MODELING A SELF-EXCITED INDUCTION GENERATOR

Marcelo G. Simoes
Electrical Engineering Department
Colorado School of Mines, Golden, CO 80401-1887

Abstract - Induction machines have been widely used in loads. As the load and speed of the SEIG changes, the demand
power generation and other industrial applications for for lagging VARs to maintain a constant AC voltage across
decades due to their numerous advantages like robust the machine terminals also changes. Fig. 1 presents a self-
design and lower cost as compared to synchronous excitation process without saturation effect. As the
machines. When a standalone induction generator like in magnetizing current increase, the output voltage will keep
isolated wind farm and micro hydro is driven by a increase, this is a positive feedback process and it will not
mechanical prime mover, the residual magnetism in the stop, because there is no intersection between B-H curve of
rotor of the machine induces an electro-magnetic field machine and capacitor V-I curve.
(EMF) in the stator windings at a frequency proportional
to the rotor speed. This EMF is applied to the capacitors
(cap-bank) connected to the stator terminals and causes
reactive current to flow in the stator windings. The cap-
bank provides necessary reactive power to run the
generator as well as reactive power to the inductive loads
connected to the generator. Multiple generators can be
connected in parallel to increase reliability and to provide
required power for increasing loads. This paper presents
a PSIM simulation model of a self-excited induction
generator (SEIG) using the well-known 𝜶𝜶𝜶𝜶𝜶𝜶 three axis
model for induction machines (IM). Transient response of
standalone generator, two SEIG generators connected in Fig. 1. Self-excitation process for induction generator neglecting
the effects of saturation
parallel, and SEIG connected to grid is also part of this
study. This paper is expected to be a learning module for When saturation effect is added, the positive feedback will
electrical engineering students to understand dynamics of stop whenever the machine B-H curve and capacitor V-I
the induction machines. intersect. Fig. 2 presents B-H characteristic curve of an
induction machine and V-I curve of a capacitor to explain
I.INTRODUCTION TO SEIG voltage buildup process in induction generators.
Induction generator is used to generate power in wind
energy farms and micro hydro because of its robust structure
and it is cheaper than synchronous generator. When a
standalone induction machine is driven by a mechanical prime
mover, the residual magnetism in the rotor of the machine
induces an EMF in the stator windings at a frequency
proportional to the rotor speed. This EMF is applied to the
capacitors connected to the stator terminals and causes
reactive current to flow in the stator windings. Hence a
magnetizing flux in the machine is established. The final value
of the stator voltage is limited by the magnetic saturation
within the machine. The induction machine is then capable of
operating as a generator in isolated locations without a grid
supply. Once the machine is self-excited and loaded, the
magnitude of the steady-state voltage generated by the SEIG Fig. 2. Self-excitation process for induction generator with the
is determined by the nonlinearity of the magnetization curve, effects of saturation
value of self-excitation capacitance connected to generator Figure 2 shows the capacitor value has an important effect on
terminal, speed, other machine parameters, and terminal the self-excitation process of the induction generator,
especially for the terminal voltage of induction generator. II. PSIM MODEL OF SEIG
From the Fig. 5, it can be observed that the final terminal
voltage is determined from the intersection point of two A PSIM model of induction machine is developed using
curves. As the value of capacitor changes, the intersection αβγ-three-axis modeling technique. The αβγ-three-axis
point will change which change the terminal voltage. Let us dynamic model represents a three-wire fed induction machine
assume that the machine is operating at point 1 in the Fig. 3. in a stationary reference frame. This model does not affect
If a capacitor of value C2, is connected, the new generated neither the stator quantities nor the three-phase model of
voltage will he determined by the machine characteristic devices connected to the stator [1, 2]. In this paper the
curve, and will evolve from point 1 to point 2 with a definite
saturation effects on the magnetizing inductance is included.
time-constant. Conversely, if the capacitor is reduced from a
value C1, at point 1 to a value C0, smaller than C1, the Simulation results have shown satisfactory results during the
magnetizing current will not be compensated by an identical self-excitation process of an induction generator and during
but leading current component. In this case, the machine will parallel operation with another generator or with grid.
first try to save the situation by dropping some of its
Induction Machine Three Phase Model (𝜶𝜶𝜶𝜶𝜶𝜶 − 𝑴𝑴𝑴𝑴𝑴𝑴𝑴𝑴𝑴𝑴)
magnetizing current to reach another stable operating point
along its magnetizing curve. This results in a lower air gap Mathematical equations describing stator and rotor are given
voltage and a consequently lower flux; the slip will increase below separately with their corresponding flux linkage
and finally the machine will stop generating (intersection at equations.
zero in Fig. 3) [4]. Size of capacitor bank used in this research
along with the magnetization curve of a real induction • Stator
machine is given in section III. 𝑎𝑎 𝑎𝑎 𝑑𝑑�𝜆𝜆𝑎𝑎
𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎 �
[𝑣𝑣𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎 ] = 𝑟𝑟𝑠𝑠 [𝑖𝑖𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎 ]+ (1)
𝑑𝑑𝑑𝑑

