Energy Conversion and Management 50 (2009) 344–353

Contents lists available at ScienceDirect

Energy Conversion and Management

journal homepage: www.elsevier.com/locate/enconman

Load frequency control strategies: A state-of-the-art survey for the researcher

H. Shayeghi a,*, H.A. Shayanfar b, A. Jalili c
a
Technical Engineering Department, University of Mohaghegh Ardabili, Ardabil, Iran
b
Center of Excellence for Power Automation and Operation, Electrical Engineering Department, Iran University of Science and Technology, Tehran, Iran
c
Islamic Azad University, Ardabil Branch, Ardabil, Iran

a r t i c l e i n f o a b s t r a c t

Article history: Global analysis of the power system markets shows that load frequency control (LFC) is one of the most
Received 13 February 2008 profitable ancillary services of these systems. This service is related to the short-term balance of energy
Accepted 12 September 2008 and frequency of the power systems and acquires a principal role to enable power exchanges and to pro-
Available online 4 November 2008
vide better conditions for electricity trading. The main goal of the LFC problem is to maintain zero steady-
state errors for frequency deviation and good tracking of load demands in a multi-area power system.
Keywords: This paper provides an overview of control strategies for researchers, as well as of their current use in
LFC
the field of LFC problems. The history of control strategies is outlined. Various control methodologies
Restructured power system
Automatic generation control
based on the classical and optimal control, robust, adaptive, self-tuning control, VSC systems, digital
Robust and adaptive control and artificial intelligent/soft computing control techniques are discussed. We make various comparisons
Intelligent/soft computing control strategy between these approaches, and the main advantages and disadvantages of the methods are presented.
Power system control Finally, the investigations of the LFC problem incorporating BES/SMES, wind turbines and FACTs devices
have also been discussed.
Ó 2008 Elsevier Ltd. All rights reserved.

1. Introduction Nowadays, the electric power industry is in a transition from a

The successful operation of interconnected power systems re-
quires matching the total generation with the total load demand
and with the associated system losses. With time, the operating
point of a power system changes, and hence, systems may experi-
ence deviations in nominal system frequency and scheduled power
exchanges to other areas, which may yield undesirable effects [1].
LFC or AGC is one of the most important issues in electric power
system design and operation for supplying sufficient and reliable
electric power with good quality. The main objectives of LFC for
a power system are
 Ensuring zero steady-state error for frequency deviations.
 Minimizing unscheduled tie line power flows between neigh-
boring control areas.
 Getting good tracking for load demands and disturbances.
 Maintaining acceptable overshoot and settling time on the fre-
quency and tie line power deviations.
Based on the above objectives, the two variable frequencies and
the tie line power exchanges are weighted together by a linear
combination to form a single variable called ACE, which is used
as the control signal in the LFC problem.
* Corresponding author. Tel.: +98 451 5512910; fax: +98 451 552904.
E-mail address: hshayeghi@gmail.com (H. Shayeghi).
vertically integrated utility scenario, where a single utility owned
and operated the generation, transmission and distribution sys-
tems and provided power at regulated rates, to the deregulated
scenario, where competitive companies sell unbundled power at
lower rates. Furthermore, various kinds of apparatuses with large
capacity and fast power consumption, such as testing plants for
nuclear fusion and steel factories, increase significantly. When
these loads are concentrated in power systems, they may cause a
serious problem of frequency oscillations. Thus, it is very impor-
tant to consider how the control services of system frequency
should be implemented. In a deregulated environment, any power
system control, such as LFC as an ancillary service, acquires a prin-
cipal role to maintain the electric system reliability at an adequate
level, and is becoming much more significant today in accordance
with the complexity of interconnected power systems [2,3]. Thus,
stabilization of frequency oscillations in an interconnected power
system becomes challenging when implemented in the future
competitive environment. A new frequency stabilization service
that emphasizes not only efficiency, reliability and economics but
also advanced and improved controls for satisfying the require-
ments of power system operation is much in demand.
The LFC problem has been augmented with valuable research
contributions from time to time, such as LFC regulator designs to
cope with parameter variations uncertainties, load characteristics,
excitation control and parallel ac/dc transmission links. The micro-
processor-based LFC Controller, robust controller, self-tuning and

0196-8904/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.enconman.2008.09.014
 H. Shayeghi et al. / Energy Conversion and Management 50 (2009) 344–353
Nomenclature
ACE area control error
AFRC automatic frequency ratio control
AGC automatic generation control
AI artificial intelligence
ANN artificial neural network
BES battery energy storage
CES capacitive energy storage
DISCO distribution companies
FD figure of demerit
GAs genetic algorithms
GENCOs generation companies
GRC generation rate constraint
HVDC high voltage direct current
IGBT insulated gate bipolar transistor
ISE integral square error
adaptive controllers designs have also been presented. The most
recent advance in this area is the application of concepts such as
neural networks, fuzzy logic and genetic algorithms to tackle the
difficulties associated with the design of LFC controllers for power
systems with nonlinear models and/or insufficient knowledge
about the system required for its accurate modeling. Apart from
advances in control concepts, there have been many changes dur-
ing the last decade or more, such as deregulation of the power
industry and use of superconducting magnetic energy storage,
wind turbines and photovoltaic cells as other sources of electrical
energy to the system. Because of these, the control philosophies
associated with the LFC problem have changed to accommodate
their dynamics and their effects on the overall system dynamic
performance. Generally, the methodologies of LFC controller de-
signs can be categorized as (i) classical methods, (ii) adaptive and
variable structure methods, (iii) robust control approaches and
(iv) AI-based methods.
In this study, the types of power system models for LFC, digital
LFC schemes and the history of various control strategies with their
salient features are outlined.
2. Power system LFC models
The LFC problem has been dealt with extensively for more than
three decades. The power systems are usually large-scale systems
with complex nonlinear dynamics. However, the major part of the
work reported so far has been performed by considering linearized
models of two/multi-area power systems [1,4–8]. The effect of
GRCs was included in these types of studies, considering both con-
tinuous and discrete power system models [9,10]. The first attempt
in the area of LFC problems has been to control the frequency of a
power system via a flywheel governor of the synchronous machine.
This technique was subsequently found to be insufficient, and a
supplementary control was included to the governor with the help
of a signal directly proportional to the frequency deviation plus its
integral. This scheme constitutes the classical approach to the solu-
tion of the LFC problem. Aggarwal et al. [11] and Cohn [12] have
illustrated that supplementary controller designs based on tie line
bias control strategy are the reason that the ACEs are regulated to
zero effectively. The standard definitions of the terms associated
with LFC of power systems were finalized in Ref. [13]. Following
that, suggestions for dynamic modeling for LFC are discussed thor-
oughly in Refs. [13–15]. Based on the experiences with actual
implementation of AGC schemes, modifications to the definition
of ACE are suggested from time to time to cope with the changing
power system environment [16–19]. Since many presently regu-
345
ITAE integral of time of absolute error
LFC load frequency control
LMI linear matrix inequality
LQG linear quadratic Gaussian
MSF multi-stage fuzzy
PI proportional plus integral
PID proportional, integral and derivative
PV photovoltaic
RBF radial biased function
SA simulated annealing
SMES super conducting magnetic energy storage
SVC static var compensator
TRANSCO transmission companies
VSC variable structure controller
VSS variable structure system
lated markets are likely to evolve into a hybrid scheme and some
deregulated markets are already of this type (e.g. Norway), the ef-
fects of deregulation of the power industry on LFC have been ad-
dressed in Ref. [20].
