Electrical Power and Energy Systems 26 (2004) 291305

www.elsevier.com/locate/ijepes

Automatic generation control with fuzzy logic controllers

in the power system including SMES units
A. Demiroren*, E. Yesil
Department of Electrical Engineering, Electric and Electronic Faculty, Istanbul Technical University 34469, Maslak, Istanbul 80626, Turkey

Abstract
This study presents a method based on fuzzy logic controllers (FLCs) for automatic generation control (AGC) of power systems including
superconducting magnetic energy storage units. The technique is applied to control systems include three areas having two steam turbines
and one hydro turbine tied together through power lines. As a consequence of continually load variation, the frequency of the power system
changes over time. In conventional studies, frequency transients are minimized by using conventional integral and proportional controllers
aiming of secondary control in AGC and zero steady-state error is obtained after sufficient delay time. In this paper, instead of this method,
the configurations of FLCs are proposed. The results obtained by using FLCs outperform than those of conventional controllers as settling
time and overshoot as shown at simulation.
q 2004 Elsevier Ltd. All rights reserved.
Keywords: Automatic generation control; Superconducting magnetic energy storage; Power conversion system; Fuzzy logic controller

1. Introduction
The automatic generation control (AGC) problem, which
is the major requirement in parallel operation of several
interconnected systems, is one of very important subjects in
power system studies. In this study, power systems with
three areas connected through tie-lines are considered.
The perturbation of frequencies at the areas and resulting
tie-line power flows arise due to unpredictable load
variations that cause mismatch between the generated and
demanded powers. The objective of AGC is to minimize the
transient deviations and to provide zero steady state errors of
these variables in a very short time although unpredictable
load variations present. In literature, for AGC, some control
strategies based on classical control theory have been
proposed [1,2]. Unfortunately, because of operating point
continuously changes depending on demand of consumers,
the selected fixed controller can unsuitable other operating
points. Therefore, many of controller with variable structure
is proposed in literature [3,4].
The AGC based on fuzzy PI-type controller is proposed
in this study. One of its main advantages is that controller
parameters can be changed very quickly by the system
dynamics because no parameter estimation is required in
* Corresponding author. Tel.: 90-212-285-6766; fax: 90-212285-6700.
E-mail address: aysen@elk.itu.edu.tr (A. Demiroren).
0142-0615/$ - see front matter q 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.ijepes.2003.10.016

designing controller for nonlinear systems. Therefore a

fuzzy logic controller (FLC), which represents a model-free
type of nonlinear control algorithms, could be a reasonable
solution.
Fuzzy PI-type controller has some advantages: (i) it
provides an efficient way of coping with imperfect
information, secondarily, (ii) it offers flexibility in decision
making processes and thirdly, (iii) it provides an interesting
man/machine interface by simplifying rule extraction from
human experts and by allowing a simpler a posteriori
interpretation of the system reasoning [5,6]. Fuzzy logic
controls are knowledge-based controllers usually derived
from a knowledge acquisition process or automatically
synthesized from self-organizing control architectures.
These controllers typically define a nonlinear mapping
from the systems state space to the control space. Thus each
FLC can be visualized as a nonlinear control surface
reflecting a process operators or a product engineers prior
knowledge. Each control surface is represented in a
knowledge base (KB) and executed by an interpreter [7].
On the other hand, in the AGC problem, the instantaneous mismatch between supply and demand of real
power in a power system can be reduced by the addition of a
fast acting superconducting magnetic energy storage
(SMES) unit [8 10]. Thus, a SMES unit having 30 MJ
capacity cooled by liquid helium is available in Bonneville
Power Administration of USA. It is used to demonstrate

