1676 IEEE Transactions on Power Systems, Vol. 10, No.

3, August 1995

OVERVIEW AND LITERATURE SURVEY OF

FUZZY SET THEORY IN POWER SYSTEMS
J.A. Momoh SM, IEEE X.W. Ma* K.Tomsovic Member, IEEE
Dept. of Electrical Engineering Electrical Engineering & Computer Science
Howard University Washington State University
Washington, D.C. 20059 Pullman, Washington 99164-2752
* Former graduate student, Dept. of EE, Howard University
A B S T R A C T : Increasing interest has been trivial. On the other hand, there are many uncertainties
seen in applying fuzzy set theory to power in various power systems problems because power
systems problems from the number of systems are large, complex, geographically widely
publications on this topic in recent years. distributed systems and influenced by unexpected events.
As a relatively new research topic a need is These facts make it difficult to effectively deal with
felt t o pay m o r e a t t e n t i o n to the many power systems problems through strict
understanding of the basic principles of the mathematical formulations alone. Therefore, expert
theory and the identification of problems systems approaches, as one area of artificial
suitable for solving by this method. This intelligence, emerged in recent years in power systems
paper p r e s e n t s a s u r v e y of r e c e n t (91*92)as a complement to mathematical approaches
publications on applications of fuzzy set and proved to be effective when properly coupled
theory to power systems and the basic together.
procedures for fuzzy set based methods to Expert systems are built based on the
solve specific power systems problems. knowledge acquired from domain experts. The expert's
Simple numerical examples are used to show empirical knowledge is generally expressed by language
t h e practical p r o c e d u r e s of problem containing ambiguous or fuzzy descriptions. As a
formulation and solution. result, classical Boolean logic may not be efficient for
the representation of such expertise. Fuzzy logic, on
1. INTRODUCTION the other hand, is a natural choice for this purpose and
many consider it promising (499).
With the penetration of fuzzy set theory into The growing number of publications on
manufacturing and computer products, applications of applications of fuzzy-set-based approaches to power
fuzzy set theory in power systems are beginning to systems indicates its potential role in solving power
receive attention from power systems researchers. systems problems. The results obtained so far are
Fuzzy sets were first introduced in solving power promising but fuzzy set theory is not widely accepted.
systems long-range decision making problems in [ 171 The reasons for its lack of acceptance include:
more than a decade ago. However, substantial interest misunderstanding of the concept, excessive claims of
in its applications to power areas is fairly recent. some researchers, lack of implemented and available
Analytical solution methods exist for many systems, and its status as a new theory. With the
power systems operation, planning and control advance of fuzzy set theory and the achievements made
problems. However, the mathematical formulations of in applications to other areas, it is beginning to receive
real-world problems are derived under certain restrictive power researchers' attention and it is felt that there is a
assumptions and even with these assumptions, the need to develop more information on this subject. In
solutions of large scale power systems problems is not this paper we aim to provide a tutorial introduction to
and a survey of the publications and applications of
95 WM 208-9 PWRS A paper recommended and approved fuzzy set theory to power systems problems.
by t h e IEEE Power System Engineering Committee of In this paper, long-range planning and control
t h e IEEE Power Engineering Society f o r p r e s e n t a t i o n problems are taken as examples to illustrate fuzzy
a t t h e 1995 IEEE/PES Winter Meeting, January 29, t o
February 2, 1995, New York, NY. Manuscript submitted methods for specific problem solving. Two examples
January 4, 1994; made a v a i l a b l e f o r p r i n t i n g used to explain the f r m y set methods are generator
December 28, 1994. maintenance schedui. 2 and power system stabilizer
design. Through these simple examples, the extension
of these approaches to practical power system problems
can be envisioned.
The paper is organized as follows.
Fundamental concepts of Fuzzy set theory are explained
in Section 2. Applications of fuzzy sets in power
system operation, planning and control are outlined and

0885-8950/95/$04.00 0 1995 IEEE

 ~

1677

discussed in Section 3. In Section 4, an example of

fuzzy set application to unit maintenance problems is
described. Section 5 presents the design procedure of a
fuzzy controller for power systems dynamic stability It should be pointed out that other operations
can be defined, in articular operations in the T-norm
improvement and suggestions for further development
follow. P
and T-conorms(* (called s-norm classes(l l)). In
addition, to perform certain mathematical operations, a
2. BASICS OF FUZZY SET THEORY crisp set (non-fuzzy) may be required. The following
definition of an a-cut can be used to create a family of
(1) Classical Sets and Fuzzv Sefs crisp sets from a given fuzzy set:
Let X be the universe of objects with elements
x, where A is called a fuzzy sub-set of X (generally
called a fuzzy set). Membership of x in classical set A
can be viewed as a characteristic function p,t, from X to
(0,l) such that
Furthermore, a particularly useful mathematical
construct is the fuzzy relation. A fuzzy binary relation
is defined as:

For a fuzzy set A of the universe X, the grade

of membership of x in A is defined as:
For a fuzzy relation R. there is the following fuzzy
composition computation :

where ~ A ( x )is called the membership function. The

value of ~ A ( x )can be anywhere from 0 to 1, and this
range is what makes it different from a crisp set. The
closer the value of ~ A ( x is ) to 1.0, the "more" x (3) FUZZY Measurn
belongs to A. Thus, the fuzzy set A has no sharp Fuzzy measures are criteria for measuring
boundary. Each of the crisp subsets of X can be shown attributes of objects. When a certain set X is
to have a one-on-one correspondence with the considered, the function g that makes subsets E and F
characteristic function, and because the membership correspond to the values in the interval [O,l J are called
functions are extensions of the characteristic functions, fuzzy measures if they have the following properties:
fuzzy sets are extensions of crisp sets(lO7).
Fuzzy set elements are ordered pairs indicating
the value of a set element and the grade of membership,
i.e.
If E E F , g ( E )5 g ( F ) ;

(2) Basic Fuzzv Set Om rations

In the following, the most common set
operators are defined. For two fuzzy sets A and B, their
union operation is defined as:

the intersection operation is:

Property (a) expresses boundary conditions, (b)
monotonicity, and (c) continuity. Fuzzy measures are
monotonic set functions. When (b) satisfies the
and the complement operation is: equality condition, g is referred to as a possibility
measure.