In deregulated power systems, the vertically integrated utility
no longer exists. However, the common LFC objectives, i.e. restor-
ing the frequency and the net interchanges to their desired values
for each control area, still remain. The deregulated power system
consists of GENCOs, TRANSCOs and DISCOs with an open access
policy. In the new structure, GENCOs may or may not participate
in their own or other areas. Thus, various combinations of possible
contracted scenarios between DISCOS and GENCOS are possible. All
the transactions have to be cleared by the independent system
operator or other organizations. Because of these, a study on sim-
ulation and optimization in a LFC system after deregulation has
been performed by Donde and Pai [21]. The concept of the DISCO
participation matrix is proposed that helps for implementation of
the contracts. Later, a generalized dynamical model for a LFC
scheme was developed in the deregulated environment by Shay-
eghi et al. [22]. Based on the idea presented in Ref. [21], the con-
cept of an augmented generation participation matrix (AGPM) to
express the effect of possible contracts in the generalized model
was proposed. The AGPM shows the participation factor of a GEN-
CO in the load following contract with a DISCO. The rows and col-
umns of the AGPM equal the total number of GENCOs and DISCOs
in the overall power system, respectively. The proposed general-
ized model helps in visualization of the contracts and introduces
added new information signals on the traditional LFC scheme dy-
namic which carry information as to which GENCO has to follow
a local demand by which DISCO.
Small signal analysis is justified for studying the systems re-
sponse for small perturbations. However, implementation of a
LFC strategy based on a linearized model of an essentially nonlin-
ear system does not necessarily ensure the stability of the systems.
Considerable attention has been paid by researchers to consider
the system nonlinearities [23–26]. Tripathy et al. [26] demon-
strated the destabilizing effect of governor dead band nonlinearity
on the conventional AGC system. It is shown that the governor
dead band nonlinearity tends to produce continuous oscillations
in the area frequency and tie line power transient response.
3. Control strategies
In many reported works on the LFC area, control schemes based
on a centralized control strategy are used for solution of the LFC
problem [27–40]. The main limitation of the works presented on
346 H. Shayeghi et al. / Energy Conversion and Management 50 (2009) 344–353
LFC considering a centralized control strategy is the need to ex-
change information from control areas spread over distantly con-
nected geographical territories along with their increased
computational and storage complexities. The decentralized LFC
concept appeared in the power system control scenario to deal
with such problems very effectively, and consequently, many re-
search papers using this concept with continuous and discrete time
system models have been presented in the literatures [41–48]. In
Ref. [44], the authors have examined the structural properties of
observability and controllability for a class of interconnected
power system models. The proposed scheme provides for complete
decentralization of a global state feedback control policy in the
sense that the area control feedback loops are completely decou-
pled. Again, a class of systematically distributed control design
methods based on (i) distributed implementations of centralized
control systems, (ii) model reduction of dynamical systems and
(iii) modeling of the interactions between the subsystems compris-
ing the global control system is presented in Ref. [44]. The salient
feature of the design is to achieve almost identical results as those
obtained with the centralized design. It should be noted that in the
dynamical operation of power systems, it is usually important to
aim for decentralization of the control actions to individual areas.
This aim should coincide with the requirements for stability and
load frequency scheduling within the overall system. In a com-
pletely decentralized control scheme, the feedback controls in each
area are computed on the basis of measurements in that area only.
This implies that no interchange of information among areas is
necessary for the LFC task. The advantage of this operating philos-
ophy is apparent in providing cost saving in data communications
and in reducing the scope of network monitoring. Because of these,
the design of decentralized load frequency controllers is based on
structured singular value and H1 norm [47–49] and [50–53]. Yang
et al. [47,48] and Shayeghi and Shayanfar [50,51] have demon-
strated that when the frequency response-based diagonal domi-
nance cannot be achieved, the structured singular values and H1
norm can be applied to design the decentralized LFC to achieve
the desired system dynamic performance, respectively, in such a
way that the stability of the overall system with the decentralized
controllers is guaranteed. In Ref. [52], using the Lyapunov function
it was illustrated that the overall system was asymptotically stable
for all admissible plant parametric uncertainties when all local
controllers were working together. Kazemi et al. [53] introduced
a suitable transformation matrix that transformed the initial refer-
ence model to an equivalent reference model, such that the conver-
gence of the output errors was guaranteed. An appropriate
adaptive law was derived for adjusting this transformation matrix.
Various LFC schemes, based on two-level [54] and multi-level
[55–57] control concepts have been reported in the literatures. A
two-level suboptimal controller has been suggested by Wang
et al. [54]. However, this approach does not ensure zero steady-
state error, and hence, a multi-level finite time optimal controller
design, ensuring zero steady-state error, has been reported in
Ref. [55]. The advantage of the hierarchical structure is reflected
in the fact that even if one of the control levels fails, the system re-
mains in operation. A global controller, which also exploits the pos-
sible beneficial aspects of the interconnections, has been applied
for the LFC problem [56], and favorable results have been achieved.
The reduction of control efforts required in LFC of the intercon-
nected power systems is sought with the help of a singular pertur-
bation approach. This can be achieved by decomposing the system
into slow and fast subsystems, designing controllers separately for
each of the subsystems and then combining the controllers to yield
a composite controller. Investigations on LFC of large power sys-
tems using this approach are available in the literature [57]. The
separate controllers were designed for slow and fast subsystems
and were combined in such a way that the slow subsystem always
interacts with only one of the fast subsystems at a time. The study
also involves the effect of parameter variations and GRCs.
4. Control technique
Generally, LFC control design methodologies can be categorized
as (i) classical methods, (ii) adaptive and variable structure meth-
ods, (iii) robust control approaches, (iv) intelligent techniques and
(v) digital control schemes.
4.1. Classical methods
Conventional control strategies for the LFC problem are those
that take the integral of the control error as the control signal. In
the classical control methodologies, to obtain the desired gain
and phase margins, Bode and Nyquist diagrams as well as root lo-
cus are usually used. Thus, the design procedure of the classical
methods for the LFC problem is straight forward, easy and amena-
ble for practical implementation. However, the investigations con-
ducted using these approaches reveal that they exhibit poor
dynamic performance, especially in the presence of other destabi-
lizing effects, such as parameter variations and nonlinearities [16]
and [58–60].
