WT releases kinetic energy instantly by reducing the speed, and therefore contributes to primary frequency regulation, while WT with no DFIG refers to conventional speed control and the WT does not participate in frequency regulation.

J = f 2 (t ) dt
0

(11)

For computational purposes, J is calculated by summing the discrete values at small intervals, over a considerable period until the steady-state is achieved, as given below.

The DFIG-WTs are at their optimal mechanical speed with maximum power obtainable from wind, and wind speed is assumed to remain constant. From the frequency plots in Fig.6, it is observed that the frequency response following the disturbance is improved, by way of lower frequency excursion when DFIG participation is considered. It is also observed that when wind penetration increases from 20% to 50%, although the peak excursion is less, the settling time increases with DFIG-WT. It can be generally concluded that with low wind penetration there is no significant improvement compared to the no-DFIG case. The plot of active power variations following the disturbance, for W = 20% and 50%, is shown in Fig.7 and Fig.8, respectively.

from wind, wind speed remaining constant during the simulation; and the conventional generator is able to supply the additional load after the transient settles. Different plots for these simulations are presented in Figs.10-13.

Figure 8 System dynamics with and without DFIG, W =50%

Fig. 6. Primary frequency response with and without DFIG

Fig. 9. DFIG mechanical speed and power response The frequency response with AGC for W = 20%, with and without DFIG-WT is shown in Fig.10. It is observed that the frequency response with DFIG is improved, by way of lower frequency peak excursion. It is also observed that by using the optimally tuned speed controller parameters, the settling time is improved. It can be generally expected that with higher values of W the settling time will be longer. Figure 11 shows that high penetration of wind energy will have a negative impact on the transient settling time. Peak frequency excursion is reduced in direct relation to the level of DFIG penetration while frequency settling time compared to the no-DFIG-based wind turbine is not improved. Figures 12 and 13 present the responses of the DFIG in providing AGC services. It is seen that when load increases at t = 0, the DFIG instantly releases kinetic energy by reducing the mechanical speed instantly, and increases its output to participate in AGC. Thereafter, the DFIG output decreases since the speed is no longer at the optimal and power extracted from the wind is

Fig. 7. System dynamics with and without DFIG, W = 20% B. Auomatic Gneneration Control (AGC) Simulations are carried out considering the dynamic model of Fig.4 for a 0.02 per unit load perturbation with and without DFIG-WT to examine its contribution in AGC service. Two levels of wind penetration, W = 20% and 50% are considered and the corresponding optimally tuned parameters for the DFIG-WT controller are used. As mentioned earlier, the DFIG-WTs operate at their optimal mechanical speeds with the output corresponding to the maximum power obtainable

reduced. As the DFIG speed controller acts and the optimal speed recovered, the DFIG power output returns to its nominal.

support services, both primary regulation and AGC. A wind generation penetration index is defined and the effect of variation in penetration is studied for the cases of 20% and 50% wind energy penetration in line with global world wind energy penetration targets of 2020 and 2050, respectively.

Fig. 10. AGC with and without DFIG-WT, W = 20% Fig. 12. DFIG generation with AGC

Fig. 11. AGC with and without DFIG, W = 50% VI. CONCLUSIONS The extraction of energy from DFIG need to be coordinated with AGC of conventional generators so as to damp the overshoot and reduce frequency excursions. With increasing penetration of wind energy in the electrical network with a target of 20% of installed power in North America and in most European countries, DFIG controllers need be optimally tuned, in order to optimize the performance of DFIG wind turbines and their performance in participation of frequency regulation. This paper presents a small perturbation, transfer function model in state-space form for isolated power systems with a mix of conventional generation and wind generation sources. The models examine both primary regulation and AGC effects in the two classes of models. DFIG based wind generation is considered for the model development of the wind generation sources. A simple parameter optimization technique based on the ISE criterion is applied to optimally tune the DFIG controller parameters. Dynamic performance analysis is carried out considering the optimally tuned DFIG controllers to examine their contribution to system frequency

Fig. 13. DFIG generation with AGC VII. REFERENCES


[1] [2] Wind Vision for Canada, Report by Canadian Wind Energy Association (CanWEA), http://www.canwea.ca/windvision_e.php J. Ekanayake and N. Jenkins, Comparison of the Response of Doubly Fed and Fixed-Speed Induction Generator Wind Turbines to Changes in Network Frequency, IEEE Transactions on Energy Conversion, Vol. 19, No. 4, December 2004. R. G. de Almeida and J. A. Peas Lopes, Participation of Doubly Fed Induction Wind Generators in System Frequency Regulation, IEEE Transactions on Power Systems, Vol. 22, No. 3, August 2007. J. M. Mauricio, A. Marano, A. Gmez-Expsito and J. L. Martnez Ramos, Frequency Regulation Contribution Through Variable-Speed Wind Energy Conversion Systems, IEEE Transactions on Power Systems, Vol. 24, No. 1, February 2009. J. G. Slootweg, S. W. H. de Haan, H. Polinder, and W. L. Kling, General Model for Representing Variable Speed Wind Turbines in Power System Dynamics Simulations, IEEE Transactions on Power Systems, Vol. 18, NO. 1, February 2003 G. Lalor, J. Ritchie, S. Rourke, D. Flynn and M. J. OMalley, Dynamic Frequency Control with Increasing Wind Generation, Proc. IEEE PES General Meeting, Denver, 2004. J. Morren, S. W. H. de Haan, W. L. Kling, and J. A. Ferreira, Wind Turbines Emulating Inertia and Supporting Primary Frequency Control, IEEE Transactions on Power Systems, Vol. 21, No. 1, February 2006.

[3]

[4]

[5]

[6]

[7]

Вам также может понравиться