26-04-2014, 12:25 PM
The art of modeling and simulation of induction generator in wind generation applications using high-order model
art of modeling and simulation.pdf (Size: 372.73 KB / Downloads: 29)
ABSTRACT
Both fixed-speed squirrel-cage induction generators and variable-speed doubly fed induc-
tion generators are used in wind turbine generation technology. Modeling and simulation
of induction machines using vector computing technique in Matlab/Simulink provides an
efficient approach for further research on wind generation system integration and control.
In this paper, the vector computing technique is applied in modeling and simulation of
induction machines. Free acceleration of squirrel-cage induction generator, active power
and reactive power control of DFIGs in a power system as well as inter-area oscillation
damping control are demonstrated using the proposed model. The modeling approach in
Matlab/Simulink makes controller design and simulation verification effective.
Introduction
Induction machines are important components that serve as power sources and common loads in power systems: pumps,
steel mills, servomotors, to name a few. With the increasing use of renewable energy in recent decades, induction generators
are found to have important applications in wind turbine generation. Wind turbine driven squirrel-cage induction generators
are usually found to interconnect with the grid together with Static Var Compensation (SVC) to support reactive power locally
[1,2]. More recently, the doubly fed induction generator (DFIG) is increasingly used in wind generation. In the case of DFIG,
rotor voltages or currents of the induction generators are being controlled. With changing wind speed, one can adjust the fre-
quency of the injected rotor voltage of the DFIG to obtain a constant-frequency at the stator [3]. The DFIG is an important type
of variable-speed constant-frequency generator. Research on DFIG has shown that it has the ability to control the active power
and the reactive power [4] and to provide frequency support [5]. More control applications of DFIG are yet to be investigated.
Efficient modeling and simulation techniques for induction generators will facilitate the research on wind turbine generation.
Matlab/Simulink is a powerful tool for time domain simulations. It has been adopted in a graduate level course on mod-
eling and simulation of machines in Purdue University [6]. Yet, in most cases, the featured vector computing technique of
Matlab has not been fully explored by electric machine researchers. Using matrix/vector concepts not only simplifies prob-
lems but also contributes to time saving in debugging.
Induction machine Simulink model
The Simulink model in terms of the state space equations (Eq. (1)) is shown in Fig. 1. In this model block, the inputs are
voltage and rotor speed and the output is a current vector. This model is quite simple and easy to understand. It saves not
only on model building time but also debugging time. The rotor speed is calculated through Eq. (7) which is shown in Fig. 2.
The rotor speed will be fed back to the input of the block in Fig. 1. The induction machine serves as a current source to the
network and the output from the network is the voltage vector. Thus, the induction machine and the power system network
are interconnected and as long as the initial condition is set, dynamic simulation can be performed.
The entire induction generator model block consists both the swing equation and the current state space model. The de-
tailed model in Simulink is shown in Fig. 3.
Interconnection of DFIG model in power systems
Modeling wind farms interconnected to the grid is important for the transient analysis of the entire system. In modeling a
power system, the network is usually treated as a Y matrix with the current and voltage relationship as I 1⁄4 YV. All generators
are treated as current sources. From the current sources, the system voltages can be computed as in Fig. 6. The voltages will
then be used in the differential equations expressed in terms of state variables I and input variables V. In the case of a DFIG,
the current to the network is the sum of the stator current and the current from the converter and the network voltage is the
stator voltage. In a power system, the voltage phasors are all based on a reference bus while the rotor angles are all based on
a reference machine.
Damping control of DFIG
Since the current loop is very fast and its bandwidth is very high compared to the damping control bandwidth, we will not
put a supplementary signal at the current control loop. Instead, we propose to add the supplementary signal at the active
power control loop. Since inter-area oscillation is a phenomenon related to the rotor angle and active power, active power
modulation is an effective method for oscillation damping in power systems.
The rotor angle difference has a good observability of the inter-area oscillation mode between the two areas [13]. The
angle difference signal can be obtained through a state-of-art Phasor Measurement Unit technology. In this paper, we as-
sume that the angle difference signal is available.
The open loop frequency responses of two different systems are compared in Fig. 12. The first system has the input/output
pair as P modulation versus the rotor angle difference. The second system has the input/output pair as Q modulation and the
rotor angle difference.
Conclusion
This paper proposes a methodology to model induction machines in Matlab/Simulink using matrix/vector concept. This
methodology greatly saves the modeling time and debugging time. Models built in this way are easy to be understood by students and engineers. The integration of the developed DFIG model with the network is presented in the paper. The inter-
connecting technique takes into consideration the DFIG vector control scheme by transferring voltage and current phasors
between two reference frames. The design of the DFIG rotor side converter control is also demonstrated in this paper. The
effectiveness of the inner current control loops, the active and reactive control loops as well as the inter-area oscillation
damping control loops is also demonstrated through simulation.