Fuzzy Logic Control of an Induction Generator as an Electrical Brake

N. Hosseinzadeh

D. Seyoum
Central Queensland University
Rockhampton, Australia
[email protected]
Fax: +61 3 92148264

P. Wolfs

ABSTRACT
Fuzzy logic systems (FLS) have been designed to control a self-excited induction generator (SEIG),
which is used as a brake. This electrical brake has been designed for the sugar cane industry in
Queensland, Australia, which uses brake vans coupled to the end of cane trains to produce a given braking
torque. The brake would be suitable for similar applications of electrical brakes in electrically driven
machines such as electrical vehicles. This project was established to investigate electrical braking as an
alternative to existing mechanical systems.
Three fuzzy logic controllers have been designed to control the output retarding torque produced by an
induction machine. One of these controllers adjusts the value of a shunt capacitance to maintain the
excitation required for the generating operation. The other two adjust the duty cycle of a PWM converter,
which drives the load of the induction machine. The duty cycle is adjusted in such a way to keep the
retarding torque, produced by the machine, fixed at a given value.
Keywords:
Fuzzy Logic Applications; Induction Generators; Intelligent Control; Electrical Machines and Drives
INTRODUCTION
Electrical braking using a self-excited induction
generator (SEIG) [1] can provide a maintenance
advantage compared with the mechanical
braking systems. The implementation of an
SEIG as a means of braking will improve the
existing mechanical braking system for the
trains that are specifically used for the sugar
cane industry in Queensland, Australia.
However, several important issues must be
addressed in the design of the electrical braking
system [2].
One of these problems is that the output voltage
of the induction generator depends greatly on
the speed and load of the train; this will cause a
significant variation in the power consumption
in the load of the machine during braking.
Therefore, the power dissipated in the load must
be adjusted with speed so that a constant
braking torque is generated.
Another problem is related to the excitation
requirements of the SEIG. In order to provide a
source of reactive power for the excitation of an
induction generator, capacitor banks may be
connected across the stator terminals [3]. These
capacitors provide the means for initiating the
self excitation of the machine. For a given
capacitor value, self-excitation can only be
achieved and maintained under certain load and
speed combinations. Therefore, the capacitor

bank has to be dimensioned according to a

defined narrow range of speed and load values
[4]. However, when the speed falls below or the
load rises beyond certain values, the machine
demagnetizes and stops generating. This means
that although the capacitive self-excitation of an
induction machine is a simple and viable
solution, an additional mechanism has to be
added in order to avoid demagnetization and to
accept the full scale of the output voltage [2].
Fuzzy logic controllers will be used in solving
the problems outlined in the previous
paragraphs. Since the time that Lotfi A. Zadeh
proposed a new approach to the analysis of
complex systems and decision processes in 1973
[5], fuzzy logic has been applied in many areas
of science and engineering. However, fuzzy
logic actually gained popularity when it was
applied to industrial problems by engineers [6].
This popularity can be attributed to the fact that
fuzzy logic provides a powerful vehicle that
allows engineers to incorporate human
reasoning in the control algorithm. In contrast
with the conventional control theory, fuzzy
logic design is not solely based on a
mathematical model of the plant; although
having an approximate mathematical model
helps in the fine tuning of a fuzzy logic
controller. A fuzzy logic controller implements
human reasoning that has been programmed into
fuzzy logic language (fuzzification, membership
functions, rules and defuzzification) for
controlling a plant [7]-[9].

This paper presents fuzzy logic systems (FLS)

to adjust the values of load and capacitance with
variations of the generator shaft speed. The
control scheme is designed to keep the output
braking torque fixed and to maintain the
magnetic field required for the operation of the
SEIG.
FUZZY LOGIC CONTROL FOR AN
INDUCTION GENERATOR
As mentioned earlier, the sugar cane industry in
Australia uses brake vans, coupled to the end of

cane trains, to produce a given constant braking

torque. Currently, brake vans operate using
compressed air, which is supplied by an onboard
compressor. The compressor activates a brake
calliper, which clamps a ventilated disc rotor on
each of the four wheel sets. This system needs
maintenance due to the wear on the brake pads
and rotors. Electrical braking can be an
alternative to this mechanical system. With the
application of electrical braking, the kinetic
energy is converted to electrical energy then
dissipated as heat energy in well ventilated
resistors as shown in Figure 1 [2].