292

A. Demiroren, E. Yesil / Electrical Power and Energy Systems 26 (2004) 291305

Nomenclature
D

f
Kp
zTP
PG
PD
T1
T3
us
R
X
Fx
L
Id0
Idmin
Edmin

the deviation
derivation of variable
frequency
transfer function gain of generator
time constant of generator
fluctuation in turbine output power
electrical load variations
time constants of hydro governor
time constants of hydro governor
control input of power system
regulation parameter
state vector
nonlinear function of x
inductances of SMES unit
operating point current of SMES units
Minimum current of SMES units
Minimum voltage of SMES units

technical and economic feasibility of SMES units in power

system applications. The energy storage requirement to
damp the frequency oscillations caused by small load
perturbations is much smaller. In such cases, the real power
transfer takes place in a very short time. To store the energy
required for AGC applications batteries are used, however,
they have some disadvantages [1]. The recent developments
on ceramic superconductors with high critical temperature
facilitated their commercial availability. Thus, to construct
and operate of SMES unit becomes more economic.
Addition of a small capacity SMES unit to the system
significantly improves transients of frequency and tie-line
power deviations against to small load disturbances. In the
SMES unit, a dc magnetic coil is connected to the ac grid
through a power conversion system (PCS), which includes
an inverter/rectifier. The superconducting coil can be
charged to a set value from the grid during normal operation
of the power system. Once superconducting coil charged,
it conducts current with virtually no losses [8 10]. The coil
is maintained at extremely low temperature. While there is a
sudden rise in the demand of load, the stored energy is
almost released through the PCS to the power system as
alternative current. As the governor and other controllers
start working to set condition, the coil changes back to its
initial value of current. Similar action becomes during
sudden release of loads. In this case, the coil immediately
gets charged towards its full value, thus absorbs some
portions of the excess energy in the system, and as the
system returns to its steady state, the excess energy absorbed
is released and the coil current attains its normal value.
During normal operation, the stored energy is taken about
20 30 MJ because of factors of superconductor stability,
mechanical forces and fatigue [8]. So, SMES unit is forced
its lower and upper current limits. Therefore SMES units are

Tdc
K1 ; K2
s
KR
TR
PR
TT
XE
TG
T2
TW
KI
Ed0
Idmax
Edmax
K0
Wsm
i

converter time delay

input scaling factors of the FLC
Laplace domain derivation term
transfer function gains of reheats of areas
reheat time constants
mechanical power during steam reheat
time constant of turbine
Governor valve position
time constant of governor
time constants of hydro governor
water starting time
integral controller gains
operating point voltage of SMES units
Maximum current of SMES units
Maximum voltage of SMES units
the gain of SMES unit
energy stored in SMES unit
area index (1,,3)

separated from the system if the inductor current reaches its

limits and SMES units are again connected only when the
control signal changes its sign. Whenever SMES unit is
disconnected, it is noticed that the power system is
equivalent to the one without SMES unit.
In this paper, the power system with three areas having
two steam turbines and one hydro turbine is considered in
simulation study. In the study, the simulation is
implemented by using MATLAB Simulink Program and
MATLAB Fuzzy Logic Toolbox (FLT) [11]. To damp out
the oscillations due to instantaneous load perturbations as
fast as possible, AGC including the FLC is used.
The results obtained show that the controller improves
effectively the damping of the oscillations after the load
deviation in one of the areas in the interconnected system
compared to conventional controllers. Moreover, the
considered interconnected power system with three areas
contains SMES units in each area. In addition, during
operation, both the effect of SMES and the effect of fuzzy
are investigated together. On the other hand, a characteristic specification of this study is that the governors at all
areas in power system have deadbands, which are
important for speed control under small disturbances.
The deadband affects the stability of the power system.
It is known that governor deadband has destabilization
effect on the transient response. Moreover, reheater effects
and boiler effects are very important for the stability of
the systems having steam turbine as known in the
literature [1]. Therefore, the effects of boiler and reheater
of each thermal area in the power system are considered
in the study.
In the simulation, firstly, a step load increase in the first
area of the power system and then step load increases in
both first area and second area of the same power system are

A. Demiroren, E. Yesil / Electrical Power and Energy Systems 26 (2004) 291305

considered. For comparison, the considered power system is

controlled by using
(i) Conventional proportional integral (PI) controllers
(ii) Fuzzy PI-type controllers
for these cases. These simulations are carried out while
SMES unit is present in each area at power system and
without SMES unit in any of the areas. The simulations
show the positive effect of SMES unit on the improvement
of the oscillation of frequency due to any load perturbation.
The results obtained show that the performance of fuzzy
controller is better than conventional integral controller, as
the main objective of the study.