Modern optimal control theory-based LFC schemes have ap-
peared in the literatures [61–67]. The feasibility of an optimal
LFC scheme requires the availability of all state variables for gener-
ating the feedback signals. This requirement may be met, if the sys-
tem state vector is observable from area measurements. However,
even if the observability condition is satisfied, the resulting con-
trollers with appropriately designed observers are normally quite
complicated, and therefore, these approaches are not suitable for
large-scale power systems where the total number of state vari-
ables is large.
Because of practical limitations in the implementation of LFC
based on feedback of all state variables, suboptimal LFC regulator
designs were considered [68,28,69]. The design method employs
model and singular perturbation techniques to affect decoupling
of the interconnection into its subsystem components. Lyapunov’s
second approach and utilization of the minimum setting time the-
ory have also been proposed by Shirai [70] for LFC regulator design.
4.2. Adaptive and variable structure methods
Adaptive control has been a topic of research for more than a
quarter of a century. Basically, adaptive control systems can be
classified into two categories: namely self-tuning regulators and
model reference control systems. The task of the adaptive control
technique is to make the process under control less sensitive to
changes in plant parameters and to un-modeled plant dynamics.
Various adaptive control techniques were proposed for LFC
schemes for dealing with plant parameter changes [71,57,72–
74,31,75]. Ross [71] described the control criteria in the LFC prob-
lem and the related practical difficulties encountered in trying to
achieve these criteria. The implementation and analysis of an
adaptive LFC strategy on the Hungarian power system have been
done by Vajk et al. [57]. Pan and Liwa [72] proposed an adaptive
controller using a proportional integral adaptation to meet the
hyperstability condition requirements considering plant parameter
changes. A multi-area adaptive control strategy [73] for a LFC
scheme and a reduced order adaptive load frequency controller
[74] for interconnected hydrothermal power system [75] have
been presented in the literatures. A self-tuning algorithm for solu-
tion of the LFC problem of interconnected power systems was re-
ported by Lee et al. [31] to provide the best control performance
for a wide range of operating conditions. Later on, a multi-variable
 H. Shayeghi et al. / Energy Conversion and Management 50 (2009) 344–353
self-tuning controller [75] was proposed by defining a cost func-
tion with a term representing the constraints on the control effort
and then minimizing that with respect to the control vector. Talaq
and Al-Basri [76] showed an adaptive fuzzy gain scheduling
scheme for conventional PI and optimal load frequency controllers.
Despite the promising results achieved by adaptive controllers, the
control algorithms are complicated and require on line system
model identification. Adaptive control strategies usually require a
perfect model following condition or explicit parameters identifi-
cation. These efforts seem unrealistic, since it is difficult to achieve
them.
Several authors [77,49,78,79] applied the concept of VSS to de-
sign load frequency controllers. Also, in Ref. [80], a fuzzy logic-
based variable structure controller was proposed for solution of
the LFC problem in electric power generation systems. This method
combines the salient features of both variable structure and fuzzy
systems to achieve high performance and robustness. It may be
noted that the VSS controllers can effectively improve transient re-
sponses due to load disturbances in the power system by properly
selecting the parameter of the controller. However, the complexity
of VSS and the associated chattering problem may be the reason
that these controllers were not fully appreciated in LFC tasks. Fur-
thermore, controllers based on the state equation of the linearized
model may require estimates of inaccessible state variables.
Observers can be designed for this, but it would involve the addi-
tional cost of data telemetry.
4.3. Robust approach
In power systems, each control area contains different kinds of
uncertainties and disturbances due to changes in system parame-
ters and characteristics, load variation, errors in modeling and lin-
earizing and environmental conditions. On the other hand, the
operating points of a power system may change very much ran-
domly during a daily cycle. Because of this, an optimal LFC regula-
tor design based on nominal system parameter values is certainly
not suitable for the LFC problem, and therefore, implementation of
these regulators on the system may be inadequate to provide the
desired system functioning. This could result in a degraded system
dynamic performance and sometimes also in the loss of system
stability. Thus, considerable efforts have been made to design
LFC controllers with better performance to cope with system
parameter changes using various robust methodologies [82–
84,36,85,86,37,87–90]. It may be noted that in the robust control
approaches, the control objectives are to design load frequency
controllers to not only meet nominal stability and nominal perfor-
mance requirements but also guarantee robust stability and robust
performance [87] in power systems on the LFC problem.
A robust controller based on the Riccati equation approach has
been proposed for solution of the LFC problem by Wang et al. [82].
Later, based on a combination of the robust control technique and
an adaptive control technique, a design of a new robust adaptive
controller was proposed for a power system LFC [83]. The motiva-
tion of combining the robust control with an adaptive control was
to use the robust control method to deal with the small parametric
uncertainties and the adaptive control technique for the large para-
metric uncertainties. In Ref. [84], a systematic control design ap-
proach based on quantitative feedback theory was suggested to
improve LFC performance, which offers an orderly environment
for design and analysis considering a physical understanding of
the system. Later, a robust load frequency controller based on
the H1 control method was proposed in Ref. [36] incorporating
parametric uncertainties. Analysis based on the l-synthesis tech-
nique shows that the proposed controller achieves good robust
performance for admissible plant parametric uncertainties. Azzam
[85] proposed a robust state feedback controller based on optimum
347
finding of the closed loop system eigenvalues that had minimum
sensitivity to system parameter changes. The design of a robust
stabilizing controller based on the Riccati equation method for all
admissible uncertainties using the combination of ‘Matching con-
ditions’ and Lyapunov stability theory has been proposed by Gos-
haidas et al. [86]. Shayeghi et al. [37] have applied l-synthesis
and the analysis technique for design of robust LFC considering
admissible parametric uncertainties. The effectiveness of the
resulting controller was compared with the proportional integral
controller optimized by the ISE index for a wide range of plant
parameter changes in the presence of GRCs. The design of a robust
controller based on Q-parameterization theory for control of the
frequency deviations was proposed by Azzam and Mohamed
[87]. This approach has several advantages such as a stable control-
ler always exists, the closed loop poles can be located in a pre-
scribed region for achieving the required transient response and
the resulting controller order is lower than the order of comparable
robust controllers, such as H1, LQG and l-analysis. The above ro-
bust control methods show good dynamical responses, however,
most of them are based on state feedback and require the availabil-
ity of all state variables for feedback. Thus, these efforts seem
unrealistic, since it is difficult to achieve them. Also, some of them
[80–84,36,85] have a centralized scheme, and the order of the
resulting controllers is high, which is not feasible for large-scale
power systems.
It may be noted that in the dynamical operation of power sys-
tems, it is usually important to aim for decentralization of control
to individual areas. This aim should coincide with the require-
ments for stability of the overall system. The main feature of the
robust decentralized scheme methods is reduction in the controller
complexity and suitability for practical implementation. Thus,
decentralized robust control has been suggested for multi-area
power system LFC [88–90]. Lim et al. [88] proposed a robust decen-
tralized load frequency controller based on the Riccati equation ap-
proach by embedding the local system parametric bounds in it for
achieving controller robustness. Using the proposed method, the
asymptotic stability of the overall power system for all admissible
uncertainties was guaranteed. A decentralized LFC based on H1
optimal control combined with an observer has appeared in Ref.