I
1678

In solving practical problems, it is important expert system approaches, more often than not expert
to adopt the appropriate fuzzy operators and even new systems serve as a complementary approach to
operations can be constructed to reflect the expert's numerical algorithms for improved performance, both in
reasoning(3). computation and in practicality. In the latter case, it is
frequently difficult to make expert systems and
(4) ApDlication of Fuzzv Set Theorv numerical algorithms coordinate efficiently. Even for
When fuzzy set theory is used to solve real the solving of the former class of problems, the crisp
problems, the following steps are generally followed: representation of human empirical knowledge, usually
expressed in natural languages and containing inherently
Step 1 Description of the original problem. The uncertain representations (such as "very likely", "quite
problem to be solved should first be stated likely", "if x is large, then y is very likely to occur", "x
mathematically/linguistically. should not be less than a", etc.), is very difficult and
lacks flexibility. Furthermore, power systems are large
Step 2 Definition of thresholds for variables. For a systems and are affected by unexpected events. In some
variable, there is a specific value with the cases these uncertainties are represented by probability.
greatest degree of satisfaction evaluated from However, it has recently been made clear that some of
empirical knowledge and a certain deviation is the uncertain factors are intrinsically of fuzzy nature and
acceptable with decreasing degree of are difficult to manage properly by probability.
satisfaction until a value that is completely There are problems in power systems that
unacceptable. The two values corresponding to contain conflicting objectives. In power syslems
the greatest and least degree of satisfaction are operation, economy and security, maximum load supply
termed thresholds. and minimum generating cost are conflicting objectives.
The combination of these objectives by weighting
Step 3 Fuzzy quantization. Based on the threshold coefficients is the traditional approach to solving this
values from step 2, proper forms of problem. Fuzzy set theory offers a better
membership functions are constructed. Many compromise( 12,13,17,21,25.43,45.46.49,51) and
forms of membership functions are available, obtain solutions which cannot be easily found by
such as linear, piece-wise linear, trapezoidal, weighting methods. The benefits of fuzzy set theory
parabolic, and so on. The membership over traditional methods are as follows:
functions should reflect the change in degree of
satisfaction with the change in variables a. It provides alternatives for the many attributes
evaluated by experts. of objectives selected.

Step 4 Selection of the fuzzy operations. In terms of b. It resolves conflicting objectives by designing
the practical decision-making process by weights appropriate to a selected objective.
human experts, a proper fuzzy operation is
selected so that the results obtained is like C. It provides capability for handling ambigiuty
those obtained by experts. The interpretation expressed in diagnostic processes which
of results using fuzzy systems is based on involves symptoms and causes.
domain experts' reasoning. Therefore, at this
level a hybrid fuzzy set-expert system scheme d It develops process control as a fuzzy relation
is desirable. It helps to remove any ambiguity between information about the condition of
that may occur in problem solving. The most the process to be controlled.
commonly used operations are Mamdani's and
Zadeh's. e. It develops intelligent robots that employ
sensors for path or position determination.

3. APPLICATIONS OF FUZZY SETS TO f. It improves human reliability models in cases

POWER SYSTEMS where many people perform multiple tasks.

In recent years, expert systems and other areas Power systems components have physical and
of artificial intelligence techniques were introduced in operational limits which are usually described as hard
power systems operation planning and control. Expert inequality constraints in mathematical formulations.
systems are typically built on domain expert's The elimination of minor violations of some
knowledge. Although some problems in power constraints (practically acceptable) usually greatly
systems, such as diagnosis (especially increases the computational burden and decreases the
transformer/generator malfunction diagnosis), alarm efficiency and may even prevent finding a feasible
processing, and others, can be solved independently by solution. In practice, certain slight violations of the
 1679

inequality constraints are permissible. This means that demand level, new station locations, environmental
there is not a clear constraint boundary and the effects, etc., have a decisive effect on the decision-
constraints can be made soft. Traditionally, this making and yet are difficult to represent
problem has been managed either by modifying the deterministically. Further, the objectives and
objective function or the underlying iterative process. constraints are uncertain or competing. In fact. the
The fuzzy set approach inherently incorporates soft decision-making process in power systems expansion
constraints and thus simplifies implementation of such planning is to a large extent, qualitative and can be
considerations. described more flexibly and intuitively by fuzzy set
In this area of power systems control, the concepts.
optimal control theory is often applied in research to
design controllers for the enhancement of power b. LongMd-Term Schedulin$a*a)
systems stability and other purposes. Since power Power systems lonumid-term scheduling
systems are large non-linear systems, simplifying problems, such as yearly maintenance scheduling,
assumptions are necessary in the design of such seasonal fuel scheduling, and mid-term operation mode
controllers. Due to model-dependence, controller's studies, are solved by various optimal and heuristic
adaptability and robustness are problematic. Recently methods. These problems contain many complications
developed fuzzy-logic-based-controllers show promise and uncertainties. Accurate formulation of the problems
for robust performance and adaptive schemes. themselves are difficult and the conventional
Table 1 is a list of the more common power optimization methods are not efficient. In addition, it is
system application areas of fuzzy set theory, with more reasonable to represent constraints as soft (the
sample references. fuzzy degree of satisfaction expression permits an
engineering acceptable violation of constraints) than as
hard. The combination of conventional methods with
Table 1. Fuzzy Sets Applications in Power Systems fuzzy sets may constitute an effective approach to solve
these problems.
Areas

related
I Problems
Planning I Planning
I Scheduling
I Reference
117-21.24
123.26-29 I
I (2) npemt ion Areas
a. Dynamic Security
~

areas Reliability 22
Operation Contingency analysis 30.33.39.5 I Dynamic security assessment (DSA)is a major
areas VARNoltage control 3 1.32,41,56,60,6 1 application in power systems operation and many
Stability evaluation 34-38 methods have been developed. Since power systems
Load forecasting 45-48 security varies according to engineering-economic
Load Management 43.44.49 considerations, the reliability of protective relays, the
Decision-making 39.54.55.57 quality of the models used, and the risk of various
support faults, it may not be reasonable to say that the system
Multi-objective 52.53 is definitely secure or not with the conventional binary
coordination
Monitoring 8c control 4 1 logic. The present practice in DSA is to perform off-
Unit commitment 58 line studies for a wide range of likely system conditions
I State estimation 50 I and network configurations. In on-line applications the
Control Fuzzy logic stabilizer 63.65-72. results of off-line studies are not directly available, but
areas 76,77,81,84 operators are allowed to rely on their own judgment and
ConverterDrives 64,80.82,83.95 knowledge acquired in off-line studies and experience
Other controls 73.75.78.79 gathered during long-time operation. Fuzzy set theory
Diagnosis Transformer diagnosis 85-87.90 may be useful in building an expert system for DSA
areas Network diagnosis 89,91,92 based on the operator's empirical knowledge.
Machine diagnosis 88
b. Load Demand Forecasting(4548)
Some of the problems are briefly described below. Power system load demand is influenced by
many factors, such as weather, economic and social
activities, and different load components (residential,
(1) Planning Related Areas industrial commercial, etc.). By analysis of only
a. Power Systems Expansion P I a r ~ n i n g t ~ ~ - ~ ~ . * ~ ) historical load data, it is difficult to obtain accurate load
There are a number of judgments based on dcmand for forecasting. The relation between load
experience and expert opinion which are crucial in demand and the independent variables is complex and it
decision-making for power system expansion planning. is not always possible to fit the load curve using
Usually it is awkward to capture experience within the statistical models. Expert system approaches have
constraint formulations of conventional optimization shown advantages over conventional methods. The
models. This is because many factors, such as load
1680