Belt or Gear
Coupling

RL

Van wheel

RL

RL

Induction
generator

Figure 1

Electrical braking system using an induction generator (adapted from [2])

Both RL and C must be adjusted with the change

in the shaft speed of the induction generator. In
this paper, it is proposed that RL be controlled

Set point Retarding Torque (Tref)

by a Mamdani-style FLS [10] and C be adjusted

by a Sugeno-type FLS [11]. The block diagram
of the control scheme is shown in Figure 2.

Tdev = Tref - Tmeas

Measured Torque (Tmeasured) -

Fuzzy Logic
System
(Mamdani)

Measured Current (I)

Voltage (V)
Speed ()

Figure 2

Torque
Calculation

Self-Excited
Induction
Generator

Fuzzy Logic
System (Sugeno)

Control scheme of a self-excited induction generator using two fuzzy logic systems

A comparator obtains the deviation of the

measured torque from a fixed set point for the
retarding torque (i.e. Terror = Tref Tmeasured).

Two samples of Terror are used to calculate its

rate of change. Then, Terror and its rate of change
are used as two inputs to a Mamdani-type FLS,

which provides a duty cycle k as its output in

order to compensate for changes in the torque
value during transient periods. This duty cycle is
added to the output of another fuzzy logic
controller, which provides a steady-state value
for the duty cycle based on the operational
speed of the machine. The combined steadystate and transient duty cycles are applied to a
Pulse Width Modulator (PWM) converter,
which is applied to a power electronic switch
that adjusts the load on the induction generator
(i.e. RL). Also, a Sugeno FLS uses the value of
shaft speed to provide a suitable value for C,
which is used to keep the magnetic field
required for the operation of the induction
generator.
Different types of fuzzy logic systems have
been used in this application because of
different requirements, which are associated
with the calculations of k and C. The FLS for
obtaining k needs a continuous-type control,
which is easily obtained with a Mamdani style
FLS with seven membership functions for its
output. On the other hand, the FLS that is used
to obtain C requires a system that imitates a
look-up table. A Sugeno style with more
singleton functions defined for the output C, in
this case 16 singletons, is more suitable for the

Figure 3

calculation of C. Sixteen singletons have been

used in this application because four deltaconnected three-phase capacitors have been
used, which will be controlled by a four-digit
binary switch. The switch has 16 different
states.
SIMULATION RESULTS
A mathematical model of an induction generator
[2] has been used to validate the performance of
the fuzzy logic control scheme proposed in this
paper. A block diagram is provided in the
appendix, which shows a MATLAB Simulink
model. It includes the mathematical model of
the induction generator, two Mamdani-type
fuzzy logic controllers that control the load on
the induction machine and a Sugeno-type fuzzy
logic controller that provides suitable values of
a capacitor bank, which is connected in parallel
to the terminals of the generator.
Figure 3 shows the voltage build-up process of
the generator, when its shaft speed is kept
constant at 1250 rpm. The shunt capacitor is
initially charged to 50 volts, which excites the
generator to start the voltage build-up process.

The initial voltage build up, with a charged capacitor and at a constant speed

In order to test the performance of the proposed

control scheme, after the voltage is built up, the
machine load is increased at t = 25 sec. Then, at

t = 40 sec, the speed of the machine is gradually

changed (firstly decreases, down to about 1107
rpm, and then increases) as shown in Figure 4.

Figure 4

A change in the machine shaft speed occurring at t = 40 sec

If the duty cycle is kept at a fixed value (i.e. an

uncontrolled load), the retarding torque that is
produced by the induction machine will also
change, as shown in Figure 5.

Figure 5

Retarding torque changes as a result of the change in speed

The purpose of the Mamdani fuzzy logic

controller, which changes the duty cycle of the
PWM converter, is to keep the retarding torque
generated by the machine constant in spite of
changes in the shaft speed of the machine,

which comes from the speed changes of the

train. Figure 6 shows the output of the fuzzy
logic controller (the duty cycle applied to the
PWM converter).