[89]. H1 control was proven to show greater effectiveness for
damping load disturbances over the conventional optimal control
by design of control system aimed at restricting the H1 norm of
its transfer function. The effectiveness of the proposed method
was investigated on three-area and six-area networks involving
both thermal and hydraulic power plants under load fluctuations.
Later, in Ref. [90], a systematic method for robust sequential
decentralized load frequency controllers using l-synthesis in a
multi-area power system was proposed. At each design step, the
information about the controllers design in the previous step is ta-
ken into account in the synthesis procedure. The investigations
were conducted on system dynamic performance in the presence
of parametric uncertainties and step load disturbances.
It should be noted that in the robust control design approaches,
it is possible to use the physical understanding of power systems
and consider the system uncertainties for the synthesis procedure.
However, the large model order, uncertain connection between
subsystems, broad parameter variations and elaborate organiza-
tional structure of power systems preclude direct application of
the standard robust control methodologies.
4.4. Intelligent approaches
In practice, many nonlinear plants, such as power systems, are
approximated by reduced order models, possibly linear, that
clearly are related by the underlying plant characteristics. How-
ever, these models are only valid within certain specific operating
348 H. Shayeghi et al. / Energy Conversion and Management 50 (2009) 344–353
ranges, and a different model may be required in the wake of
changing operating conditions, or the control system should adopt
the new system model parameters. On the other hand, due to the
complexity and multi-variable conditions of the power system,
classical and nonflexible LFC schemes do not represent good en-
ough solutions. Thus, for the purpose of evaluating the perfor-
mance of such systems, a flexible method was developed. In
recent years, the advent of modern intelligent methods, such as
ANNs, fuzzy logic and GAs, has solved the above-mentioned prob-
lems to a great extent.
The human ability to control complex plants has encouraged
researchers to pattern controls on human neural network systems.
ANNs, with their massive parallelism and ability to learn any type
of nonlinearities, are now being used in the area of nonlinear con-
trol problems, especially when the system is operating over the
nonlinear range. The applications of ANNs for solution of the LFC
problem are reported in Refs. [32,33,91,35,92–94,39]. Franoise
et al. [32] proposed multi-layer nonlinear networks for system
control which were trained by a back propagation through time
algorithm. The proposed ANN controller in comparison with classi-
cal PI controllers on single-area and two-area power systems has
achieved good performance. A new LFC scheme to incorporate
the nonconforming load problem was presented by Douglas et al.
[33], in which an effort had been undertaken to develop algorithms
capable of discriminating between non-controllable short-term
excursions and controllable long-term excursions. Of the two tech-
niques described, one was developed using a neural network algo-
rithm for pattern recognition of controllable signals, and the other
technique was based on the detection of the controllable signal in
the presence of a noisy random load using a random signal proba-
bility model. Test results reveal that the neural network-based AGC
implementation had a significant improvement over the modern
AGC implementation. The LFC system performance was evaluated
with a nonlinear neural network controller using a generalized
neural structure to yield better system dynamic performance than
the individual neurons [91]. A four-area interconnected power sys-
tem model with a reheat nonlinearity effect of the steam turbine
and upper and lower constraints for generation rate nonlinearity
of the hydro turbine was considered for the investigation in Ref.
[92]. It has been shown in Ref. [93] that the AGC problem can be
viewed as a stochastic multi-stage decision-making problem or a
Markov Chain control problem, and algorithms have been pre-
sented for designing AGC based on a reinforcement learning
approach.
Recently, application of the ANN technique based on robust
control methodologies for solution of the LFC problem in intercon-
nected power system has appeared in the literatures [94,39]. Shay-
eghi and Shayanfar [94] have used the idea of the H1 robust
control technique for training of RBF neural networks for improve-
ment of the performance of the proposed controller under various
operating conditions. In Ref. [39], the idea of l-synthesis control
techniques has been used for training an ANN-based LFC controller
too. These approaches combined the advantage of neural networks
and robust control techniques to achieve the desired level of robust
performance under large parametric uncertainness and lead to a
flexible controller with relatively simple structure.
Nowadays, fuzzy logic is used in almost all sectors of industry
and science. One of them is power system control. On the other
hand, their robustness and reliability make fuzzy controllers useful
for solving a wide range of control problems in power systems. Un-
like traditional control theory, which is essentially based on math-
ematical models of the controlled plants, the fuzzy control
methodology tries to establish the controller directly from domain
experts or operators who are controlling the plant manually and
successfully. To consider various power system aspects many stud-
ies have been reported for design of a fuzzy logic-based LFC regu-
lator in the literature [95–100]. Researches on the LFC problem
show that the fuzzy PI controller is simpler and more applicable
to remove the steady error [96,97]. The fuzzy PI controller is
known to give poor performance in system transient response. In
view of this, Yesil et al. [98] have proposed fuzzy PID methods to
improve the performance of the LFC problem. It should be pointed
out that they require a three dimensional rule base. This problem
makes the design process more difficult. The problem of decompo-
sition of multi-variable systems for the purpose of distributed fuz-
zy control design was proposed in Ref. [99]. This method has
reduced the number of interactive fuzzy relations among subsys-
tems. The combined intelligent technique using ANN and fuzzy lo-
gic theory has also been presented to utilize the novel aspects of
both designs to a single hybrid LFC system [100].
These days, GAs are the most popular and widely used algorithm
of all the intelligent algorithms. GAs are search algorithms based on
the mechanism of natural selection and natural genetics which oper-
ate without knowledge of the task domain and utilize only the fitness
of evaluated individuals. They can be considered as a general pur-
pose optimization method and have been widely used to solve many
complex nonlinear engineering optimization problems in general
and especially for the solution of LFC problems [101–106]. In Ref.
[101], the optimum adjustment gains of the integral controller using
GAs through performance indices ISE and ITAE are investigated. A
reinforced GA has been proposed as a suitable method to automate
tuning of the membership function and rule sets of fuzzy gain sched-
uling load frequency controllers to improve the dynamic perfor-
mance of multi-area power systems in the presence of system
nonlinearities, such as GRCs and governor dead band [101]. Later,
contrary to the trial and error selection of the variable structure-
based LFC feedback gains, a GA-based method was used for finding
optimal feedback gains [102]. The test results show that not only
the dynamical performance has been improved but also the control
effort is dramatically reduced. Karnavas et al. [100] have presented a
comprehensive study on the LFC problem of an autonomous power
system using combined advent of ANN, fuzzy logic and GA tech-
niques to achieve the desired level of robust performance. A higher
order robust dynamic performance is achieved for LFC designs based
on GA and LMIs [104]. The PI control parameters have been obtained
by coordinating the GA with a linear matrix inequalities control tool-
box for optimization of the H1 norm-based performance index. In
Ref. [105], a new GA/GA–SA-based fuzzy LFC scheme of a multi-area
thermal generating system is developed. A function like FD has been
used as the fitness function for evaluating the fitness of GA/hybrid
GA–SA optimization. This function directly depends on transient
performance characteristics such as overshoots, undershoots, set-
tling times and time derivative of the frequency. The hybrid GA–SA
technique yields more optimal gain values than the GA method.