numerical aspects and uncertainties of this problem programming(100). integer programrning(lOl), etc.
appear suitable for fuzzy methodologies. Since
generator maintenance scheduling problem concerns
(3) Control Areas several complications and uncertainties, accurate
a. FUZZY Logic ~ontrollers(63-84) formulation of the problem itself is difficult and the
In traditional controller design, a system model above methods require extensive computation. An
needs to be constructed and control laws are derived from extension of the dynamic programming approaches is to
the analysis of the model. Because of the non- include fuzzy sets for the constraints. This involves the
linearities it is nearly always necessary to linearize the fuzzification of all constraints by proper membership
system model and then the linear controllers are used to functions. This problem was solved in (271 by fuzzy
control the non-linear system. Fuzzy logic controllers dynamic programming and the membership functions
have received much attention in r a n t years, since they were defined as weighted piece-wise linear functions.
are more model-independent, show high robustness, and which is used in general-formulation.
can adapt. Fuzzy logic controllers are mainly used for
power systems excitation and converter controls. (1) Qrirrinal Problem Formulation
A state (1,T) is defined by the I-th maintained
(4) Diagnosis generator and its maintenance starting at T-th time
a. Diagnosis(85-92) interval. A recursive algorithm to compute the
Human experts play central roles in trouble minimum cost for generator I to be maintained in time
shooting or fault analysis. In power systems, it is interval T is exprcsscd as:
required to diagnose equipment malfunctions as well as
disturbances. Equipment malfunctions are caused by
many factors and the information available to perform
diagnosis is incomplete. In addition, the conditions that
induce faults may change with time. Subjective where: Fcost(I,T)= least total cost increase to get to
conjectures based on experience are necessary. state (I,??
Accordingly, h e expert systems approach has proved to Pcos[(I,T) = production cost increase for
be useful. As stated previously, fuzzy set concepts can generator I maintained starting at T-th time interval;
lend itself to the representation of knowledge and the (J) = set of feasible states at stage T- 1.
building of an expert system.
Reserve marnin consrraint
b. Fuzzy Logic and Artificial Neural network^(^^-^^^) For secure operation of a power system, it is
Fuzzy logic and artificial neural networks necessary that the system always have a specified
(ANN) both have an interpolative quality(94). The generation reserve margin. that can be expressed as
combination of fuzzy sets with multi-layered neural follows:
networks can improve the performance of A N N s ( ~ ~ ) .
Through this approach, multi-layered ANNs are able to I

be used for those problems for which numeric training C G ,IPi- Pd - P, - Pb, (1 1)
samples cannot be constructed and human expertise is i=l
already available. On the other hand, the introduction of
fuzzy concepts can increase the amount of information where: Pd = system load demand: P, = reserve margin
available from ANN'S output. Applying fuzzy set capacity; Pi = total generation capacity; CGi =
concepts may render it possible to make better use of total maintenance capacity; Pioss = network
the ANN output information. Fuzzy logic based ANNs real power loss.
are used for transformer fault diagnosis and control in
(90) and (95). m o w e r COnStralnf
For a scheduled maintenance to be feasible, the
4. GENERATOR MAINTENANCE needed manpower hours must be less than maximum
SCHEDULING BY FUZZY DYNAMIC manpower available from the power utility, i.e.:
PROGRAMMING

Generator maintenance is a routine scheduling task for

power utilities. If done propcrly, can lead to increased
reliability, reduced generation cost, extcnded gcnerator
lifetime and delayed need for new installations. Several where: MHpi=manpower hours to maintain generator
generation maintenance scheduling techniques have bcen
1;
developed, such as, successive application dynamic
 1681

MHpmax(T)= total manpower hours satisfaction of 0.0). The membership function is

available; shown in Fig. 1. The reserve margin is represented by
n = number of generators maintained at the the following membership function:
same time interval T.

tram
To avoid large amounts of transmission loss
and decrease in reserve capacity in one region, the
geographical constraint is used to limit the number of
maintained generators in each region at time interval T

N , ( T ) I N,"'"(T) (13) Where: Rm(I,T) = the least residual capacity b reach

state (1,T.l;
where: N, ( T ) = number of maintained generators in R m a O = the allowable minimum capacity of
region m; reserve margin;
N,"" (T) = maximum number of generators R m s O = the expected security capacity of
allowed to maintain simultaneously in region reserve margin.
m.
Due to maintenance generator I in the time
interval T, the increase in generation cost should not be
The latest and earliest maintenance starting greater than Fa and it is expected to be less than pa
times for different generators are expressed as follows: based on experience. The membership value is 1.0
when the increase in generation cost is less than Ea;
-
the membership value is 0.0 when the increased
generation cost is greater than Pa; the membership
where: Ti = the starting time to maintain generator I;
. Tie = the earliest possible starting time to value is less than 1.0 but greater than 0.0 if the increase
maintain generator I; in generation cost Pa E (ea,Fa). This
Til = the latest permitted starting time to membership function is shown in Fig. 2.
maintain generator I.
(2) F ~ z Oumt
v ization of Obiectives
For the generator maintenance scheduling
problem, two fuzzy objectives are used, the reserve
margin and the production cost increase. The former
should be maximized and the latter minimized

-
_p P P( 1.n
a a a
Fig.2 Membership function for production cost increase

From Fig. 2, the membership function for

production cost increase due to the maintenance of
RmaC) Rms(T) RmO,T) generator I in the time interval T is expressed as:
Fig.1 Membership function for reserve margin

According to the operation guidelines of a

power utility, the reserve margin is specified as: an
expected reserve margin of Rm (with the degree of
satisfaction 1.O), an allowable reserve margin slightly
less than,,R (degree of satisfaction less than 1.O) but a
minimum reserve margin not less than Rm, (degree of

I
1682
where: Pa(I.T) = production cost increase to maintain
generator I in time interval T;
P (I)= expected cost increase due to the
-a
maintenance of generator I;
Fa
( I ) = allowable cost increase due to the
maintenance of generator I.

(3) Fuzzification of Constraints

Fuzzv manpower constraint

For a scheduled maintenance to be feasible, the Te Tb n
required manpower hours should not exceed the Fig 4 Membership function for maintenance window
maximum manpower available from the utility. If the
average available manpower hours in time interval T are The membership function in Fig. 4 can be
M H p and the maximum MHpmax; while the expressed by the following function:
membership value is in the interval (0.1) if the required
manpower hours are in the interval [MHp,MHpmaX].
This membership function is depicted in Fig. 3. The
analytical expression is:

0;if MH,(Z,T) 2 MH,"""(T)

k ! ( Z , T )= I MH, (Z,T )- MH," (T)

M H Y X(T)- MH, (T)

I i f MH,"(T)S M H , ( Z , T ) S MH,""(T)
&if MH,(Z,T) L MH,"""(T)
(17)
where: Tb(I) = the best starting time to maintain
generator I;
Te(1) = the earliest starting time to maintain
where: MHp(I,T) = the least manpower hours required generator I;
to reach state (1,T); Tl(I) = the latest starting time to maintain
MHpmaX(T) = maximum manpower hours generator I.
provided possibly by the utility in time
interval T; Fuzzv r Dhicd constraint;
MHpav(T) = the available manpower hours in For a large interconnected power system,
the utility in time interval T. power transactions are usually scheduled and
constrained. or power transfer over a long distance is not
Fuzzv maintenance window constraint expected. Therefore, the capacity (or number of
Based on the operation history of a specified generators) that can be used for maintenance in region j
generator, a maintenance expert may assume the best in the time interval T is limited. This constraint is also
time to maintain this generator is Tb but should not be soft and a slight violation is permissible. If we use
later than TI. However, according to the load demand, MgmaxCj,T)and Mg,,T) to express the maximum and
this generator is required at least up to Te. That is, the available maintenance capacity in time interval T, the
earliest time to maintain the generator. The membership function can be described intuitively by
membership function for maintenance window Fig. 5 .
constraints can be depicted by Fig. 4.