𝑑𝑑𝜆𝜆𝑎𝑎𝑎𝑎
𝑣𝑣𝑎𝑎𝑎𝑎 = 𝑟𝑟𝑠𝑠 𝑖𝑖𝑎𝑎𝑎𝑎 + 𝑑𝑑𝑑𝑑
𝑑𝑑𝜆𝜆𝑏𝑏𝑏𝑏
𝑣𝑣𝑐𝑐𝑐𝑐 = 𝑟𝑟𝑠𝑠 𝑖𝑖𝑏𝑏𝑏𝑏 + 𝑑𝑑𝑑𝑑
𝑑𝑑𝜆𝜆
𝑣𝑣𝑐𝑐𝑐𝑐 = 𝑟𝑟𝑠𝑠 𝑖𝑖𝑐𝑐𝑐𝑐 + 𝑑𝑑𝑑𝑑𝑐𝑐𝑐𝑐

• Stator flux linkage

,𝑎𝑎
𝜆𝜆𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎 = (𝐿𝐿𝑙𝑙𝑙𝑙 + M)[𝑖𝑖𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎
𝑎𝑎 ] + [𝐿𝐿𝑠𝑠𝑠𝑠 ]�𝑖𝑖𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎 � (2)
𝑑𝑑[𝜆𝜆𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎 ] 𝑑𝑑[𝑖𝑖𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎 ] 𝑑𝑑[𝑖𝑖𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎 ] 𝑑𝑑𝑑𝑑
𝑑𝑑𝑑𝑑
= 𝐿𝐿𝑙𝑙𝑙𝑙 𝑑𝑑𝑑𝑑
+M 𝑑𝑑𝑑𝑑
+ [𝑖𝑖𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎 ]
𝑑𝑑𝑑𝑑
,
𝜆𝜆𝑎𝑎𝑎𝑎 = 𝐿𝐿𝑙𝑙𝑙𝑙 𝑖𝑖𝑎𝑎𝑎𝑎 + 𝑀𝑀(𝑖𝑖𝑎𝑎𝑎𝑎 + 𝑖𝑖𝛼𝛼𝛼𝛼 )
,
𝜆𝜆𝑏𝑏𝑏𝑏 = 𝐿𝐿𝑙𝑙𝑙𝑙 𝑖𝑖𝑏𝑏𝑏𝑏 + 𝑀𝑀(𝑖𝑖𝑏𝑏𝑏𝑏 + 𝑖𝑖𝛽𝛽𝛽𝛽 )
,
Fig. 3. Magnetizing curve of the induction generator with V-I 𝜆𝜆𝑐𝑐𝑐𝑐 = 𝐿𝐿𝑙𝑙𝑙𝑙 𝑖𝑖𝑐𝑐𝑐𝑐 + 𝑀𝑀(𝑖𝑖𝑐𝑐𝑐𝑐 + 𝑖𝑖𝛼𝛼𝛼𝛼𝛼𝛼 )
curve of capacitor