Later, for optimization of PID gains in designing a Sugeno fuzzy
logic-based LFC scheme, a particle swarm optimization technique
was reported [106]. PSO, as one of the modern heuristic algorithms,
is a population-based evolutionary algorithm, which is motivated by
simulation of social behavior instead of survival of the fittest. The
proposed PSO algorithm establishes the true optimality of transient
performance similar to those obtained by the GA–SA-based optimi-
zation technique, but it is very much faster than the GA–SA algo-
rithm. The Sugeno fuzzy logic technique is very simple to
implement and acts much faster than the classical fuzzy logic
system.
4.5. Digital approaches
Since digital control is more accurate and reliable, compact in
size, less sensitive to noise and drift and more flexible, researchers
have focused their attention on proposing digital LFC control
schemes [107–109,62,110–112]. Ross and Green [107] were
 H. Shayeghi et al. / Energy Conversion and Management 50 (2009) 344–353
probably the first to present a comprehensive direct digital LFC
regulator for power systems. The results, incorporating dynamic
control criteria for performance evaluation, of the digital control
system based on field tests were outlined. As the ACE representing
generation mismatch in an area can be derived in discrete mode by
sampling the tie line power flow and system frequency deviations
and then transferring it over the telemetry links and unlinks in a
continuous time system, the control vector in the discrete mode
is constrained to remain constant between sampling instants.
Based on this fact, Bohn and Miniesy [62] have analyzed the effect
of the sampling period on the system’s dynamic behavior using a
discrete model of a single area power system. An informative study
on digital LFC modeling, including the criterion for evaluation of
system dynamic performance with the help of indices that mea-
sure the effectiveness of control relative to control efforts, is dis-
cussed by Demello et al. [110]. Kothaeri et al. [111,112] have
investigated more realistic modeling of LFC schemes in the discrete
mode, i.e. considering that the system is operating in continuous
mode and the controller is operating in discrete mode [111]. In
Ref. [112], a discrete mode LFC of an interconnected power system
with reheat thermal plants considering a new ACE is described. The
new ACE is derived from tie line power deviation, frequency devi-
ation, time error and inadvertent interchanges. Optimum gains of
the integral and PI controllers using the concepts of stability mar-
gin and the ISE index based on the conventional and new ACEs
have been obtained, and their dynamic performance was compared
for a step load disturbance.
5. LFC scheme with DC links
High-voltage direct current transmission has emerged on a
power scenario, due to its numerous technical and economic
advantages, for a large chunk of power transfer over large distances.
Besides other applications, the commissioning of an HVDC link in
parallel with existing ac links has shown beneficial effects from
the stabilization point of view of the system. Considerable research
works have appeared to investigate the damping effects of the dc
link on the frequency control of interconnected power systems
[113–117,40,118]. The effects of an automatic frequency ratio con-
trol on HVDC transmission to the automatic frequency control on ac
systems have been investigated in Ref. [113] when AFRC is applied
to a random load disturbance in a steady state. The frequency
improvement and reduction effects of the output power of regulat-
ing power stations by AFRC are analyzed. Sanpei et al. [114] have
developed a new dc control system based on multi-variable control
for frequency control of the two ac systems optimally while main-
taining their stability. A decentralized robust LFC based on the
Ricatti equation for a multi-area interconnected power system with
ac as well as frequency controllable HVDC links has been proposed
in Ref. [115]. Kumar and Ibraheem [116,117,40] have conducted a
comprehensive study for optimal load frequency controller design
of two-area power systems with parallel ac/dc links. Optimal con-
troller designs by considering the incremental dc link power flow
as an additional state, as well as control variable, were analyzed.
The results reveal that the system transient performance has signif-
icantly improved, considering the incremental dc link power flow
as an additional state variable in comparison to that obtained when
the system interconnection is through the ac link only.
6. LFC considering BES, SMES, SVC, SSSC, wind turbine and PV
systems
It should be noted that most of the proposed control strategies,
so far, for solution of the LFC problem have not been implemented
due to system operational constraints associated with thermal
349
power plants. The main reason is the non-availability of the re-
quired power, other than the stored energy in the generator rotors,
that can improve the performance of the system in the wake of
sudden increases in load demands. Also, due to the persistence of
the system frequency and tie line deviations for a long duration
in the case of small-load disturbances in the presence of suitable
supplementary controllers, the governor system may no longer
be able to absorb the frequency fluctuations due to its slow re-
sponse. On the other hand, electromechanical oscillations in a
power system can be effectively damped by fast acting energy stor-
age devices, because additional energy storage capacity is provided
as a supplement to the kinetic energy storage in the moving mass
of the generator rotor. The energy storage devices share the sudden
changes in power requirement in the load. Thus, in a power sys-
tem, the instantaneous mismatch between supply and demand of
real power for sudden load changes can be reduced by the addition
of active power sources with fast response such as BES, SMES and
CES devices. Some researchers [118–120] have investigated LFC
performance improvement of power systems including a BES unit.
A fast acting BES can effectively dampen electromechanical oscilla-
tions in a power system because it provides storage capacity in
addition to the kinetic energy of the generator rotor, which can
share the sudden changes in the power requirement. In Ref.
[119], it has been revealed that BES is helpful in meeting sudden
requirements of real power load, and the system transient perfor-
mance is significantly improved. Also, it is effective in reducing the
peak deviations of frequency and tie line power flows. The effects
of BES on LFC have been studied considering governor dead band
nonlinearity and GRCs by Lu and Liu [120]. The salient features
of SMES units such as low discharge rate, fast acting, increased
time required for power flow reversal and maintenance require-
ment have lead to its application as load frequency stabilizers
[120–130]. The performance of the adaptive controlled SMES is
compared with that of non-adaptive SMES, keeping the supple-
mentary control as the conventional one with the integral control-
ler discussed in Ref. [122]. It has been observed that when the
SMES control is adaptive, the performance is almost insensitive
to control gain parameter variation. In Ref. [123], the effects of
SMES on the LFC problem have been studied considering governor
dead band nonlinearity, steam reheat constraints and boiler
dynamics. Parameter optimization of the controller was performed
by the second method of Lyapunov, which ensures stability of the
system. The results reveal that the use of ACE for the control of
SMES units considerably reduces the tie line power deviation,
and the action of the SMES is localized with diminished contribu-
tion for load disturbances in the other area as compared to using
frequency deviations as the control signal. Later, the feasibility of
using an IGBT converter instead of the thyristor converter as a
power with SMES was achieved [124]. Demiroren [125,126] has
investigated the effects of a SMES unit on the LFC problem using
ANN and fuzzy logic-based supplementary controllers. The results
reveal that including SMES in each control area is effective in
reducing the peak deviations of frequency and tie line power flow
and in improving system transient response. Recently, in Ref.