I I
1a.X Mg6.T) Mg %k MgO.T)
~ b V C T ) W PCr) MHdI,T)
Fig. 3 Membership function for manpower ho.urs Fig. 5 Membership function for geographical constraint
 1683

Corresponding to Fig. 5 , the following Table 2 S stem Conditions

membership function is obtained:

Manpower hours
Time duration

Time 1 2 3 4 5 6 7 8 9 lo7
.interval
Laad 8 3 80 81 22 41 35 44 23 72 69
value
where: Mg(I,T) = the least maintenance capacity to
reach state (1.T); The given thresholds are:
Mg(i.T) = the available maintenance capacity - Each maintenance can be scheduled advanced r
of the region j in time interval T; deferred by one time interval;
Mg(i)mm = the maximum permitted - The manpower hours available are limited U)
maintenance capacity of region j in time 200 for all time intervals;
interval T. - The net reserve capacity limits, Rma and
Rms, are specified as 10% and 15% of the load
(4) Fuzzv Dvnamic Programming demand of the time interval under
With fuzzy constraint membership functions consideration.
and fuzzy objective membership functions, the optimal For generator 3 maintained at time interval
generator maintenance scheduling. depoted by optimal 4.5, or 6. the thresholds are Rma(4)=2.2, Rms(4) = 3.3;
decision PD, is the interaction of fuzzy sets hi,pc2, Rma(5)=4.1, Rms(5)=6.15; Rma(6)=3.5, Rms(6)=5.25;
pC3.p g l . and pg2. An optimal decision by fuzzy respectively. To maintain generator 3 in time interval
dynamic programming is the one with the highest 4.5, or 6, the reserve margins are 28, 9, and 15,
membership value of p~ at the last stage, The respectively, all of which are grater than the expected
recursive algorithm is expressed as: reserve margin Rm. So the membership values are:
Pg l(3.4) = 1.0
Pg1(3,5) = 1.0
pgl(3.6) = 1.0
Since the available manpower hours are
200(>120, assuming no other generator is undergoing
maintenance) pc1(3,4) = j ~ ~ i ( 3 . 5=) pc1(3,6) = 1.0.
So, the fuzzy decision for each state of this stage 1 is:
where: p~(1,T)= the highest membership value to pD(3,4)= min{p,1(3,4),p,2(3,4),p,,(3,4)) = 0.0
reach state (1.T);
(J) = set of feasible states at stage T-1. ~ ~ 3 .=
5 ~)n { p = ~ ( 3 . 5 ) , p = 2 ( 3 . ~ ) . p , 1=( 31.0
,5)}
The optimal decision for generator main@nance p0(3,6)= ~ n { p c 1 ( 3 . 6 ) . p c 2 ( 3 . 6 ) , ~ , ~ (=3 0.0
,6)}
scheduling is set of states (1.T) determined by the
following equation: Using the same process, the states of the
second (for generator 1) and the third (for generator 2)
are:

PD(1,l) = 0.0 ~ ~ ( 2 . =7 0.0

)
where: opt = set of optimal state (1,T); PD(1.2) = 0.5 ~ ~ ( 2 . =8 1.0
)
(T) = set of maintenance starting time PD(1,3) = 0.25 ~ ~ ( 2 .=90.0)
intervals;
(G) = set of generators to be maintained. According to the fuzzy decision of fuzzy
dynamic programming, the optimal generator
(3) A Sirn~leNumerical Example maintenance schcduling is:
Consider a 3 generator system with system
conditions and forecast peak load for each time interval
shown in Tables 2 and 3. For simplicity, production
cost and geographical constraint are neglected.
1684

where: o and ooare measured speed and set point,

respectively. Then , A h can be computed as:

- ACO(kT)
1)r]
ACO[(k i-
A b [ ( k + 1)T]= (24)
T
where: T is the sampling interval
Therefore, the optimal fuzzy scheduling is to . .
maintain generator 1 at time interval 2; generator 2 at 8; (2) Fuzzv Loeic Control Rules F u z z v Ouanuzauon)
and generator 3 at 5 . Each control rule consists of an "IF situation
In order to handle some of the non linearity THEN action" pair. The " I F portion of the rule is
which occurs in treating membership functions some called the premise or antecedent, while the "THEN" part
tuning scheme is appropriate for handling problems is the consequent. The premise defines a region of the
with non linearity. Examples of tuning can be found in input space wherein the rule is valid. This region is
control applications, signal processing and industrial characterized by a fuzzy set on the discourse of the input
applications. A good reference summary for tuning variable. The consequent defines the desired control
schemes is provided by M. Gupta in [106]. action to be taken for the input space delimited by the
premise.
Define the following fuzzy sets for input
5. A FUZZY POWER SYSTEM signals A W and AW: "large positive (LP)", "medium
STABILIZER(FPSS) positive (MP)", "small positive (SP)", "very small
(VS)", "small negative (SN)", "medium negative
Power system stabilizers (PSS) are necessary (MN)", and "large negative (LN)". According to the
to provide appropriate damping of undesirable control principle of the PSS, the control rules can be
oscillations caused by disturbances. Current PSS constructed based on the values of Ao and Ak. For
design can be categorized into two groups: example:
conventional PSS and adaptive PSS. Conventional
PSS lacks the ability to adapt to changes in the
system's operating condition. Adaptive controllers IF Am value is LP and A h value is LN,
demonstrate a certain effectiveness. However, most THEN the control action U should be VS.
adaptive PSS are based on linearized system models and,
therefore, cannot adapt to wide ranges of power system The above IF-THEN rule is a fuzzy description
operating conditions. Some AI based PSSs have shown of the control logic representing the human expert's
promise. qualitative knowledge. For different A m and A h
Fuzzy logic controllers belong to the category values, a set of fuzzy control rules similar to the above
of AI-based controllers. One advantage of fuzzy logic can be obtained. An example set of rules is shown in
over other forms of knowledge-based controllers lies in Table 4.
the interpolative nature of fuzzy control rules. The
overlapping fuzzy antecedents to the control rules Table 4 Fuzzy Logic Control Rules
provide transitions between the control actions of
different rules. Because of this interpolative quality,
fuzzy logic controllers usually require far fewer rules
than other knowledge-based controllers.

(1) Basic Considerations in PSS Design

The PSS is aimed at damping the oscillation
of generator speed (system frequency) to zero. I t is
reasonable to choose the generator speed deviation Am
and its derivative ACh as the input signals to the In order to properly describe the input signals,
controller. A o can be computed directly from real-time it is necessary to specify certain thresholds for Am and
measurements and the set point as: A h according to the expert's knowledge and to define
their membership values as is shown in Table 5 .
A&(?) = a(?)
- 0, (23) According to Table 4, the fuzzy relation matrix
can be constructed. However, since the fuzzy relation
matrix usually has very high dimension (49x7 for the
 1685

example here), it is inconvenient and unnecessary to determined from this table. Using these membership
store the relation mauix directly. values and the fuzzy relation matrix defied by equation
By virtue of basic results in theory of fuzzy (25). the following computation of fuzzy mauix gives
relational equations, the fuzzy relation of the control the labeled membership value:
rules is given by:
PU (4
(26)
= mYE aU [ min(p(Aw),p( Ah),R( Aw,Ah))]

According to Table 6, the fuzzy output can be

where: A = fuzzy set for Am; transformed into a crisp (defuzzified) control signal as
-0.8.
B = fuzzy set for ~ h ; An alternative way to defuzzify the output is
U = fuzzy set for input ui. by using a center of inertia method, which is defined as:
For appropriate control signals, we define crisp
values for the control output corresponding to different
fuzzy subsets LN. MN. SN, VS. SP, MP, or LP.
These crisp values are given in Table 6.

for continuous membership functions; or

Table 5 Membership values of Am and A#

for discrete membership functions.