Once the voltage buildup process is completed and generator • Rotor

terminal has rated voltage, load can be connected across the 𝑑𝑑�𝜆𝜆,𝑎𝑎
𝑎𝑎𝑎𝑎𝑐𝑐𝑟𝑟 �
,𝑎𝑎
0 = r𝑟𝑟, �𝑖𝑖𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎 �+ 𝑑𝑑𝑑𝑑
(3)
terminals. However, if the load is inductive it will need
reactive power to function which should be supplied from 𝑑𝑑𝜆𝜆,𝛼𝛼𝛼𝛼 �𝜆𝜆,𝛽𝛽𝛽𝛽 −𝜆𝜆,𝛾𝛾𝛾𝛾 �
additional cap-bank. It is because induction generator lacks 0 = r𝑟𝑟, 𝑖𝑖𝛼𝛼𝛼𝛼
,
+ 𝑑𝑑𝑑𝑑
+
√3
field winding and it cannot generate reactive power for the 𝑑𝑑𝜆𝜆,𝛽𝛽𝛽𝛽 �𝜆𝜆,𝛾𝛾𝛾𝛾 −𝜆𝜆,𝛼𝛼𝛼𝛼 �
0= ,
r𝑟𝑟, 𝑖𝑖𝛽𝛽𝛽𝛽 + 𝑑𝑑𝑑𝑑
+
load. Hence the size of capacitor connected across the √3
generator needs to large enough to supply reactive power both 𝑑𝑑𝜆𝜆,𝛾𝛾𝛾𝛾 �𝜆𝜆,𝛼𝛼𝛼𝛼 −𝜆𝜆,𝛽𝛽𝛽𝛽 �
0 = r𝑟𝑟, 𝑖𝑖𝛾𝛾𝛾𝛾
,
+ +
for the generator operation and for the load. 𝑑𝑑𝑑𝑑 √3

The following sections present mathematical equations • Rotor flux linkage equations
describing abovementioned characteristics of induction
generators and a simulation model based on the equations. 𝜆𝜆,𝑎𝑎 , ,𝑎𝑎 𝑇𝑇 𝑎𝑎
𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎 = �𝐿𝐿𝑙𝑙𝑙𝑙 + M��𝑖𝑖𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎 � + [𝐿𝐿𝑠𝑠𝑠𝑠 ] [𝑖𝑖𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎 ] (4)
Also, parallel operation of two identical SEIG and operation
of SEIG with grid are analyzed. 𝑑𝑑�𝜆𝜆,𝛼𝛼𝛼𝛼𝛼𝛼𝛼𝛼 � ,
𝑑𝑑�𝑖𝑖𝛼𝛼𝛼𝛼𝛼𝛼𝛼𝛼 � 𝑑𝑑[𝑖𝑖𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎 ] 𝑑𝑑𝑑𝑑
𝑑𝑑𝑑𝑑
= 𝐿𝐿,𝑙𝑙𝑙𝑙 𝑑𝑑𝑑𝑑
+M 𝑑𝑑𝑑𝑑
+ [𝑖𝑖𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎 ]
𝑑𝑑𝑑𝑑
 𝜆𝜆,𝛼𝛼𝛼𝛼 = L,𝑙𝑙𝑙𝑙 𝑖𝑖𝛼𝛼𝛼𝛼
, ,
+ 𝑀𝑀(𝑖𝑖𝑎𝑎𝑎𝑎 + 𝑖𝑖𝛼𝛼𝛼𝛼 )
, , , ,
𝜆𝜆𝛽𝛽𝛽𝛽 = L𝑙𝑙𝑙𝑙 𝑖𝑖𝛽𝛽𝛽𝛽 + 𝑀𝑀(𝑖𝑖𝑏𝑏𝑏𝑏 + 𝑖𝑖𝛽𝛽𝛽𝛽 )
𝜆𝜆,𝛾𝛾𝛾𝛾 = L,𝑙𝑙𝑙𝑙 𝑖𝑖𝛾𝛾𝛾𝛾
, ,
+ 𝑀𝑀(𝑖𝑖𝑐𝑐𝑐𝑐 + 𝑖𝑖𝛼𝛼𝛼𝛼𝛼𝛼 )

• Electromagnetic torque (T𝑒𝑒 )

P
Te = √3( )M(ias iγr − ics iαr ) (5)
2
2 𝑑𝑑𝜔𝜔𝑟𝑟 2
𝑇𝑇𝑒𝑒 = � � 𝐽𝐽 + � � 𝐵𝐵𝑚𝑚 𝜔𝜔𝑟𝑟 + 𝑇𝑇𝐿𝐿
𝑃𝑃 𝑑𝑑𝑑𝑑 𝑃𝑃

The symbols and scripts used in equations denote:

• Symbols: [v] is phase voltage vector; [i] is line-current

vector; [λ] is flux vector; [Lsr] is the stator and rotor Fig. 5. Flux calculation circuit
mutual inductance matrix; r is resistance; L is inductance;
• M is the air gap magnetizing inductance; Lms is the
mutual inductance between stator (or rotor) windings of
two distinct phases; ωr is the rotor electrical speed in rd/s
in relation to the reference axis; ωmr is the rotor
mechanical speed in rpm; and P is the number of poles.
• Subscripts: ‘abc’ refers to the a, b and c phases
respectively, ‘s’ refers to the stator, ‘r’ refers the rotor, ‘l’
indicates leakage inductances.
• Superscripts: ‘a’ refers to the abc reference-frame model,
written in machine variables and a single inverted comma Fig. 6. Coupling of electromechanical torque
indicates rotor variables and parameters referred to the
number of turns of the stator windings. M is magnetizing inductance in Henry. It can be obtained
• In order to make easier to understand how above from various test performed on the machine. The magnetizing
equations are modeled in PSIM, step by step description current of the induction machine [𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼] in (2) and (4) is the
is presented below with figures. The Fig. 4 shows stator result of the addition of the stator to the rotor currents, as
and rotor voltage equations (1) to (4) in PSIM. indicated in the Fig. 4.
The saturation of the air-gap magnetizing inductance (M) is
considered, considering that the machine is magnetically and
electrically symmetrical, and hence the saturation, caused by
the air-gap flux, affects equally all windings and is
independent on the direction of the flux at any instant,
depending only on its magnitude [2]. The magnetizing current
space-vector, as defined in following equation (6), is the
quantity which determines the saturation degree of the air-gap
magnetizing inductance. Magnetizing current (𝐼𝐼𝐼𝐼) is the
Fig. 4. PSIM model of stator and rotor equations phasor sum of three phase currents flowing through
magnetizing branch in Fig. 4.
𝑖𝑖2𝜋𝜋 𝑖𝑖2𝜋𝜋
2
The dependent voltage source components are modeled as 𝐼𝐼𝑚𝑚 = (𝑖𝑖𝑎𝑎𝑎𝑎 + 𝑖𝑖𝑏𝑏𝑏𝑏 𝑒𝑒 3 + 𝑖𝑖𝑐𝑐𝑐𝑐 𝑒𝑒 − 3 ) (6)
3
in the Fig 5. The equation (5) for electromechanical torque is
given in the Fig. 6. Where capacitor 𝐶𝐶𝐶𝐶 represents rotor Fig. 7 represents simulation circuit to calculate magnetizing
moment of inertia constant (J), resistance 𝑅𝑅𝑅𝑅 represents inductance of machine.
inverse of rotational viscous friction constant (1/B), and
constant voltage source on the right represents rotor speed.
Mechanical load torque at the motor shaft (TL) in case of
generator would be zero so it is not included in the model.
 J = 0.002kg ∙ 𝑚𝑚2 J = 0.002kg ∙ 𝑚𝑚2
Excitation capacitor = Excitation capacitor =
340𝜇𝜇𝜇𝜇 (star connection) 500𝜇𝜇𝜇𝜇 (star connection)
Voltage = 208V rms Voltage = 208V rms
(line -line) (line -line)

Two induction machines parameters are tabulated as below

[4]. The magnetizing inductance M is provided in following
graph. This graph is plotted by performing various tests on an
induction machine.

Magnetization curve of a induction

machine
Fig. 7. Magnetizing current and magnetizing inductance 0.5

Magnetizing inductance (H)

calculation circuit 0.45
0.4
A graph showing relationship between Im_rms and M is 0.35
used in this research to evaluate value of M at different value 0.3
of current. This graph will be given in the section III. In 0.25
𝑑𝑑𝑑𝑑 0.2
equation (2) and (4) [𝐼𝐼𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎 ] represents emf induced due to 0.15
𝑑𝑑𝑑𝑑
current flowing through magnetizing inductance (M) which is 0.1
0.05
represented in the following model.
0
0 1 2 3 4
Magnetizing current (A)

Fig. 9. Magnetizing current (𝐼𝐼𝐼𝐼) vs magnetizing inductance (M)

Magnetizing curve of a induction generator

and V-I curve of capacitor
Voltage p.u. 500uF

0.8
0.7
Terminal voltage p.u.