2. SMES system
The schematic diagram in Fig. 1 shows the
configuration of a thyristor controlled SMES unit
[8 10]. The SMES unit contains DC superconducting
coil and converter which are connected by Y D/Y Y
transformer. The control of the converter firing angle
provides the dc voltage Ed appearing across the inductor
to be continuously varying within a certain range of
positive and negative values. The inductor is initially
charged to its rated current Id0 by applying a small
positive voltage. Once the current reaches the rated
value, it is maintained constant by reducing the voltage
across the inductor to zero since the coil is
superconducting [10]. Neglecting the transformer and
the converter losses, the DC voltage is given by
Ed 2Vd0 cos a 2 2Id Rc

293

In AGC operation, the dc voltage Ed across the

superconducting inductor is continuously controlled
depending on the sensed area control error (ACE) signal.
In literatures, it is taken to be dependent on the sensed Df
signal or ACE defined as Bi Dfi DPi i 1; ; 3 [12].
The expression of DPi is given in Appendix A and Bi
parameters are chosen as 1=Kpi 1=Ri as generally
mentioned in literature [13]. In this study, as in recent
literature, inductor voltage deviation of SMES unit of
each area is based on ACE of the same area in power
system [9]. Moreover, the inductor current deviation is
used as a negative feedback signal in the SMES control
loop. So, the current variable of SMES unit is intended to
be settling to its steady state value. If the load
demand changes suddenly, the feedback provides the
prompt restoration of current. The inductor current must
be restored to its nominal value quickly after a system
disturbance, so that it can respond to the next load
disturbance immediately [8 10]. As a result, the
equations of inductor voltage deviation and current
deviation for each area in Laplace domain are as follows
DEdi s K0i

1
B Df s DPi s
1 sTdci i i

2 KIdi
DIdi s

1
DI s
1 sTdci di

1
DEdi s;
sLi

i 1; ; 3

where KIdi is the gain for feedback DIdi ; Tdci is converter

time delay, K0i (kV/unit ACE) is gain constant and Li (H)
is inductance of the coil. The deviation in the inductor
real power of SMES unit is expressed in time domain as
follows

where Ed is DC voltage applied to the inductor (kV), a is

firing angle (8), Id is current flowing through the inductor
(kA). Rc is equivalent commutating resistance (V) and
Vd0 is maximum circuit bridge voltage (kV). Charge and
discharge of SMES unit are controlled through change of
commutation angle a: If a is less then 908, converter acts
in converter mode and if a is greater than 908, the
converter acts in an inverter mode (discharging mode).

This value is assumed positive for transfer from ac grid to

dc. Fig. 2 shows the block diagram SMES unit.The energy
stored in SMES at any instant in time domain is given as
follows

Fig. 1. The schematic diagram of SMES unit.

Fig. 2. The block diagram of SMES unit.

DPsmi t DEdi Idi0 DIdi DEdi ;

Wsmi t

Li Idi2
2

MJ;

i 1; ; 3

i 1; ; 3

294

A. Demiroren, E. Yesil / Electrical Power and Energy Systems 26 (2004) 291305

3. Modelling for automatic generation control

of the power system with three areas
An interconnected power system is considered as being
divided into control areas, which are connected by tie-lines.
In each control area, all generators are assumed to form a
coherent group. Some of the areas in the power system are
considered having load perturbations having same
magnitudes. The power system investigated in the study,
consisting of three areas is shown in Fig. 3.
The power system is assumed to contain two reheat
turbine type thermal units and a hydro unit. The detailed
block diagram of the interconnected power system is given
in Fig. 4 and system data is given in Table 1.
Each area supplies its user pool, and tie-lines allow
electric power to flow between areas. Therefore, each area
affects others, that is, a load perturbation in one of the areas
affects the output frequencies of other areas as well as power
flows on tie-lines. Due to this, the control system of each
area needs information about the transient situation in all
areas to bring the local frequency to its steady state value.
While the information about each area is found in its
frequency, the information about the other areas are in the
perturbations of tie-lines power flows.
In conventional system, turbine reference power of each
area is tried to be set to its nominal value by an integral
controller and the input of the integral controller of each
area is Bi Dfi DPi i 1; ; 3 called as ACE of the area.
Each of two areas including steam turbines contains
governor, reheaters stage of steam turbine and generation
rate constraints. All of the governors have deadband effects
that are important for speed control under small
disturbances [1]. The reheater effect to the steam turbines
is considered in the state space model equations.
3.1. Governor deadband
To represent the governor deadband is used describing
function approach as in literature [1]. An adequate
description of hysteresis type of nonlinearity is expressed
as: to describe the governor deadband nonlinearity, a
describing function approach [12] is used as explained
below. The governor deadband nonlinearity tends to
produce a continuous sinusoidal oscillation of natural
period of about 2 s. A describing function approach is
used to linearize the governor deadband in terms of change