[127], LFC of an interconnected hydrothermal power system
including SMES has been investigated. The optimum value of inte-
gral controller gain in the control areas in the presence of different
GRCs considering an SMES unit in either thermal or hydro area, as
well as both areas through minimizing the ISE performance index
based on the frequency and tie line power flow deviations was ob-
tained. The results reveal that considering SMES in both areas and
SMES in either of them improves the system dynamic performance
significantly. However, from the point of economy, including SMES
in either of the areas may be preferred.
The effects of CES units on the LFC problem considering gover-
nor dead band nonlinearity and boiler dynamics have been studied
350 H. Shayeghi et al. / Energy Conversion and Management 50 (2009) 344–353
by Tripathy [128]. It has been verified that CES provides the same
kind of dynamic performance improvement of the power system as
SMES, but compared to SMES, it is less expensive, has dielectric
losses and its energy density is low. The solid state phase shifter
systems or SSSC (Static Synchronous Series Compensators) have
also been found to aid in damping the oscillations of power system
dynamic response. New techniques of LFC regulator design, based
on SSSC, have been proposed in Refs. [129–131]. Because of the
high-speed performance of SSSC to the governor system, the SSSC
compensates the peak value of the transient frequency deviation
for a sudden load perturbation. The dynamics of the governor sys-
tem was eliminated in the control design process for the phase
shifter. In Ref. [129], a feedback signal composed of the area fre-
quency and tie line deviations has been used to stabilize the power
system. The coefficient of these deviations is obtained by using
pole placement of a reduced order system based on the aggrega-
tion of inertia center mode. In Ref. [130], the parameters of the
lead-lag controllers considered for the frequency stabilizer of SSSC
are automatically optimized by a micro-genetics algorithm using
evaluation of the fitness function based on a linear combination
of damping ratio of the inter-area mode and the multiplicative sta-
bility margin for taking system uncertainties into account. The re-
sults have confirmed the high robustness of this strategy against
various load disturbances with changing frequency in the vicinity
of the inter-area mode.
Currently, small wind turbines are among the candidate sys-
tems envisioned to operate in parallel with the utility’s generators.
The combined effect of the customer’s load demand and the wind
turbine fluctuating power output will develop a new load diversity
curve for the utility system. As a result, the regulation or LFC
requirements could differ significantly from the present ones.
Therefore, studies relating to LFC of power systems incorporating
the dynamics of such systems are reported in the literature
[131–133]. A method to analyze the effects that small wind tur-
bines may have on the utility’s LFC process has been developed
by Curtice and Reddoch [131]. Wind turbine output scenarios,
varying in frequency and magnitude, are combined with system
load variations to test the effectiveness of present AGC strategies.
The change in the system performance from the base case is as-
sessed using ACE values, time between zero crossings, inadvertent
accumulation and control pulses sent to regulating units.
An AGC scheme for a wind farm in the north of Spain with 37
variable speed wind turbines was developed in Ref. [132]. The con-
trol scheme is based on two control levels: a supervisory system
controls active and reactive power of the wind farm by sending
out set points to all wind turbines, and a machine control system
ensures that set points at the wind turbine level are reached. The
advancements in wind turbine and micro-hydro technology have
made it possible to introduce decentralized hybrid electric power
systems. A variable structure LFC of isolated wind-diesel micro-hy-
dro hybrid power systems was developed and dynamic perfor-
mance has been investigated by Bhatti and Kothari [133]. Static
VAR compensation systems have also been found to aid in damping
the oscillations of power system dynamic responses. A new tech-
nique of AGC regulator design, based on SVC, has been suggested
in Ref. [134]. A feedback signal composed of frequency deviation
and reactive power variation has been used to stabilize the electri-
cal power system. The coefficient of these deviations is a function
of the system and controller parameters. The influence of a PV sys-
tem on LFC has also been described in Ref. [135]. Besides the other
observations, it has been revealed that an electrical power system
containing a 10% contribution from PV stations would require a
2.5% increase in LFC capacity over a conventional system. More-
over, BES, SMES and CES units and a favorable effect of integrating
a 10% fuel cell into the power system dynamic model on power
system dynamic performance have also been evident [136].
7. LFC in a deregulated environment
Nowadays, the electric power industry is in transition to a com-
petitive energy market. In the new structure, GENCOs may not par-
ticipate in the LFC task and DISCOs have the liberty to control any
available GENCOs in their own or other areas. Global analysis of
the power system markets show that the LFC is one of the most
profitable ancillary services of these systems, and from the mecha-
nism used to manage the provision of this service in ancillary mar-
kets, the bilateral contracts or competitive offers stand out [22]. On
the other hand, the real world power system contains different
kinds of uncertainties and disturbances, and coming deregulation
significantly increases the severity of this problem. Under this con-
dition, the classical controller is certainly not suitable for the LFC
problem. In recent years, several control scenarios based on opti-
mal, robust and combined intelligent approaches have been pro-
posed for the LFC system in deregulated power systems. Some
researches are contained in Refs. [20,21] and [137–149]. In a dereg-
ulated environment, the independent system operator receives
contracts with GENCOs to provide area regulation. This is required
because of unscheduled generator and load changes and inconsis-
tent frequency bias existing in the system [137]. The governor re-
sponse is defined as area regulation contracts, and the cost of area
regulation is allocated among the players by the ratio of their par-
ticipation. Besides addressing the operational structures likely to
result from deregulation, the possible approaches to LFC and asso-
ciated technical issues; i.e. standards and algorithms, were de-
scribed by Christie and Bose [20]. Two alternative methods of LFC
of interconnected power systems of Norway and Sweden are intro-
duced by Bakken and Grande [138]. In some of the reported strate-
gies, attempts have been made to adapt well-tested classical LFC
schemes to the changing environment of power system operation
under deregulation [138,140]. A robust decentralized controller
based on H1 control theory and mixed H2/H1 control techniques
have been reported in Refs. [141–143] for solution of the LFC prob-
lem in a deregulated power system. To achieve decentralization, in
each control area, the effects of contracted signals and interfaces
between control areas are treated as a set of new disturbance sig-
nals. The proposed control strategy is formulated as a general LMIs
optimization problem and solved by the LMI control toolbox pro-
vided by MATLAB software. The results evaluation reveals that
these approaches achieve good dynamical performance, but the
stability of the overall system with local controllers was not dis-
cussed. Also, the order of the resulting controllers is high in general.