0.6
0.5
0.4
Fig. 8. Circuit describing emf induced due to current flowing 0.3
through magnetizing inductance (M) 0.2
0.1
0
III. PSIM SIMULATION 0 1 2 3 4

To validate the model developed above, real induction Magnetizing current (A)
machine parameters are collected and used for simulation.
Fig. 10. Magnetizing curve of an induction generator with V-I
Table I Induction machine ratings and parameters curve of capacitor
Generator 1 (G1) Generator 2 (G2) Fig. 10 shows how terminal voltage of the induction
Power = 0.5kW Power = 0.5kW generator increases as magnetizing current increases. The V-I
𝑅𝑅𝑠𝑠 = 3.2Ω 𝑅𝑅𝑠𝑠 = 4.2Ω graph for different capacitor ratings is also included in this
𝐿𝐿𝑙𝑙𝑙𝑙 = 15𝑚𝑚𝑚𝑚 𝐿𝐿𝑙𝑙𝑙𝑙 = 22𝑚𝑚𝑚𝑚 figure which explains what size of capacitor should be
𝑅𝑅𝑟𝑟 = 3.32Ω 𝑅𝑅𝑟𝑟 = 4.32Ω connected across the generator in order to achieve rated
𝐿𝐿𝑙𝑙𝑙𝑙 = 18𝑚𝑚𝑚𝑚 𝐿𝐿𝑙𝑙𝑙𝑙 = 22𝑚𝑚𝑚𝑚 voltage. Data from Table I, Fig. 9 and Fig. 10 are used in the
P=4 P=4 simulation model explained in the section II above. The
following sections describe how the induction generator
behaves during isolated condition serving loads, during
parallel operation with another induction generator, and
during parallel operation with the grid.
A) Isolated Induction Generator

The Fig. 11 is the output voltage at the terminal of induction

generator. The self-excitation of the machine as an induction
generator was simulated using a Y-connected 340μF capacitor
bank and three sinusoidal-voltage sources, which were Fig. 13. Two identical generators in parallel. G1 running, load 1
disconnected at 15ms, to represent the residual flux effects connected at 1s, G2 with no terminal voltage connected at 2 s, and
during the self-excitation process. The self-excitation was load 2 connected at 3s.
carried out setting the mechanical load equivalent to 1800
rpm.

Fig. 11. Voltage build up at the SEIG terminal without electrical

load connected Fig. 14. Two generators both at the same voltage level. G1
running, load 1 connected at 1s load 2 connected at 2s, G2
connects at 4s.

Fig. 12. Output voltage as the load is connected at t= 2s, showing a (a)
voltage dip

B) Two SEIG in Parallel Operation

Two identical SEIG at different voltage level The voltage

gas at 2s is due to voltage difference between two generators
G1 and G2 at same voltage level. As given in the Fig. 14,
because of the loads 1 and 2 connected across G1 the terminal
voltage has some dips. This is due to the fact that only G1 does
not have enough power to drive both the loads, as the G2
comes online at time 4s the terminal voltage goes back to
original position. Adding G2 with G1 provides additional
(b)
power required by increased load. This behavior of SEIG can
be explained in detail with the following figure Fig. 15 (a) and Fig. 15 Two identical SEIG, (a) G1 running, L1 connected at 1s,
(b). load 2 connected at 2s, G2 comes online at 4s; (b) G1 running, G2
with equal terminal voltage connected at 4s, load 1 connected at
5s, load 2 connected at 6s. Small or no voltage dip
C) SEIG and Grid Parallel Operation
Another important behavior of induction generator that needs
be analyzed would be how SEIG reacts with load change
while it is operating in parallel with the grid. The following
figure explains how an induction generator reacts with load
change when it is connected with the grid. Grid voltage is 208
V rms (line – line).
(a)

Fig. 16. SEIG and grid: G1 running, grid connected at 2s, load 1
connected at 3s, G2 connected at 4s, load 2 connected at 5s.
(b)
Fig. 17. (a) Simulink model of stator and rotor equations (b) Flux
From the Fig. 16 it is clear that SEIG follows grid voltage and calculation circuit
frequency. No additional capacitor needed for G2 because
grid provides reactive power required. G1 and G2 do not
affect grid voltage or frequency and they are not affected by
load change. These simulation results match with the results
presented in the reference [5].