Figure 6

The fuzzy logic controller adjusts the duty cycle to maintain a constant torque

As seen from the figure, the duty cycle increases

when the speed decreases and also when Terror
has a positive value (which means the produced
torque is less than a set-point torque). This will
increase the power consumption in the load and
will prevent the retarding torque from
decreasing. As Terror comes down to a value
close to zero and does not change any more, the
duty cycle settles down to a steady state value
which is correspondent to a normal load (a
specific value for RL, decided by the designer of
the system according to the specifications of the
machine for any specific speed). In other words,
the duty cycle has two components. The first

Figure 7

component is a steady state value for a given

speed. The second component is a value for
transient periods and is controlled by the torque
error and its rate of change. The latter will settle
down to zero when Terror settles down to zero.
As a result of the controller action, the torque
remains fairly constant in spite of the changes in
speed. This is demonstrated in Figure 7, which
corresponds to the speed change of Figure 4 and
control action of Figure 6. Note that the initial
change in the torque value (prior to t = 40 sec) is
during the voltage build-up process.

Torque output is kept fairly constant after the initial voltage build up

Comparing Figure 7 with Figure 5 demonstrates

the improvement made by the fuzzy logic

control in providing a fixed braking torque for

the brake van application.

CONCLUSIONS
Three fuzzy logic controllers have been
designed to control the output retarding torque
produced by an induction machine, operating in
the generating mode. One of these controllers,
which is of a Sugeno type, adjusts the value of a
capacitor bank in order to keep the required
excitation for the machine.
The other two controllers provide the duty cycle
required for a PWM converter, which drives the
load on the machine. One of these controllers
adjusts the value of the duty cycle according to

changes in the speed of the machine and the

other one responds to changes in the torque
level. The combined duty cycle adjusts the level
of power consumption to keep the produced
retarding torque as close as possible to a setpoint value.
The simulation results are promising as was
demonstrated in the paper. The fuzzy logic
controllers are going to be implemented on a
digital signal processing (DSP) board and the
authors are arranging performance tests on a
physical machine.

REFERENCES
[1]

[2]

Elder, J. M., Boys, J. T. and Woodward,

J. L., Self-excited induction machine as
a low-cost generator, IEE Proceedings
Part C, vol. 131, no. 2, 1984, pp. 33-41
Seyoum, D. and Wolfs, P., Self excited
induction generators for brake van
applications, The Proceedings of the
Australasian Universities Power
Engineering Conference (AUPEC),
Brisbane, Australia, September 2004

[3]

Basset, E.D. and Potter, F.M.,

Capacitive excitation of induction
generators, Transactions of the
American Institute of Electrical
Engineers, vol. 54, 1935, pp. 540-545

[4]

Shridhar, L., singh, B., Jha, C. S., Singh,

B. P. and Murthy, S. M., Selection of
capacitance for the self regulated short
shunt self excited induction generator,
IEEE Transactions on energy
Conversion, vol. 10, no. 1, March 1995,
pp. 10-17

[5]

Zadeh, L. A.: Outline of a new approach

to the analysis of complex systems and
decision processes, IEEE Trans. on
Systems, Man and Cybernetics, vol.
SMC-3, no. 1, pp. 28-44, 1973

[6]

Mamdani, E. H. and Assilian, S.: An

experiment in linguistic synthesis with a
fuzzy logic controller, International
Journal of Man-Machine Studies, vol. 7,
1975, pp 1-13

[7]

Yager, R. R. and Filev, D. P.: Essentials

of fuzzy modelling and control, John
Wiley&Sons, 1994

[8]

Mendel, J., M., Fuzzy logic systems for

engineering: a tutorial, Proceedings of
the IEEE, vol. 83, no. 3, March 1995, pp.
345-377

[9]

Tsoukalas, L. H. and Uhrig, R. E.,

Fuzzy and neural approaches in
engineering, 1997

[10]

Mamdani, E. H., Ostergaard, J. J. and

Lembessis, E.: Use of fuzzy logic for
implementing rule-based control of
industrial processes, Fuzzy Sets and
Decision Analysis, Amsterdam, North
Holland, 1984, pp 429-445

[11]

Sugeno, M., Industrial applications of

fuzzy control, Fuzzy Sets and Decision
Analysis, Amsterdam, North Holland,
1985

Appendix
MATLAB Simulink model of a self-excited induction generator, which is controlled by fuzzy logic controllers