One development after deregulation of the power industry is the
necessity of a communication infrastructure to support an increas-
ing variety of ancillary services for effective implementation of LFC
schemes. An article focusing on the communication network
requirements for third party LFC services in an interconnected
power system was written by Bhowmik et al. [144]. Data commu-
nication models based on queuing theory have been proposed in
the study. The design of decentralized robust LFC applying the H1
norm and structured singular value was proposed by Shayeghi
and Shayanfar [50,51,145]. It has been shown that when frequency
response-based diagonal dominance cannot be achieved, subject to
a condition based on the H1 norm and structured singular value,
each local control area can be designed independently such that
the stability of the overall closed loop system is guaranteed. The
developed design strategy combines the advantages of H2 and H1
control synthesis and gives a powerful multi-objectives design ad-
dressed by LMI techniques by considering the effect of plant param-
eter changes, system nonlinearities and disturbances [51].
It may be noted that the main capability of robust control meth-
odologies-based LFC schemes [141–143,50,51,145] is in the possi-
bility of controller design based on a more complete model of the
system considering uncertainties, too. This fact is of great
 H. Shayeghi et al. / Energy Conversion and Management 50 (2009) 344–353
importance knowing that the power system has a variable struc-
ture and is subject to different types of uncertainties and distur-
bances. However, due to the complexity of the actual
uncertainties and the large model order of the real world power
system, these approaches yield complex controllers whose size will
be very large in general. Recently, in order to overcome these
drawbacks and retain the advantages of the robust control tech-
nique, a robust decentralized neural network-based LFC in a dereg-
ulated power system has appeared in the literature [22,146] by
Shayeghi and Shayanfar. The motivation for using a robust control
strategy for training a neural network-based controller was to take
large modeling uncertainties into account, cover physical con-
straints on control action and minimize the effects of area load dis-
turbances. The developed approaches combined the advantage of
the neural networks and robust control technique and led to a flex-
ible controller to provide robust performance [22]. Recently, a mul-
ti-stage fuzzy PID type controller with a fuzzy switch has been
proposed to solve the AGCs problem in a deregulated power sys-
tem using the generalized LFC dynamic model [147]. This approach
is a form of behavior-based control where the PD controller be-
comes active only when certain conditions are met. The resulting
structure is a controller using two-dimensional inference engines
(rule base) to perform reasonably the task of a three-dimensional
controller. The proposed method requires fewer sources for the
operator, and its role in the system response is more apparent
[148]. In a fuzzy control system, optimization of membership func-
tions plays a key role because it is a computationally expensive
combinatorial optimization problem. To reduce fuzzy system effort
and save cost, a GA-based MSF PID controller has been reported by
Shayeghi et al. [149] for solution of the LFC problem. The GA works
off line, and in order to improve the algorithm convergence speed
and guarantee an optimum value, a modified GA based on the hill
climbing method has been used. The salient feature of the pro-
posed method is that it does not require an accurate model of
the LFC problem, and the design process is less demanding com-
pared to other fuzzy PID controllers. The evaluation of system per-
formance, characterized in terms of ITAE and FD indices, revealed
that it can be a promising control scheme for solution of the LFC
problem and ideal for real world power systems.
The uncovered subject material on LFC is available in reviews
and state-of-the-art lectures [150–152].
8. Conclusion
Load frequency control is one of the important issues in power
systems operation and control for supplying sufficient and reliable
electric power with good quality. Especially in the deregulated
electricity market, it will serve as an ancillary service and acquires
a principal role to enable power exchanges and to provide better
conditions for electricity trading. LFC goals, i.e. frequency regula-
tion and tracking load demands, maintaining tie line power inter-
changes to specified values in the presence of modeling
uncertainties, system nonlinearities, complexity and multi-vari-
able condition of power system, determine LFC synthesis as a mul-
ti-objective optimization control problem. This paper is focused on
the recent research in the area of LFC and intends to be a useful ref-
erence and search tool as well as a critical account of the up-to-
date use of AI technologies in the LFC problem. Emphasis has been
given to categorizing various LFC strategies reported in the litera-
ture and their salient features and disadvantages.
Among the discussed categories of LFC strategies, robust control
and AI-based methods have shown an ability to give better perfor-
mance in dealing with the system nonlinearities, modeling uncer-
tainties and area load disturbances under different operating
conditions. The main capability of robust control approaches is
351
alleviation of the impossibility of controller design based on a more
complete model of the system that considers uncertainties and
physical constraints, too. The salient feature of the AI technique
is that it provides a model-free description of the control system
and does not require an accurate model of the plant. In conclusion,
we can say that the robust and AI techniques, like all other control
techniques, have relative advantages and disadvantages. There are
no rules as to when a particular technique is more suitable for the
LFC problem. It is envisaged that this paper will serve as a valuable
resource to any further worker in this important area of research.
References
[1] Kothari DP, Nagrath IJ. Modern power system analysis. 3rd
ed. Singapore: McGraw-Hill; 2003.
[2] Raineri R, Rios S, Schiele D. Technical and economic aspects of ancillary
services markets in the electric power industry: an international comparison.
Energy Policy 2006;34(13):1540–55.
[3] Kumar J, Hoe NG, Sheble G. AGC simulator for price-based operation. Part I:
Modeling. IEEE Trans Power Syst 1997;12(2):527–32.
[4] Saadat H. Power system analysis. USA: McGraw-Hill; 1999.
[5] Cohn N. Considerations in the regulation of interconnected area. IEEE Trans
Power Syst 1967;86:1527–38.
[6] Van Ness JE. Root loci of load frequency control systems. IEEE Trans Power
Appl Syst 1963;82(5):712–26.
[7] Wood A, Wollenberg B. Power generation, control and operation. 2nd ed. John
Wiley & Sons; 1996.
[8] Elgerd OI, Fosha C. Optimum megawatt frequency control of multi-area
electric energy systems. IEEE Trans Power Appl Syst 1970;89(4):556–63.
[9] Kwatny HG, Kalnitsky KC, Bhatt A. An optimal tracking approach to load
frequency control. IEEE Trans Power Appl Syst 1975;94(5):1635–43.
[10] Elgerd OI. Electric energy system theory: an introduction. 2nd ed. New
York: McGraw-Hill; 1982.
[11] Aggarwal RP, Bergseth FR. Large signal dynamics of load-frequency control
systems and their optimization using nonlinear programming: I & II. IEEE
Trans Power Appl Syst 1968;87(2):527–38.
[12] Cohn N. Techniques for improving the control of bulk power transfers on
interconnected systems. IEEE Trans Power Appl Syst 1971;90(6):2409–19.
[13] IEEE Committee Report. Standard definitions of terms for automatic
generation control on electric power systems. IEEE Trans Power Appl Syst
1970;89.
[14] IEEE PES Committee Report. Dynamic models for steam and hydro-turbines in
power system studies. IEEE Trans Power Appl Syst 1973;92.
[15] IEEE PES Working Group. Hydraulic turbine and turbine control models for
system dynamic. IEEE Trans Power Syst 1992;7(1):167–74.
[16] Kundur P. Power system stability and control. New York: McGraw-Hill; 1994.
[17] IEEE PES Committee Report. Current operating problems associated with
automatic generation control. IEEE Trans Power Appl Syst 1979;98:98.
[18] Jaleeli N, Vanslyck LS, Ewart DN, Fink LH, Hoffmann AG. Understanding
automatic generation control. IEEE Trans Power Appl Syst 1992;7(3):1106–22.