and rate of change in the speed. The governor deadband is

defined, as the total magnitude of a sustained speed change
within which there is no change in valve position.
The nonlinearity of hysteresis is expressed as [1,12]
y Fx; dx=dt

For a basic assumption, the variable x is taken as a

sinusoidal oscillation
7

x < A sin v0 t

where A is the amplitude of oscillation, v0 is the frequency

of oscillation and v0 2pf0 p with f0 0:5 Hz. Such an
assumption is quite realistic. It has been known that the
backlash nonlinearity tends to give continuous sinusoidal
oscillation with a natural period of about 2 s [12]. Fx; dx=dt
function can be evaluated as a Fourier series as follows [1]
Fx; dx=dt F0 N1 x

N2
dx=dt
v0

For a reasonable approximation, it is enough to consider the

first three terms. As the blacklash nonlinearity is symmetrical about the origin, then F0 is equal to zero and
Fx; dx=dt N1 x

N2
dx=dt DBx
v0

where DB denotes the deadband [13]. For the analysis in

this paper, backlash of approximately 0.5% is chosen [12].
As result, the fourier coefficients are obtained as N1 0:8
and N2 20:2:
3.2. Boiler system
In this study, the effect of the boiler at each steam area
in the power system is also considered as detailed
configuration. This configuration for each steam area is
represented as given by Fig. 5 [12]. This includes the
long-term dynamics of fuel and steam flow on boiler drum
pressure. Representation for combustion controls is also
incorporated. The model is basically for a drum type boiler,
similar responses have been observed for once-through
boilers and pressurized water reactors. The model can be
used to study the responses of coal fired units with poorly
tuned combustion controls and well tuned oil or gas fired
units. In conventional steam units, changes in generation
are initiated by turbine control valves and the boiler
controls respond with necessary immediate control action
upon sensing changes in steam flow and deviations in
pressure [12].
As a result, the state space equations of the power system
with SMES units are written as following
x_ Ax Bu G

Fig. 3. The schematic diagram of power system with three areas.

10

where G is a vector containing nonlinear terms. The state

variables and input for the power system with three areas
including SMES units in the case of using conventional PI

295

Fig. 4. The block diagram of power system with three areas having SMES units.

A. Demiroren, E. Yesil / Electrical Power and Energy Systems 26 (2004) 291305

296

A. Demiroren, E. Yesil / Electrical Power and Energy Systems 26 (2004) 291305

Table 1
The parameter values of the power system (2000 MW area capacity)
KP1;2
TP1;2
Ts1;2
KR1;2
TR1;2
TP3
T1 (3 area)
T2 (3 area)
KId1;2;3
Tdc1;2;3
Id01;2;3;4
Edmax1;2;3
Idmax1;2;3

120
20 s
0.001 s
0.333
10 s
13 s
48.7 s
0.513 s
0.2
0.026 s
4.5 kA
21.3 kV
6.21 kA

TG1;2
KI1;2;3
R1;2;3
TE1;2;3
DPD1;2;3
KP3
T3 (3 area)
TW (3 area)
K01;2;3
L1;2;3
Ed01;2;3
Edmin1;2;3
Idmin1;2;3

0.2 s
0.001
2.4
0.3 s
^0.01 p.u. MW
80
10 s
1s
50
2H
0
3 kV
4.05 kA

controller are
xT Df1 ;DPR1 ;DPG1 ;DPref 1 ;DxE1 ;Df2 ;DPR2 ;DPG2 ;DPref2 ;
DxE2 ;Df3 ;DxE3 ;DPR3 ;DPG3 ;DPref3 ;DP12 ;DP23 ;DP31 ;
DEd1 ;DId1 ;DEd2 ;DId2 ;DEd3 ;DId3 
uT DPD1 ;DPD2 ;DPD3 
the parameters in equation above are given in List of
Symbols. In the case that no SMES unit is contained, the
state variables DEdi and DIdi i 1;;3 are not available.
Moreover, in order to project physical constraints,
a generation rate limitation of 0.1 p.u./min
(i.e. 0.0017 p.u. MW/s) for thermal areas and 4.5%/s for
hydro unit are considered [14]. By considering the matters
mentioned above, the state space equations are obtained in

Fig. 5. The detailed configuration of boiler.

discrete time domain as given in Appendix A. The variables

used in Appendix A are given in List of Symbols.

