Electrical Machines

LABORATORY T3

Three-phase Induction Generator

Stand-Alone and grid-connected operations

2021/2022

Prof. João Fernandes and Prof. Paulo Branco 1

1. THREE-PHASE INDUCTION GENERATOR

The purpose of this experiment is to verify the operation of the induction machine as a generator
when connected to the electric grid (with a fixed voltage and frequency) and when operated stand-alone
(no grid connection).
This guide is divided into three parts:
a) Explanation of the generator principle of three-phase induction machine in off-grid and on-
grid applications.
b) Practical application of induction generators for a household.
c) Experimental tests.

2. GENERATOR PRINCIPLE OF THREE-PHASE INDUCTION MACHINES

2.1. GRID-CONNECTED (ON-GRID)

When connected to an electrical machine, an induction machine can operate in generator mode when
its rotational speed exceeds the synchronous speed. Figure 1 shows the torque-speed characteristics of
the laboratory induction machine (2.2kW, 400V, 50Hz, 1420 rpm) when operating with a slip between
-1 and 1 ([-1 0] is generator mode and [0 1] is motor mode). As it can be seen, the torque becomes
negative when the rotor speed is higher than the synchronous speed of 1500 rpm. The torque-speed
characteristic is not exactly symmetric because only the active part of the current is reversed. The
reactive part of the current remains positive due to the need for magnetic energy to create the rotating
magnetic field.

Figure 1 – Torque-speed characteristic of Induction Machine: motor and generator mode.

Prof. João Fernandes and Prof. Paulo Branco 2

 To understand this behavior, the machine's active and reactive power and current for the motor-
generator modes are shown in Figure 2. The active power, shown in black in Figure 2(a), shows similar
behavior to the torque, reverting its signal when the rotor speed exceeds the synchronous speed.
However, the reactive power is always positive, i.e., it is always being consumed by the induction
machine. This is because to create the rotating magnetic field; magnetic energy is always required
(remember that 𝑄 = 𝜔𝐿𝑀𝑠 𝐼𝑀𝑒𝑓 2 = 𝑈𝑀𝑠𝑒𝑓 2 /(𝜔𝐿𝑀𝑠 )). The magnetic energy required by the induction
machine is mostly given due to the magnetizing induction coefficient, 𝐿𝑀𝑠 , which may increase or
decrease depending on the level of saturation of the magnetic core (remember Lab1A).
In Figure 2, in blue are marked the limits of the safety operation zone of this machine, given by the
limit of stator current allowed in the machine (rated current of 4.7A).

(a) (b)
Figure 2 – (a) Active power and reactive power VS speed characteristic and (b) current-speed
characteristic of Induction Machine in motor and generator mode.

In Figure 3 are shown the active and reactive power and 1current for the allowed working zone. The
reactive power and reactive stator current remain positive and almost constant for both motor and
generator modes. This means that the motor and generator modes are not completely symmetric from a
current point of view. The active power and current are almost linear with the speed and change signal
when the speed is higher than the synchronous speed.

1 Remember that:
𝑃𝑆 = 𝑈𝑒𝑓 𝐼𝑒𝑓 cos(𝜑) = 𝑈𝑒𝑓 𝐼𝑎 → 𝐼𝑎 = 𝐼𝑒𝑓 cos(𝜑) (active current)
𝑄𝑆 = 𝑈𝑒𝑓 𝐼𝑒𝑓 sin(𝜑) = 𝑈𝑒𝑓 𝐼𝑞 → 𝐼𝑞 = 𝐼𝑒𝑓 sin(𝜑) (reactive current)

Prof. João Fernandes and Prof. Paulo Branco 3

 (a) (b)
Figure 3 – Zoom of (a) the active power and reactive power VS speed characteristic and (b) active
and reactive stator current for the allowed working zone (current lower than the rated one).

To facilitate understanding this behavior, Figure 4 presents the power flow for both motor and
generator modes. Quantities Ps and Qs are the stator active and reactive power, PRs and PRr are the stator
and rotor Joule losses, Pmec is the mechanical power, Qσs and Qσr are the reactive power consumed by
the stator and rotor flux leakages, and QMs is the reactive power consumed to create the rotating magnetic
field. Note that the reactive power flow remains the same between the two modes, while the active power
flow is symmetric. Therefore, with an induction machine, it is possible to be operating in the generator
mode (PS<0), but it is still required to consume reactive power (QS>0).