IV. MATLAB/SIMULINK MODEL OF SEIG

The SEIG is also modeled using Matlab/Simulink using

the same αβγ-three-axis modeling technique. The same
parameters used in PSIM are used in the Matlab/Simulink (a)
model. This section describes how the mathematical
equations described in theory are implemented in
Matlab/Simulink to get final SEIG model. The following
figure shows stator and rotor voltage equations (1) to (4) in
Simulink. Figure 18 (b) shows that magnetizing current 𝐼𝐼𝐼𝐼 is
connected to low pass filter it is because 𝐼𝐼𝐼𝐼𝐼𝐼, 𝐼𝐼𝐼𝐼𝐼𝐼, 𝐼𝐼𝐼𝐼𝐼𝐼 in the
equation (6) are instantaneous values which might have some
high frequency noise (spikes) which can make 𝑑𝑑𝑑𝑑/𝑑𝑑𝑑𝑑 very
large, resulting in possible modeling ill performance,
dependent on the simulation step-size.. In other words,
derivative of a signal that has high frequency (or low time)
could be "infinite". Adding a low-pass-filter before and after (b)
M eliminates such a singularity in the model. Fig. 17 (b)
represents dependent voltage source which computes flux Fig. 18. (a) Torque calculation circuit (b) Magnetization current
and magnetizing inductance computation circuit
linkage in the equation (3). The equation (5) and (6) for
electromechanical torque and magnetization current is given
in the Fig. 18 (a) and (b) respectively.
 its terminal voltage. Effect of load change on terminal voltage
were analyzed in depth in PSIM. These results were compared
with the experimental and simulation results presented in the
literature and it is concluded that the SEIG model developed
in this paper works as expected. It is also expected that this
PSIM model will be a learning tool for undergraduate
electrical engineering students who would want to understand
dynamics of an induction machine. The second part of this
research is to develop a simulation model using
Matlab/Simulink using the same αβγ-three-axis modeling
technique. The section IV shows how SEIG model can be
implemented in Matlab. Standalone mode of operation of the
generator is analyzed and compared with PSIM results. The
Fig. 19. Circuit describing emf induced due to current flowing amplitude of terminal voltage is close to that was obtained in
through magnetizing inductance (M) PSIM however, the output frequency is not as expected.
Because of the time constraint, different modes of operation
Self-excited induction generator was simulated using the of SEIG in Simulink are not studied in this paper. It is
parameters described in the section III above. Following expected that once the students understand the theory of αβγ
figure shows terminal voltage of the generator. Grid was modeling and PSIM simulation model they would be able to
connected for 0.2 s to represent residual magnetic flux and improve the Matlab/Simulink model to get results for
400uF capacitor bank is connected in parallel with the stator different modes of operation of SEIG.
terminal of the machine which would supply reactive power
required to for the machine as well as for the load connected.
As shown in the Fig. 20, the amplitude of output voltage of
the generator from Simulink model matches results from REFERENCES
PSIM model however the frequency is lower than that from [1] M. Godoy Simões, F. A. Farret; Modeling Power
PSIM. Electronics and Interfacing Energy Conversion Systems,
IEEE/Wiley. ISBN: 978-1-119-05826-7
[2] E. G. Marra and J. A. Pomilio, "Self-excited induction
generator controlled by a VS-PWM bidirectional converter
for rural applications," in IEEE Transactions on Industry
Applications, vol. 35, no. 4, pp. 877-883, July-Aug. 1999.
[3] B. Palle, M. G. Simoes and F.A. Farret, "Dynamic
simulation and analysis of parallel self-excited induction
generators for islanded wind farm systems," in IEEE
Transactions on Industry Applications, vol. 41, no. 4, pp.
1099-1106, July-Aug. 2005.
[4] E. Suarez and G. Bortolotto, "Voltage-frequency control
Fig. 20. Output voltage at no load. Grid disconnected at 0.2 s. The
of a self-excited induction generator," in IEEE Transactions
cap-bank of size 400uF provides necessary reactive power to
on Energy Conversion, vol. 14, no. 3, pp. 394-401, Sept. 1999.
maintain voltage.
[5] M. Godoy Simões, F. A. Farret, Modeling and Analysis
with Induction Generators – 3rd Edition, Taylor and Francis,
V. CONCLUSION CRC Press, ISBN 978-114-8224467-0

A simulation model of the SEIG is developed using the

well-known same αβγ-three-axis modeling technique. PSIM
and Matlab/Simulink simulation software have been used for
simulation. The parameters for the machine are used from real
lab test data of an induction machine. The results from PSIM
model are the same as expected. Three separate modes of
operation of SEIG have been analyzed in PSIM: stand alone
mode, two SEIG in parallel, and SEIG and grid operation.
Simulation results presented above show that SEIG takes
reactive power either from cap-bank or from grid to maintain