It should be pointed out that the above
description of fuzzy PSS design is a simple explanation
of fuzzy controller design and cannot be directly used for
a practical power system stabilization without tuning.
Another point that so far has been ignored in the
Table 6 Crisp outputs w.r.t. fuzzy subsets literature is that multiple input signals (more than two)
can be used in the design of fuzzy controllers, that
makes a fuzzy controller capable of accommodating the
control of systems of greater complexity.
Additional illustrations of applications of fuzzy
set theory can be found in [13-161.

6. CONCLUSIONS
After the membership function values are
constructed appropriately for input and output signals The application of fuzzy set theory to power
(shown in Tables 5 and 6) and the fuzzy relation is systems is a relatively new area of research. while the
defined in equation (25), the fuzzy logic controller can interest in this field is increasing, more attention is
be used for power systems stability enhancement. It needed on basic principles and suitable problems for this
has been reported that fuzzy-logic-based PSS performs method. By using two simple numerical examples. the
better than conventional PID controllers (refer to the procedures of how LO use this new tool are shown. It is
literature listed in "applications in power system our expectation that the advantages of introducing fuzzy
control" in the Reference section). To date, extensive set methods to solve power systems problems has been
comparisons to adaptive methods have been made. Clarilied.
Next, we demonstrate the computation of the Based on the number of publications on the
output signals. Consider, the input signals are subject, power system control problems are the most
Am=-0.2 and A#=-0.1, from Table 5 they are identified popular areas for fuzzy set based approaches. In fact
as "medium negative (MN)" and "small negative (SN)" many achievements in the field of fuzzy control have
respectively, their membership values can also be been seen in other industries.
1686

It appears that fuzzy sets can play an important A. Kandel: Fuzzy Mathematical Techniques with
role in power system operation and planning Applications (1986), Addison-Wesley, Reading,
optimization problems. Several publications have used MA.
fuzzy sets to manage conflicting objectives and soft A. Kaufman and M. Gupta: Introduction to Fuzzy
Arithmetic (19852 Van Nostrand Reinhold.
constraints. This scheme not only makes the problem New York NY.
formulation more flexible, but if applied correctly, can H.J. Zimmermann: Fuzzy Sets, Decision Making
improve the computational efficiency. and Expert Systems (1987). Kluwer Academic
Finally, we conclude the paper with the Publishers.
statement that fuzzy set theory is a very useful G.J. Kiir and T.A. Folger: Fuzzy Sets, Uncertainfy
supplement to the conventional mathematical methods and Information (1988). Prentice-Hall.
in solving power systems problems and applying fuzzy A. Kaufmann and M. Gupta: Fuzzy Mathematical
sets is not to fuzzify the problem but to clarify a Models in Engineering and Management
problem that is fuzzy in nature through fuzzy set Sciences (1989), North Holland.
theory. J.C. Bezdek: Pattern Recognition with Fuzzy
Objective Function Algorithms (1981). Plenum
Press.
ACKNOWLEDGMENTS
Applications In Power System Planning a n d
The authors would like to express their Scheduling
gratification to the National Science Foundation for the
financial support under grant No. ESC-8657559,and S.B. Dhar: "Power System Long-Range Decision
Mr.Patrick Lusaka, who assisted in the preparation of Analysis under Fuzzy Environment". IEEE Trans.
this document. on PAS, Vol. PAS-98. No. 2. 1979.
A.K. David and R.D. Zhao: "An Expert System
with Fuzzy Sets for Optimal Planning", IEE
Trans. on Power Systems. Vol. 4, No. 3, Aug.
REFERENCES 1989.
A.K. David: "Expert System with F u u y Sets for
General Optimal Long-Range Operation Planning", IEEE
Trans. on Power Systems, Feb. 1991. pp. 59-65.
D. Dubois and H. Prade: Fuzzy Sets and Systems: V. Miranda and M.A. Matos: "Distribution
Theory and Applications (1980), Academic Press, System Planning with Fuzzy Models and
INC. Techniques". Proceedings of CIRED 1989,
W. Pedryoz: Fuzzy Control and Fuzzy System Brighton, Aug. 1989.
(1989). John Wiley & Sons, Inc. J.T. Saraiva, V. Miranda and L. Pinto: "Impact on
H.J. Zimmermann: Fuzzy Set Theory and Its Some Planning Decisions from a Fuzzy
Applications (1984), Kluwer-Nijhoff Publishing. Modeling of Power Systems", Proceedings of the
Constantin Virgil Negoita: Expert Systems and 1993 PICA, Scottsdale, Arizona, May 4-7, 1993.
Fuzzy Systems (1984). The Benjamin/Cummings pp. 327-333.
Publishing Company, INC. V. Miranda: "Using Fuzzy Reliability Indices in a
W.H. Janko. M. Roberts and H.J. Zimmermann: Decision Aid Environment for Establishing
Progress in Fuzzy Sets and Systems (1990), Interconnection and Switching Location Policies,
Kluwer Academic Publishers. Proceedings CIRED PI. AIM-IEE. s.6.21. Liege,
J. Dombi: "A General Class of Fuzzy Operators, Belgium, April 1991.
the DeMogio Class of Operators and Fuzziness H.J. Haubrich, T.H. Seitz and A. Bovy: "Fuzzy
Induced by Fuzzy Operators", Fuzzy Sets and Sets in Reliability Analysis of Power
Systems, 11, 115 (1983). Distribution Systems". I993 Expert System
J. Dombi: "Membership Function as An Applications to Power Systems, Melbourne. Jan.
Evaluation". Fuzzy Sets and Systems, 35. 1 1993. pp. 212-217.
(1 990). K. Tomsovic, G.Skugge, D. Sjelvgren and J.
M. Gupta and J. Qi: "Theory of T-norms and Bubenko: "Seasonal Planning for Best Fuel Mix
Fuzzy Interference Method, Fuzzy Sets and Under Multiple Objectives Using Fuzzy
Systems, 40. 431 (1990). Mathematics", 1993 Expert Systems
H. Prade: "A Computational Approach to Applications to Expert Systems. Melbourne, Jan.
Approximate and Plausible Reasoning with 1993, pp. 112-118.
Applications to Expert Systems", IEEE Trans. on Z.G. Hu: "Power System planning with Fuzzy
Pattern Analysis and Machine Intelligence, Programming". 1993 Expert System
PAMI-7. 3 (1985-6). Applications io Power Systems, Melbourne. Jan.
L A . Zadeh: "Fuzzy Sets", Information and 1993. pp. 107-111.
Control, 8, 338 (1965). V. Miranda and J.T. Saraiva: "Fuzzy Modeling
of Power System Optimal Power Flow", IEEE
 1687