4. Fuzzy logic controllers

The AGC based on FLC is proposed in this study. One of
its main advantages is that controller parameters can be
changed very quickly by the system dynamics because no
parameter estimation is required in designing controller for
nonlinear systems. Therefore a FLC, which represents a
model-free type of nonlinear control algorithms, could be a
reasonable solution.
There are many possibilities to apply fuzzy logic to the
control system. The fuzzy logic structure for the all
controller design can be seen in Fig. 6 [5]. There are four
main structures in a fuzzy system: the fuzzifier, the inference
engine, the KB, and the defuzzifier. The first stage in the
fuzzy system computations is to transform the numeric into
fuzzy sets. This operation is called fuzzification. From the
point of view of fuzzy set theory, the inference engine is
the heard of the fuzzy system. It is the inference engine that
performs all logic manipulations in a fuzzy system. A fuzzy
system KB consists of fuzzy IF THEN rules and membership functions characterizing the fuzzy sets [5]. The result of
the inference process is an output represented by a fuzzy set,
but the output of the fuzzy system should be a numeric
value. The transformation of a fuzzy set into a numeric
value is called defuzzification. In addition, input and output
scaling factors are needed to modify the universe of
discourse. Their role is to tune the fuzzy controller to
obtain the desired dynamic properties of the processcontroller closed loop [5].
In this paper, the inputs of the proposed Fuzzy controllers
are ACE, and change rate in ACE (DACE) as shown in
Fig. 7, which is indeed error e and the derivation of the
error _e of the system, respectively. This gives us a fairly
good indicator of the general tendency of the error.
Many fuzzy controller structures based on various
methods have been presented. The most widely used
methods in the practice is the Mamdani method proposed
by Mamdani and his associates [14] who adopted the
min max compositional rule of inference based on an
interpretation a control rule as a conjunction of the
antecedent and consequent. It is natural to apply the
conventional theory, to solve the nonlinear problem of
fuzzy controller and much work has been done in this
direction. In literature, firstly, compared the linear rule type
fuzzy controller with classical PID controller in their
input output relationship. Later, a fuzzy controller of the
crisp type can approximate any real-valued continuous
function to any degree of accuracy is proved [15].
Conventional controllers are derived from control theory
techniques based on mathematical models of the open-loop
process to be controlled. For instance, a conventional
proportional integral (PI) controller can be described by

A. Demiroren, E. Yesil / Electrical Power and Energy Systems 26 (2004) 291305

297

Fig. 6. Components of a Fuzzy System.

the function
u Kp e Ki

e dt

11

According to the conventional automatic control theory, the

performance of the PI controller is determined by its
proportional parameter Kp and integral parameter Ki [16].
The proportional term provides control action equal to some
multiple of the error, while the integral term forces the
steady state error to zero.
Since the mathematical models of most industrial process
systems are type 0, obviously there would be steady-state
error if classical PD fuzzy controller controls them [16].
Whenever the steady-state error of the control system is
eliminated, it can be imagined substituting the input DACE
of the fuzzy controller with the integration of error. This will
result in the fuzzy controller behaving like a parameter
time-varying PI controller; thus the steady-state error is
removed by the integration action. However, these methods
will be hard to apply in practice because of the difficulty of
constructing fuzzy control rules. Usually, fuzzy control
rules are constructed by summarizing the manual control
experiences of an operator who has been controlling the
industrial process skillfully and successfully. The operator
intuitively regulates the executer to control the process by
watching the error and the change rate of the error between
output of the system and the set-point value given by the
technical requirement. It is no practical way for the operator
to observe the integration of the error of the system.
Therefore it is impossible to explicitly abstract fuzzy control
rules from the operators experience. Hence, it is better to
design a fuzzy controller that possesses the fine characteristics of the PI controller by using only ACE and DACE.