PRs Qσr PRr

Ps
Qs
QMs
Qσs
(1− s ) R 'r Pmec
s

(a)

PRs Qσr PRr

Ps
Qs
QMs
Qσs
(1− s ) R 'r Pmec
s

(b)
Figure 4 – Power flow in (a) motor mode and (b) generator mode.

Prof. João Fernandes and Prof. Paulo Branco 4

 2.2. STAND-ALONE APPLICATIONS
In stand-alone applications, there is no access to the electric grid.

Therefore, how can we supply the reactive power to the induction machine required to create the
rotating magnetic field?

One possible solution is to install a three-phase capacitor bank, C, at the stator terminals of the
induction machine to supply the required reactive power. This solution is called Self-Excited Induction
Generator2 (SEIG), Figure 5.
The capacitor has the function to supply the reactive power required by the induction machine and
the load. All active power generated by the induction machine is supplied to the load. However, note
that no electric voltage and frequency are applied to the induction machine; only the rotor speed (slip s)
is applied. So, if no voltage is applied to the system, the capacitors cannot supply the reactive power,
and, therefore, no rotating magnetic field is originated in the SEIG. Then, how do the SEIG works?
Due to asymmetries in the magnetic reluctances of the induction machine with the rotor’s
position, a small stator voltage is induced, US0, even without any current injected into the
machine, when the rotor is rotating. The capacitor uses this initial voltage to create an initial
current for the induction machine. This process will cascade until stability is reached. This is
going to be explained later.

Ps=PL Ps
PRs Qσr PRr
QL Qs

Qc QMs
Qσs
RL C (1− s ) R 'r Pmec
load s

Figure 5 – Self-excited induction generator (SEIG) scheme.

Let us consider the SEIG working without load at a constant speed in Figure 6. If no load is
connected to the SEIG stator terminals, then the mechanical power supplied to the SEIG is equal to the
SEIG losses, Pmec = PRs + PRr (let us neglect the iron losses to now simplify the explanation). Considering

2 For more information you can check the following paper: https://doi.org/10.1016/j.enconman.2019.01.016

Prof. João Fernandes and Prof. Paulo Branco 5

no load and the leakage induction coefficients are very low compared to the magnetizing one, then Qc
≈ QMs and Ic ≈ IM.

Ps=0 PRs Qσr PRr

Qs

Qc QMs
Qσs
C (1− s ) R 'r Pmec
s

Figure 6 – Self-excited induction generator (SEIG) power flow at no load.

If IC ≈ IM and if we neglect the voltage drop in the stator impedance (𝑈𝑠 ≈ 𝑈𝑚 → the capacitor is in
parallel to 𝐿𝑀𝑠 ), then

𝑈𝑠𝑒𝑓 ≈ 𝑈𝑀𝑠𝑒𝑓 = 𝜔𝐿𝑀𝑠 𝐼𝑀𝑒𝑓 ≈ 𝜔𝐿𝑀𝑠 𝐼𝐶𝑒𝑓

𝐼𝐶𝑒𝑓
𝑈𝑠𝑒𝑓 = 𝑈𝐶𝑒𝑓 =
𝜔𝐶

As can be seen, the stator voltage depends on the magnetizing inductance and the capacitor
value. There is a linear relationship between the capacitor voltage and current. However, the
magnetizing inductance 𝐿𝑀𝑠 changes with the level of saturation of the magnetic machine core
(when it reaches saturation, its value decreases, remember Lab1A).

Finally, Figure 7 is shown the excitation process of the SEIG, with the magnetizing curve
given by 𝑈𝑠𝑒𝑓 ≈ 𝜔𝐿𝑀𝑠 𝐼𝐶𝑒𝑓 and the capacitance curve is given by 𝑈𝑠𝑒𝑓 = 𝐼𝐶𝑒𝑓 /𝜔𝐶. The process
starts by applying the small-induced machine voltage (US0) to the capacitor. Due to the small
voltage, the capacitor will produce a small reactive current, IC1, injected into the SEIG, which
will amplify the initial voltage to a new voltage US1. The new voltage US1 is now applied to the
capacitor, which increases the capacitor current, IS2. This process repeats itself until the two
curves intersect (in point O). This process is called voltage build-up.

Prof. João Fernandes and Prof. Paulo Branco 6

 s) O
LM
US e(
rv
Cu
g
in
tiz
ne
ag ve
M ur
cec
a n
a cit
p
US2 Ca
US1
US0

Start IC1 IC2 IC ≈ IMs

Figure 7 – Self-excited induction generator (SEIG) excitation process.