Tram. on Power Systems, Vol. 7, No. 2, May Startup Scheduling and Operation Support in
1992. Fossil Power Plants", 1993 Expert Systems
V. Miranda. M. Matos and J.T. Saraiva: "Fuzzy Applications to Power Systems, Melbourne. Jan.
Load Flow - New Algorithms Incorporating 1993. pp. 593-599.
Uncertain Generation and Load Representation". K. Tomsovic and J.M. Ling: "A Proposed Fuzzy
Proceedings of the 10th PSCC, Graz, Aug. 1990, Set and Measurement Approach to Monitoring
pp. 621-627. and Coneol of Real Power", IASTED Power High
C.E. Lin. C.J. Huang, C.L. Huang and S.Y. Lee: Tech '91. Tainan. Taiwan, Mar. 1991.
"An Optimal Generator Maintenance. Approach R. Yokoyama, T. Niimura and Nakanjahi: "A
Using Fuzzy Dynamic Programming". 1992 Coordinated Control of Voltage and Reactive
IEEEIPES Summer Meeting, 92 SM 401-0 PWRS, Power by Heuristic Modeling and Approximate
Seattle, July 1992. Reasoning", 1992 IEEEIPES Winter Meeting, 92
Y. Terasawa and S. Iwamoto: "Optimal Power WM 114-9 PWRS. New York, Feb. 1992.
Flow Using Fuzzy Mathematical Programming", G. L. Torres, L.E.B. Da Silva and D. Mukhedkar:
Electrical Engineering in Japan, Vol. 108, No. 3, Distribution Feeder Load Modeling: An
1988. pp. 46-54. Approach Using Fuzzy Sets". European
K.H. Abdul-Rahman and S.M. Shahidehpour: Conference on CircuU Theory and Design,
"Application of Fuzzy Sets to Optimal and ECCTD-89, IEE Symposium, Brighton, U.K.,
Reactive Power Planning with Security Sept. 1989. pp. 639-643.
Constraints". Proceedings of 1993 PICA, G. L. Torres, L.E.B. Da Silva and D. Mukhedkar:
Scottsdale. Arizona, May 4-7. 1993, pp. 124- Fuzzy Load Management in Distribution
130. Systems", IASTED Power High Tech '89,
Valencia, Spain, June 1989, pp. 425-430.
Applications i n Power Systems Operation G. L. Torres, L.E.B. Da Silva and D. Mukhedkar:
A Fuzzy Knowledge Base for Load Forecasting",
Y.Y. Shu and H.C. Kuo: "Fuzzy Set Based 1991 Expert System Applications io Power
Contingency Ranking", IEEE Trans. on Power Systems. TokyoKobe. April 1991. pp. 439-446.
Systems, Vol. I. No. 3. Aug. 1992. G. L. Torres, L.E.B. Da Silva and D. Mukhedkar:
K. Tomsovic: "A Fuzzy Linear Programming A Fuzzy Knowledge Based System for Bus Load
Approach to the Reactive PowerNoltage Control Forecasting". 1992 IEEE International
Problem", IEEE Trans. on Power System, Conference on Fuzzy System, Mar. 8-12. 1992,
Nov. 7. No. 1, Feb. 1992, pp. 287. San Diego, California, pp. 1211-1218.
K.H. Abdul-Rahman and S.M. Shahidehpour: "A K.L. Ho. Y.L. Hsu and F.F. Wu: "Fuzzy Expert
Fuzzy Based Optimal Reactive Power Control", Systems: An Application to Short Term Load
Paper #92 WM 114-9 PWRS. IEEEIPES 1992 Forecasting", 1992 IEE Proceedings of
Summer Meeting, Seattle, WA. July 1992. Generation. Transmission and Distribution, PL
K. Tomsovic and J.M. Ling: "A Proposed C. Vol. 139, Nov. 1992, pp. 471-477.
Information Approach to Power System E. Economakos: "Application of Fuzzy Concepts
Security", In ESAPS-91, pp. 422-432. to Power Demand Forecasting". IEEE Trans. SMC,
J.L. Souflis. A.V. Machias and B.C. Papadias: Vol. 9. 1979, pp. 651-657.
"An Application of Fuzzy Concepts for Transient H.C. Kuo and Y.Y Hsu: "Distribution System Load
Stability Evaluation", IEEE Trans. on Power Estimation and Service Restoration Using a Fuzzy
Systems, Vol. 4. No. 3. Aug. 1989. Set Approach". 1992 IEEEIPES Winter Meeting,
X.W. Ma: "Fuzzy Pattern Recognition of Power 93 WM 086-9 PWRD. Columbus, Feb. 1993.
System Transient Stability". International K.D. Labuda and S.M. Shahidehpour: "A Fuzzy
Conference on Advances in Power System Multi-Objective Approach to Power System State
Control, Operation and Management, Estimation", 1993 Expert Systems Applications
Hong Kong, Nov. 1991. to Power Systems. Melbourne. Jan. 1993, pp.
A.V. Machias and J.L. Souflis: "A Fuzzy 21 8-223.
Approach in the Determination of Unstable A.V. Machias and G.D. Shikos: "Fuzzy Risk
Machine". IEE Proceedings, Vol. 137, Pt. C. No. Index of Wind Sites", 1992 IEEEIPES Winter
2, Mar. 1990. Meeting, 92 WM 023-2 EC, New York, Jan.
M. Maios. J. Lopes and F. Barbosa: "Fuzzy 1992.
Classification for On-Line Transient Stability T. Mackaw. K. Yamuda and R. Yokohama: "Fuzzy
Assessment of Electric Power Systems", Coordination of Economy and Quality in Power
Proceedings of the 10th PSCC, Graz, Australia, System Generation Control", The Third
Aug. 1990, pp. 865-871. International Conlerence on Power System
H. Matsumoto, et al.: "An Expcrt System Based Moniforing ond Control, London. June 1991. pp.
on Fuzzy Reasoning for Startup Scheduling and 174-179.
Operation Support in Fossil Power Plants". T. Mackaw. K. Yamuda and R. Yokohama, H.
IEEEIPES 1991 Winter Meeiing, 91 WM 127-1 Obtsuki and Y. Mitzukami: "Fuzzy Coordination
EC, 1991. New York. of Multi-Objective Reflecting Operator's
C. Tingusly, A. Germond and C.C. Liu: "An Intention for Dynamical Generator
Expert System Based on Fuzzy Reasoning. for Rescheduling", 1991 Expert System Applications