Fig. 7. The PI-type fuzzy controller.

Fig. 8. The proposed PI-type fuzzy controller.

One way is to have an integrator serially connected to the

output of the fuzzy controller, as shown in Fig. 7 [17].
The control input to the plant can be approximated by
ub

ut dt

12

where b is the integral constant, or output scaling

factor. Hence, the fuzzy controller becomes a parameter
time-varying PI controller. The controller is called as
PI-type fuzzy controller, and the fuzzy controller without
the integrator as the PD-type fuzzy controller. In a PI-type
fuzzy control system, the steady-state error is zero, but when
the integral factor is small the response of the system is

Fig. 9. The membership functions of ACE and DACE.

Table 2
Lookup table of fuzzy rules

298

A. Demiroren, E. Yesil / Electrical Power and Energy Systems 26 (2004) 291305

slow, and when it is too large there is a high overshoot and

serious oscillation [16].
In this paper, a controller structure that simply connects
the PD-type and PI-type fuzzy controller together in parallel

is proposed. The fuzzy controller is shown in Fig. 8 [9],

where a is the weight on the PD-type fuzzy controller and b
is that on the PI-type fuzzy controller. Whenever the value
of a=b is large, it means that the derivative control is more

Fig. 10. The results for a step load perturbation at only first area in the case of without SMES units in none of the areas.

A. Demiroren, E. Yesil / Electrical Power and Energy Systems 26 (2004) 291305

effective than the integral control, and vice versa.

From another point of view, the fuzzy controller behaves
as a time-varying PID controller.
The type of the FLC obtained is called Mamdani type
which has fuzzy rules of the form

299

If ACE is Ai and DACE is Bi THEN u is Ci i

1; 2; 3; ; n:
Here, Ai ; Bi ; Ci ; are the fuzzy sets. The triangle
membership functions for each fuzzy linguistic values of

Fig. 11. The results for a step load perturbation at only first area in the case of with SMES units in all of the areas.

300

A. Demiroren, E. Yesil / Electrical Power and Energy Systems 26 (2004) 291305

Fig. 11 (continued )

the ACE and DACE are shown in Fig. 9 [10], in which NB,
NM, NS, Z, PS, PM, and PB represent negative big,
negative medium, negative small, zero, positive small,
positive medium, and positive big, respectively. Also, the
set of fuzzy rules is shown in Table 2.

5. Simulation results
In this study, the application of Fuzzy PI-type controller
to AGC in the power system with three areas having two
steam turbines and one hydro turbine tied together
through power lines is investigated. Moreover, each area
in the power system includes SMES unit. So, affirmative
effect of SMES on AGC is also taken into account in the
study. The effect of SMES in the improvement of AGC at
power systems is known in literature.
The load perturbation or perturbations having amplitude
of 0.01 p.u. MW to the given areas is applied and
the frequency oscillations and tie-line power flows are
investigated. The simulation incorporates detailed model for

the boiler dynamics, steam reheat process and governor

deadband nonlinearity effects using the describing function
approach and the limits of the rate of generating power
belonging to each thermal unit area. The oil/gas-fired boiler
used in thermal units of the power system considered is
having the fastest response among boilers as known in
literature. Governor deadband negative effect on settling
time and on the amplitude of oscillation is known therefore
governor effects are taken into account. Moreover, the
physical boundaries such as both generating limits and the
limits of operation of SMES unit at each area are taken into
account. The simulations are implemented by using
MATLAB Simulink program and FLT. The conventional
PI controllers and fuzzy logic PI-type controllers of
Mamdani type are investigated as comparison due to load
perturbation/or perturbations in the same power system.
In this paper, to damp out the oscillations as fast as
possible, AGC is implemented by using the fuzzy logic
PI-type controllers and SMES units in each area at the
power system. The results obtained show that the controller
improves effectively the damping of the oscillations after

A. Demiroren, E. Yesil / Electrical Power and Energy Systems 26 (2004) 291305

301

Fig. 12. The results for step load perturbations at first two areas in the case of without SMES units in none of the areas.