Of course, different capacitors will impose different stable-operation points. In Figure 8 is shown
the operation points (O1, O2, and O3) for different capacitors (C1, C2, and C3, respectively). Capacitor
C3 is not capable of exciting the SEIG with proper values of voltages, O3. However, for higher values
of capacitors, like C1, the induction machine may reach the magnetic saturation zone (flat zone in the
magnetizing curve), which will induce high current values, damaging the machine. Therefore, the
selection method of the excitation capacitors is important to allow the excitation of the SEIG without
overloading (in the next figure C2 is the most appropriate).

C3 C2

O2 O1 C1
US

)
urv e (L M s
etizing C
Magn

O3 C1>C2>C3
US0

IC ≈ IMs
Figure 8 – Self-excited induction generator (SEIG) excitation process for different capacitors.

Prof. João Fernandes and Prof. Paulo Branco 7

 So how can we estimate the correct value of the capacitor?
There are several methods to do this. However, a simple one can be used to estimate this value. This
method is based on the reactive power balance between the capacitors and the induction machine: Qc
≈ QMs.

1. First, compute the magnetizing induction coefficient of the induction machine by supplying
several levels of voltages between the rated voltage UN and 20%UN at a fixed voltage. Data
from the laboratory machine:

Figure 9 – Evolution of the magnetizing inductance with the stator voltage at a fixed frequency of
50Hz.

2. Next, considering a negligible stator voltage drop (𝑈𝑠 ≈ 𝑈𝑚 ) and the reactive leakage
power Qc ≈ QMs, compute the capacitor value for the rated voltage induction coefficient 𝐿𝑀𝑠 :

𝑄𝑀𝑠 = 𝜔𝐿𝑀𝑠 𝐼𝑀𝑒𝑓 2 ≈ 𝜔𝐿𝑀𝑠 𝐼𝐶𝑒𝑓 2

1
𝐼𝐶𝑒𝑓 2 }→𝐶=
𝑄𝐶 = 𝜔2 𝐿 𝑀𝑠
𝜔𝐶

For the laboratory induction machine, 𝐿𝑀𝑠 = 210 mH, and C = 48 μF.

3. PRATICAL APPLICATIONS CASES

In this laboratory, we will study two application cases where a load representing a household
is supplied by an auxiliar induction generator coupled to a small wind turbine or a small water
turbine, Figure 10. The two cases investigated are: a) with an induction generator connected to

Prof. João Fernandes and Prof. Paulo Branco 8

the grid, Figure 11, and b) with an off-grid induction generator supplying the whole load, Figure
12.
Wind turbine

On-Grid

Prime House
Water turbine mover IM (Load)

Off-grid

Figure 10 – Case studies investigated: on-grid or off-grid.

Please note that, in the on-grid application, the active power generated by the induction
machine may not be the same required by the load. There are three possible cases of power
flow:
1. The active generating power is higher than the load, and the excess of power is
injected into the electric grid (PS > PL, Pgrid > 0);
2. The active generating power is the same as the load, and no active power is injected
or required from the electric grid (PS = PL, Pgrid = 0), and;
3. The active generating power is lower than the load, and the electric grid supplies the
remaining power (PS < PL, Pgrid < 0).

In all three cases, the electric grid always supplies the reactive power required by the
induction machine.
House (Load)
PS On-grid Pgrid
Electric
Prime
mover IM Grid
400V, 50Hz
QS PL QS
RL
Representative
resistive load

Figure 11 – Electrical wiring and power flow of the on-grid case study investigated.

The induction generator must always supply the load for the off-grid case study because no
other power sources are present. In fact, if the prime mover does not have sufficient power to

Prof. João Fernandes and Prof. Paulo Branco 9

supply the active power required by the load instantaneously, it will decelerate, reducing the
frequency and voltage of the system. If the load is of an electronic characteristic, it may be
disconnected for safety reasons.

House (Load)
PS Off-grid
Prime
mover IM
PL
QS RL
C
Excitation Representative
capacitors resistive load

Figure 12 – Electrical wiring and power flow of the off-grid case study investigated.

4. CONDUCTING THE EXPERIMENT

The previous on-grid and off-grid applications will be tested in the laboratory, where a DC
motor will be used to simulate a wind or water turbine. In Figure 13 is shown a photo of the
laboratorial setup, with the induction machine, the DC machine and the resistive load, and the
electric grid for the on-grid case and the capacitors for the off-grid case.