I
 to Paver Systems, Tokyo/Kobe, April 1991, Based Self-Tuned Power System Stabilizer“,
pp. 470-477. IEEEIPES 1992 Summer Meeting. Seattle. WA.
H. Matsumoto, et al.: “An Operation Support July 12-16, 1992, Paper No. 92 SM 474-7 EC
Expert System Based on On-Line Dynamics M.A.M. Hassan, O.P. Malik and G.S. Hope: “A
Simulation and Fuzzy Reasoning for Startup Fuzzy Logic Based Stabilizer for a Synchronous
Schedule Optimization in Fossil Power Plants”. Machine”, IEEE Trans. on Energy Conversion,
1992 IEEEIPES Winter Meeting, 93 WM 143-8 Vol. 6, No. 3, Sept. 1991.
EC. Columbus, Feb. 1993. T. Hiyama. T. Sameshima and C.M. Lim:
T. Niimura and R. Yokoyama: “An Approximate “Fuzzy Logic Stabilizer with Digital
Reasoning Approach for Optimal Dynamic Compensation for Stability Enhancement of
Dispatch of Thermal Generating Units Including Multi-Machine Power System“, Third
Auxiliary Control”, IEEE Trans. on Power Symposium on Expert System Application to
Systems, Vol. 6, No. 2, May 1991, pp. 651-657. Power Systems, April 1-5, 1991. Tokyo/Kobe,
T. Niimura, Y. Nakanishi, K. Yasuda and R. Japan.
Yokoyama: “Multi-Attribute Voltage/Reactive A. Ishigame, T. Imoto, S. Kawamoto and T.
Power Control Based on Approximate Tanihuchi: “Fuzzy and Optimal Control for
Reasoning”, 1991 Expert System Applications to Stabilizing Power Systems“, Electrical
Power Systems, TokyoKobe, April 1991, pp. Engineering in Japan, Vol. 1 1 1 , No. 2, 1991.
462-469. T. Hiyama and C. M. L m : “Application of Fuzzy
0. Satoh. T. Oyama and Y. Serizawa: “Fuzzy Logic Control Scheme for Stability Enhancement
Dccision-Making of Restoration Target”, 1993 of a Power System“, IFAC Symposium on Power
Expert System Application to Power Systems, Systems and Power P l m Control, Aug. 1989.
Melbourne, Jan. 1993, pp. 582-587. Singapore.
C.C. Su and Y.Y. Hsu: ‘‘Fuzzy Dynamic M.A.M. Hassan. O.P. Malik and G.S. Hope: “A
Programming: An Application to Unit Fuzzy Logic Based Self-Tuned Power System
Commitment”, IEEE Trans. on Power Systems, Stabilizer”, The Third International Conference
Vol. 6. No. 3, Aug. 1991. on Power System Monitoring and Control,
C.N. Lu, R.C. Leou, K.C. Liu and M. Unum: “A Sponsored by IEE, London, U.K., June 1991.
Load Flow and Fuzzy Linear Programming Based T. Hiyama and S. Samasima: “Fuzzy Logic
External Network Modeling Approach”. Control Scheme for On-Line Stabilization of
IEEEIPES 1993 Summer Meeting. Vancouver, Multi-Machine Power Systems“. Fuzzy Sets and
B.C., Canada, July 18-22 1993, 93 SM 551-2 EC. Systems, Vol. 39, Mar. 1991, pp. 181-194.
K.H. Abdul-Rahman. S.M. Shahidehpour: C.S. Chang. Y .C. Liang S. Biangovan: “Fuzzy
Reactive Power Optimization Using Fuzzy Load Logic Control of Static VAR Sources for
Representation”, IEEE Trans. on Power Systems. Enhancing Transient Stability of Interconnected
Accepted for publication to appear in 1993. Power Systems“, I993 Expert System
K.H. Abdul-Rahman, S.M. Shahidehpour: Applications to Power Systems. Melbourne,
“Optimal Reactive Power Dispatch with Fuzzy Jan. 1993. pp. 399-405.
Variables”, 1993 IEEE International Symposium L. Chao and B.T. Kunnel: “Dealing with
on Circuits and Systems, Chicago, IL, May Uncertainty and Fuzzy Control of PCSEX”. 1993
1993, pp. 2188-2191. Expert System Applications to Power Systems,
C. Wang, S.M. Shahidehpour and R. Adapa: “A Melbourne. Jan. 1993. pp. 412-417.
Fuzzy Set Approach to Tie Line Flow D. Chui and M.A. Laughton: “Power System
Computation in Optimal Power Scheduling”, Control Using Fuzzy Reasoning“. Proceeding of
1992 American Power Conference, Chicago, IL, the 10th Power System Computation
April 1992. pp. 1443-1450. Conference,, Graz. Australia, Aug. 1990.
T. Hiyama: ”Real Time Control Micro-Machine
Applications in Power System Control System Using Micro-Computer Based Fuzzy
Logic Power System Stabilizer“, 1992 IEEEIPES
[63] Y.Y. Hsu: “Design of Fuzzy Power System WiMer Meeting. 92 WM 128-9 EC. Columbus,
Stabilizers for Multi-Machine Power Systems”, Feb. 1993.
IEE Proceedings, Pt. C. Vol. 137, No. 3, May T. Hiyama. S . Samasima and C.M. Lim: “Fuzzy
1990. Logic Stabilizer with Digital Compensation for
[ 641 F. Cheng and S. Yeh: “Application of Fuzzy Stability Enhancement of Multi-Machine Power
Logic in the Speed Control of AC Servo System System“, Third Symposium on Expert System
and an Intelligent Invertor”, Proceedings of IEEE Application to Power Systems, April 1-5, 1991.
92 Summer Meeting. Paper No. 92 SM 555-3 EC. Tokyo/Kobe. Japan, pp. 455-461.
[65] E. Handschin. W. Hoffman, F. Royer and Y.Y. Hsu and C.H. Cheng: ”Load Frequency
Stephanblome: “A New Method of Excitation Control Using Fuzzy Logic”, IASTED Power High
Control Based on Fuzzy Set Theory”, Proceeding Tech ’91, Tainan. Taiwan, Mar.1991.
of the 1993 IEEE PICA, Scottsdale. Arizona, May M. Nerezian, G . Anderson and K. Tomsovic:
4-7, 1993, pp. 34-40. “Power System Damping Control Design with
[66] M.A.M. Hassan and O.P. Malik: “Implementation Fuzzy Logic“, 1993 Expert System Applications
and Laboratory Test Results for a Fuzzy lQgic to Power Systems, Melbourne. Jan. 1993, pp.
 1689