the load deviation in one of the areas in the interconnected

system compared to conventional controllers. In addition,
during operation, both the effect of SMES and the effect of
FLC are investigated together.
In this work, firstly, a step load increase in the first area of
the power system and then step load increases in both first

area and second area of the same power system are

considered. For comparison, the considered power system
is controlled by using
(i) Conventional PI controller
(ii) Fuzzy PI-type controller

302

A. Demiroren, E. Yesil / Electrical Power and Energy Systems 26 (2004) 291305

for these cases. These simulations are carried out while

SMES unit is present in each area at the power system and
without SMES unit in any of areas, separately. In Fig. 10a f,
the deviations of frequencies of each area and the deviations

of tie-line power flows for a step load perturbation at only

first area in the case of without SMES units in any of
the areas are shown, respectively. In the figures, the results
are given as compare with the cases of conventional PI

Fig. 13. The results for step load perturbations at first two areas in the case of with SMES units in all of the areas.

A. Demiroren, E. Yesil / Electrical Power and Energy Systems 26 (2004) 291305

303

Fig. 13 (continued )

controller and Fuzzy PI-type controller. The dashed lines

represent the case of fuzzy PI-controllers, and dotted lines
represent the case of conventional controllers. In Fig. 11a f,
the deviations of the same variables for the same operation
case with SMES unit at each area are shown. Also, in
Fig. 11g i, the currents of SMES units is shown for the
same case. In Fig. 12a f and in Fig. 13a i, the deviations of
the same variables are shown for step load increases in both
of first two areas, by using conventional PI and Fuzzy
PI-type controllers.
The aim of the study is to investigate the effects of
Fuzzy PI-type controller and SMES unit on improving
dynamic performance of the power system. The results
obtained show that the performance of Fuzzy PI-type
controller is better than conventional PI controller as the
main objective of the study. Moreover, the simulations
represent the positive effect of SMES unit on the
improvement of the oscillation of frequency due to any
load perturbation.

6. Conclusions
This study is an application of FLC to AGC in power
system with three areas for both cases, with SMES units
available in all areas and without SMES unit in any of
areas. In this work, transient behaviours of the frequency
of each area and tie-line power deviations in a power
system including three areas are considered due to the
load perturbations. In practice, power systems generally
have more than two areas and each area has different
properties from others. Because of this, in the study, the
power system with three areas of which consisting of
two thermal units and the other one consisting of a hydro
unit is considered. In the simulation, detailed model
incorporates the boiler dynamics, steam reheat process
and governor deadband nonlinearity effects using the
describing function approach. Moreover, the physical
boundaries such as both generating limits and the limits
of operation of SMES unit at each area are taken into
account.

304

A. Demiroren, E. Yesil / Electrical Power and Energy Systems 26 (2004) 291305

According to deviation of the power system energy

demand, the SMES unit releases the needed energy or
absorbs residue energy from power system. The model
including generating unit and SMES unit together represents
realistic performance of the power system. So, the nonlinear
state space equations of the power system are used directly
during the control of the power system by FLC and by
conventional proportional integral controller.
In the last 10 years, fuzzy controllers are applied,
successfully, to many industrial processes, which are mostly
nonlinear. Since the power systems are also inherently
nonlinear, nonlinear controllers are needed, so FLCs give a
good solution. In this paper, a new FLC, called fuzzy PI-type,
is applied to AGC in the power system having three areas.
The design of the proposed Fuzzy controller is very simple
and effective. Since it is a model-free type of controller, it can
be implemented to a power system.
From the obtained results, it is shown that the performance
of Fuzzy-PI controller outperforms than that of conventional
PI controller at AGC in power system. Also, the positive
effect of SMES unit on AGC is represented, in the study.

2KI3
B3 Df3 s DP3 s
s

DPref3 s

A10

The expression of deviations of tie-line power in the power

system as follows
1
2pT12 Df1 s 2 2pT12 Df2 s
2
1
DP23 s 2pT23 Df2 s 2 2pT23 Df3 s
s
DP12 s

DP31 s

1
2pT31 Df3 s 2 2pT31 Df1 s
s

A11
A12
A13

The last state variables for each of these areas are DPi i
1; ; 3 and the state space equation related to the variables
are different for each area depending on the system
configuration (Fig. 2). These are as following
DP1 s DP12 s a31 DP31 s