Figure 13 – Laboratory setup.

4.1. ON-GRID OPERATION

In this experiment, the induction machine is connected directly to the electric grid. The
induction machine is mechanically coupled to a DC machine, as shown in Figure 14. The DC

Prof. João Fernandes and Prof. Paulo Branco 10

machine will operate as a motor, supplied by a rectifier connected to an autotransformer (dotted
connection).

House (Load)
Ia On-grid Igrid
 Electric
Ua DC IM Grid
= IL
400V, 50Hz

Prime mover
RL

Figure 14 – On-grid laboratory wiring diagram.

a) Supply only the excitation circuit of the DC machine (armature circuit not connected)
and start the induction motor. Verify the direction of rotation and the signal of the
induced voltage in the DC armature circuit.
b) Stop the Induction machine by disconnecting it from the electric grid. Supply the DC
motor with the rated excitation current (When operating in motor mode, you can
NEVER disconnect the excitation circuit of the DC motor). Then, supply the DC
motor armature circuit with a small voltage and check the direction of its rotation and
the signal of the applied armature voltage. If any direction is opposite from step 1,
check it with your professor.
c) Now reduce the DC machine armature voltage to zero, keeping the excitation circuit
connected. Plug the induction machine into the grid to start it at no-load. Check if the
induction machine current is around 3.0A. After achieving the steady-state, act upon the
DC motor by SLOWLY increasing its voltage to bring the induction machine speed
higher than the synchronous speed. Stop when the DC machine armature current reaches
Ia = 12A. Keep the induction machine stator current always lower than 4.7A.
d) Set the load resistances to RL = 200 Ω. Adjust the DC motor armature voltage to reach
an armature current of Ia = 12A (in this case PS > PL, Pgrid > 0). Register the armature
voltage and current values, Ua e Ia, and the excitation current, Iexc, in the DC motor, the
rotation speed, Nr, and one stator phase current and voltage, IS and US, the active, PS,
and reactive, QS, power in the induction generator and the resistive load current, IL.

Prof. João Fernandes and Prof. Paulo Branco 11

 e) Repeat the previous experiment with RL = 100 Ω and RL = 600 Ω. This simulates an
increase and decrease of load that is typical during a typical household day. Repeat
the measurements.
f) Repeat the previous experiments with a DC motor armature current of Ia = 7A, for RL =
[100 200 600] Ω. Repeat the measurements.

4.2. OFF-GRID OPERATION

In this experiment, the induction machine is operated off-grid to supply the load. The
induction machine is connected to a capacitor bank to supply the required reactive power and a
resistive load, Figure 15. The prime mover is the same as the previous experiment: a DC motor
supplied by a rectifier connected to an autotransformer and the electric grid. Note that there is
no electric grid in the household.

Prime mover House (Load)

Ia IS Off-grid
 UFF
Ua DC IM
=
RL
C

Electric
Grid
400V, 50Hz

Figure 15 - Induction machine connection diagram.

a) Set the value of the capacitor’s bank to zero, 𝐶 = 0, and the resistive load to 𝑅𝐿 = ∞
(no-load).
b) Supply the excitation circuit of the DC motor with the rated excitation current. Act upon
the DC motor by increasing its armature voltage to bring the induction machine speed
to near 1500 rpm. Keep the induction machine stator current lower than 4.7A.
c) In the oscilloscope, register the stator voltage and current evolution during the voltage
build-up process without resistive load. To do this, change the value of the capacitors
to C = 48 μF and visualize the voltage build-up process in the oscilloscope and save
their waveforms in a USB pen.

Prof. João Fernandes and Prof. Paulo Branco 12

d) Set the load resistances to RL = 200 Ω. Adjust the armature voltage of the DC motor to
get an electric frequency of 50Hz at the induction machine. Register the armature
voltage and current values, Ua e Ia, and the excitation current, Iexc, in the DC motor, the
rotation speed, Nr, and one stator phase current and voltage, IS and US, in the induction
generator. Also, visualize the induction machine stator current and voltage in the
oscilloscope and save their waveforms in a USB pen.
e) Now set the load resistances to RL = 150 Ω (this simulates an increase of load) and RL =
300 Ω (this simulates a decrease of load) and repeat the measurement without changing
the armature voltage of the DC machine. What happens to the frequency?
f) With RL = 150 Ω and RL = 300 Ω, adjust the frequency for 50Hz and repeat the
measurements.

Prof. João Fernandes and Prof. Paulo Branco 13