406-4 1 1. 1993 Canadian Conference on Electrical and

L.E.B. Da Silva, G.L. Torres, G. Oliver, G.E. Computer Engineering, Sp. 14-17, 1993
April and D. Mukhedar: "A Fuzzy-Forth Rule Base Vancouver, Canada
Production System for Power Converter Control". P. Jervanhaura. P. Verho and J. Partanen:
1991 Expert System Applications to Power "Using Fuzzy Sets to Model the Uncertainty in
Sysfems, Tokyo/Kobe, Japan, April 1991, pp. the Fault Location Process of Distribution
447-454. Networks", IEEEPES 1993 Summer Meeting,
A. Ishigame, T. Ueda, S. Kawamoto and T. Vancouver, B.C.. Canada, July 18-22 1993. 93
Taniguchi: "Design of Electric Power System SM 4168 PWRD
Stabilizer Based on Fuzzy Control Theory", I992
IEEE International Conference on Fuzzy System, Others
Mar. 8-12, 1992. San Diego, California. pp. 973-
980. K. Tomsovic: "Current Status of F u u y Set Theory
A.M. Sharaf and T.T. Lie: "A Fuzzy Logic Rule Applications to Power Engineering Problems", T.
Based Position Controller for PM-Dc Motor IEE Japan, Vol. 113-b. No. 1. 1993.
Drives", 1993 Canadian Conference on Electrical LA. Zadeh: Interpolative Reasoning in Fuzzy
and Computer Engineering, Sept.. 14-17. 1993. Logic and Neural Network Theory". 1992 IEEE
Vancouver, Canada. lntertdonal Conference on Fuzzy Systems, Mar.
H.M. Mashaly and A.M. Sharaf: "Fuzzy Logic 8-12 1992, San Diego, California.
Controller for Maximum Power Tracking in Line- K.B. Cho: "A New Learning Fuzzy Logic
Commutated Photovoltaic Inverter Scheme", Controller (LFLC) Using Artificial Neural
1993 Canadian Conference on Electrical and Network for Inverter Fed Variable Speed Drive".
Computer Engineering, Sept.. 14-17, 1993. 1993 Expert System Applications to Power
Vancouver, Canada. S y s t e m , Melbourne, Jan. 1993, pp. 588-592.
T. Hiyama: "Robustness of Fuzzy Logic Power T. Iwata. K. Machida and Y. Toda: "Fuuy Control
System Stabilizers Applied to Multi-Machine Using Neural Network Technique", Proceedings of
Power Systems", IEEEIPES Summer Meeting. International Conference on Neural Networks,
Vancouver. B.C.. Canada, July 18-22 1993, 93 Vol. 3, 1990.
SM 551-2 EC. K. Saitoh and S. Iwamoto: "Application of Neural
Network Based Fuzzy Control to Power System
Applications in Power System Diagnosis Generators", 1991 ANN Applications to Power
Sysfems. Seattle, July 1991, pp. 144-148.
H.E.D. Dijk: "Exformer: An Expert System Z Z . Zhang. G.S. Hope and O.P. Malik: "Expert
Stabilizer for Transformer Fault Diagnosis", Systems in Electric Power Systems - A
Proceeding of the 9th Power System Computation Bibliographical Survey". IEEE Trans. on Power
Conference, Lisbon, Sept. 1987, pp. 715-721. Systems, Vol. 4. No. 4. Oct. 1989.
C.E. Lin. J.M. Ling and C.L. Huang: An Expert T. Taylor and D. Lubkeman: "Applications of
System for Transformer Fault Diagnosis and Knowledge Based Programming to Power
Maintenance Using Dissolved Gas Analysis", Engineering problems", IEE Tram. on Power
1992 IEEEIPES Winter Meeting, 92 WM 243-6 Systems, Vol. 4. No. 1, Feb. 1989.
PWRD, New York, July 1992. H.H. Zurn and V.H. Quintana: "Generator
K. L. Tomsovic. M. Topper and T. Ingvarec: Maintenance Scheduling via Successive
"A Fuzzy Information Approach to Integrating Approximation Dynamic Programming", IEEE
Different Transformation Diagnostic Methods", Trans., PAS, Vol. PAS-94, No. 2. Mar./April
1992 IEEEPES Winter Meeting, 92 WM 243-5 1975. pp. 665-671.
PWRD, New York, July 1992 J.E. Dopazzo and H.M. Merrill: "Optimal
J. Xu, S. Zhang and y. Liu: "Fuzzy Diagnosis Maintenance Scheduling Using Integer
Method in Machinery Condition Monitoring Programming", IEEE Trans. on PAS, Vol. PAS-
and Failure Diagnosis". Proceedings of the 2nd 94. No. 5, SeptJOct. 1975. pp. 1537-1545.
International Machinery Monitoring and W.K. Christiane and A.H. Palmer: "A Technique
Diagnostic Conference, Los Angeles. Sept. for the Automated Scheduling of the Maintenance
1990. pp. 227-233 of Generating Facilities", IEE T i m . on PAS,
K. Choi, K. Nimblya and J. Hasagawa: "Fuzzy Vol. PAS-91, No. 2, Feb. 1975, pp. 137-144.
Decision Making of Deloing Countermeasures R.N. Adams and N. Kagan: "Electrical Power
Against Snow Accretion", 1991 Expert Distribution Systems Planning using Fuzzy
System Applications IO Power Systems, Mathematical Programming". 1993 PSCC. pp.
Tokyo/Kobe, April 1991, pp. 433--438. 163-172,
J. J. Dukarm: "Transformer Oil Diagnosis R. Tanabe, K. Yasuda and R. Yokoyama: "An
Using Fuzzy Logic and Neural Networks", Algorithm for the Optimal Generation Mix
1993 Canadian Conference on Electrical and Accounting for Fuzziness of Both Decision
Computer Engineering, Sp. 14-17, 1993 Making and Planning Parameters". 1993 PSCC.
Vancouver, Canada pp. 593-600.
J. L. Chen: "A Fuzzy Expert System for Fault I. Hela], J. Smolinski, A.M. Sharaf. K.S. Brown,
Diagnosis in Electric Distribution System". W.K. Marshall, B.E. Milton. E.F. Hill:
1690

"Prospective Use of Fuzzy Logic in Power System Engineering Education, and the American Association for
Reliability". 1993 PSCC, pp. 839-846. Artificial Intelligence.
[ 1061 M. Gupta: Fuzzy Computing, Theory, Hardware was born in China in 1962. He
and Applicufion (1988). North Holland. received his B.S. degree from Hafai University of
New York, NY. Technology, P.R. of China, in 1983 and the M.S. degree
[ 1071 T. Terano. K.Asai and M. Sugeno: Fuzzy Sysfems from the Graduate School, Electric Power Research Institute
Theory and Its Applications (1987). Academic [EPRI(CHINA)]. P.R. of China, in 1985. respectively, all
Press, San Diego, California. in Electrical Engineering. He is presently doing research
towards the Ph.D. degree at Howard University in the areas
BIOGRAPHIES of dynamic security assessment, optimal power flow and
applications of artificial intelligence techniques to power
A. Momoh received the B.S.E.E. degree systems.
from Howard University in 1975. the M.S.E.E. degree from Mr. Ma worked as an engineex at EPRI(CHINA)
Camegie Mellon University in 1976. the M.S. degree in from 1985 to 1992 and was engaged in the research of
systems engineering from the University of Pennsylvania power system computerized control, m particular in the
in 1980 and the Ph.D. in Electrical Engineering from development of Energy Management System and Operator
Howard University in 1983. Training Simulator software.
Dr.Momoh is a professor in the Electrical
Engineering Department at Howard University and also the &vin Tomsovic was born in Slidell, LA, in
director of the program in power systems engineering. 1963. He received the B.S. from Michigan Tech
Previously, he has been a member of the technical staff at University, Houghton, in 1982. and the M.S. and Ph.D.
Bell Laboratories, and a research scientist at Westinghouse degrees from University of Washington, Seattle, in 1984
Research and Development. His current research activities and 1987. respectively. all in Electrical Engineering. He
are concentrated on stability analysis. system security and had held positions at McDonnell Douglas Corp., IBM
expert systems design for utility firms and government Corp., and MIT Lincoln Laboratory. Visiting and adjunct
agencies. e has published more than 20 technical papers in appointment include Boston University. National Cheng
controls, systems theory and power systems engineering. Kung University and National Sun Yat-Sen University in
In 1986 and 1987. Dr. Momoh received the Taiwan, and the Royal Institute of Technology in Sweden.
departmental outstanding teacher award and in 1987 Currently he is with Washington State University in
received a national Science Foundation Presidential Young Pullman, WA. His research interest include expert systems,
Investigator Award. and fuzzy set applications to power system control and
Dr.Momoh is an active member of the Power security.
Engineering Society of the IEEE, the American Society for