A14

DP2 s a12 DP12 s DP23 s

A15

DP3 s a23 DP23 s DP31 s

A16

Parameters of thermal turbine and hydro turbine are

given in List of Symbols.
Appendix A
The state space equations of the power system is given as
follows; for thermal unit in discrete time domain equations
KPi
Dfi s
DPGi s 2 DPDi s
1 sTPi
A1
2 DPi s 2 DPsmi s
1
Dx s
A2
DPRi s
1 sTTi Ei
DPGi s

1 sKRi TRi
DPRi s
TRi

2KIi
Bi Dfi s DPi s
s


1
1
DxEi s
DPrefi s 2
Dfi s
1 sTGi
Ri

DPrefi s

A3
A4
A5

where subscript i represents each thermal area in power

system i 1; 2:
For hydro turbine state space equation in Laplace domain
is given as below
KP3
DPG3 s 2 DPD3 s
1 sTP3
2 DPsm3 s 2 DP3 s


1
1
DxE3 s
DPref3 s 2
Df3 s
1 sT1
R3
Df3 s

A6
A7

DPR3 s

1 sT2
Dx s
1 sT3 E3

A8

DPG3 s

1 2 sTW
DPR3 s
1 0:5sTW

A9

References
[1] Lu C-F, Liu C-C, Wu C. Effect of battery energy storage system
on load frequency control considering governor deadband and
generation rate constraints. IEEE Trans Power Syst 1995;10(3):
555 61.
[2] Cavin JK, Budge MC, Rosmunsen P. An optimal linear system
approach to load-frequency control. IEEE Trans Power Apparatus
Syst 1971;PAS-90:2472 82.
[3] Ai-Homouz ZM, Abdel-Magid YL. Variable structure load frequency
controllers for multiarea power systems. Int J Electr Power Energy
Syst 1993;15(5):293 300.
[4] Chang CS, Fu W. Area load frequency control using fuzzy gain
scheduling of PI controllers. Electr Power Syst Res 1997;42:
145 52.
[5] Lee CC. Fuzzy logic in control systems: fuzzy logic controller, parts
I II. IEEE Trans Syst, Man Cyber 1990;20(2):40418. see also
419 435.
[6] Das D, Nanda J, Kothari ML, Kothari DP. Automatic generation
control of an interconnected hydrotermal system in continous and
discrete modes considering generation rate constraints. IEE Proc
1983;3(Pt D 1):46171.
[7] Pedrycz W. Fuzzy control and fuzzy system. New York: Wiley;
1993.
[8] Mitani Y, Tisuji K, Murakami Y. Application of superconducting
magnet energy storage to improve power system dynamic
performance. IEEE Trans Power Syst 1988;PWRS-3(4):1418 25.
[9] Tripathy SC, Balasubramania R, Chanramohanan Nair PS. Effect of
superconducting magnetic energy storage on automatic generation
control considering governor deadband and boiler dynamics. IEEE
Trans Power Syst 1992;PWRS 3(7):126673.
[10] Tripathy SC, Balasubramania R, Chanramohanan Nair PS. Adaptive
automatic generation control with superconducting magnetic energy
storage in power system. IEEE Trans Energy Convers 1992;EC 7(3):
434 41.

A. Demiroren, E. Yesil / Electrical Power and Energy Systems 26 (2004) 291305

[11] Wesley Hines J. Matlab supplement to fuzzy and neural approaches in
engineering. New York: Wiley; 1997.
[12] Tripathy SC, Chanramohanan Nair PS. Automatic generation control
with superconducting magnetic energy storage in power system.
Electr Machines Power Syst 1994;31784.
[13] Elgerd OI. Electric energy system theory: an introduction. New York:
McGraw-Hill; 1971.

305

[14] Mamdani EH, Assilian S. An experiment in linguistic synthesis with a

fuzzy logic controller. Int J ManMachine Stud 1975;7:113.
[15] Patyra MJ, Mlynek DM. Fuzzy logic, implementation and applications. New York: Wiley; 1996.
[16] Benjamin CK. Automatic control systems. Englewood cliffs, NJ:
Prentice-Hall; 1987.
[17] Arzen KE, Johansson M, Babuska R. A Survey on Fuzzy Control.
FAMIMO Deliverable D1.1b, Lund; 1997.