[19] Green RK. Transformed automatic generation control. IEEE Trans Power Syst
1996;11(4):1799–804.
[20] Christie RD, Bose A. Load frequency control issues in power system operations
after deregulation. IEEE Trans Power Syst 1996;11(3):1191–200.
[21] Donde V, Pai A, Hiskens IA. Simulation and optimization in a LFC system after
deregulation. IEEE Trans Power Syst 2001;16(3):481–9.
[22] Shayeghi H, Shayanfar HA, Malik OP. Robust decentralized neural networks
based LFC in a deregulated power system. Electr Power Syst Res
2007;47:241–51.
[23] Concordia C, Kirchmayer LK, Szymanski EA. Effect of speed governor dead-
band on tie-line power and frequency control performance. Am Inst Elect Eng
Trans 1957;76:429–35.
[24] Wu FF, Dea VS. Describing-function analysis of automatic generation control
system with governor dead band. Electr Power Syst Res 1978;1(2):113–6.
[25] Oni B, Graham H, Walker L. Investigation of nonlinear tie-line bias control of
interconnected power systems. IEEE Trans Power Appl Syst
1981;100(5):2350–6.
[26] Tripathy SC, Bhatti TS, Jha CS, Malik OP, Hope GS. Sampled data automatic
generation control analysis with reheat steam turbines and governor dead
band effects. IEEE Trans Power Appl Syst 1984;103(5):1045–51.
[27] Fosha CE, Elgerd OI. The megawatt frequency control problem: a new
approach via optimal control theory. IEEE Trans Power Appl Syst
1970;89(4):563–77.
[28] Aldeen M, Trinh H. Load frequency control of interconnected power systems
via constrained feedback control schemes. Int J Comput Elect Eng
1994;20(1):71–88.
[29] Ismail A. Robust load frequency control. In: Proceedings of the IEEE
conference on control applications, vol. 2, New York, Dayton, OH,
September 1992. p. 634–5.
[30] Yamashita K, Miyagi H. Load-frequency self-tuning regulator for
interconnected power systems with unknown deterministic load
disturbances. Int J Control 1989;49(5):1556–68.
352 H. Shayeghi et al. / Energy Conversion and Management 50 (2009) 344–353
[31] Lee KA, Yee H, Teo CY. Self-tuning algorithm for automatic generation control
in an interconnected power system. Electr Power Syst Res
1991;20(2):157–65.
[32] Franoise B, Magid Y, Bernard W. Application of neural networks to
load-frequency control in power systems. Neural Networks
1994;7(1):183–94.
[33] Douglas LD, Green TA, Kramer RA. New approaches to the AGC
nonconforming load problem. IEEE Trans Power Syst 1994;9(2):619–28.
[34] Chaturvedi DK, Satsangi PS, Kalra PK. Load frequency control: a generalized
neural network approach. Electr Power Energy Syst 1999;21(6):405–15.
[35] Demiroren A, Sengor NS, Zeynelgil HL. Automatic generation control by using
ANN technique. Electr Power Compon Syst 2001;29(10):883–96.
[36] Karrari M, Shayeghi H, Abedi M, Menhaj MB. Design of H1 controller load
frequency control in electrical power systems. Amirkabir J Sci Technol
1999;11(41):79–88.
[37] Shayeghi H, Karrari M, Menhaj MB. Theory of l synthesis for power systems
load frequency control. J Electron Eng 2000;51:258–63.
[38] Shayeghi H, Shayanfar HA. Application of ANN technique for interconnected
power system load frequency control. Int J Eng Trans B 2003;16(3):247–54.
[39] Shayeghi H, Shayanfar HA. Application of ANN technique based on l-
synthesis to load frequency control of interconnected power system. Electr
Power Energy Syst 2006;28:503–11.
[40] Ibraheem , Kumar P. A novel approach to the matrix Riccati equation solution:
an application to optimal control of interconnected power systems. J Electr
Power Compon Syst 2004;32(1):33–52.
[41] Kawabata H, Kido M. A decentralized scheme of load frequency control power
system. Electr Eng Jpn 1982;102(4):100–6.
[42] Park YM, Lee KY. Optimal decentralized load frequency control. Electr Power
Syst Res 1984;7(4):279–88.
[43] Calovic MS. Automatic generation control: decentralized area-wise optimal
solution. Electr Power Syst Res 1984;7(2):115–39.
[44] Aldeen M, Marsh JF. Observability controllability and decentralized control of
interconnected power systems. Int J Comput Elect Eng 1990;16(4):207–20.
[45] Aldeen M, Marsh JF. Decentralized proportional-plus-integral control design
method for interconnected power systems. IEE Proc Gen Transm Distri
1991;138(4):263–74.
[46] Aldeen M. Interaction modeling approach to distributed control with
application to power systems. Int J Control 1991;53(5):1035–54.
[47] Yang TC, Cimen H, Zhu QM. Decentralized load-frequency controller design
based on structured singular values. IEE Proc Gen Transm Distri
1998;145(1):7–14.
[48] Yang TC, Ding ZT, Yu H. Decentralized power system load frequency control
beyond the limit of diagonal dominance. Electr Power Energy Syst
2002;24(3):173–84.
[49] Kumar A, Malik OP, Hope GS. Variable-structure-system control applied to
AGC of an interconnected power system. IEE Proc Gen Transm Distri
1985;132(1):23–9.
[50] Shayeghi H, Shayanfar HA. Design of decentralized robust LFC in
a competitive electricity environment. J Electron Eng 2005;56(9–10):
225–36.
[51] Shayeghi H, Shayanfar HA. Robust decentralized LFC design in a restructured
power system. Int J Emerging Electr Power Syst 2006;6(2). Article 4.
[52] Lim KY, Wang Y, Guo G, Zhou R. A new decentralized robust controller design
for multi-area load frequency control via incomplete state feedback. Optim
Control Appl Meth 1998;19:345–61.
[53] Kazemi MH, Karrari M, Menhaj MB. Decentralized robust adaptive-output
feedback controller for power system load frequency control. Electron Eng
2002;84(2):75–83.
[54] Wang Y, Zhou R, Wen C. Robust load-frequency controller design for power
systems. IEE Proc Gen Transm Distri 1993;140(1):111–6.
[55] Wang Y, Zhou R, Wen C. New robust adaptive load frequency control with
system parameter uncertainties. IEE Proc Gen Transm Distri
1994;141(3):184–90.
[56] Ray G, Rani CS. Stabilizing decentralized robust controllers of interconnected
uncertain power systems based on the Hessenberg form: simulated results.
Int J Syst Sci 2001;32(3):387–99.
[57] Vajk I, Vajta M, Keviczky L. Adaptive load frequency control of Hungarian
power system. Automatica 1985;21(2):129–37.
[58] Elgerd OI, Fosha C. Optimum megawatt frequency control of multi-area
electric energy systems. IEEE Trans Power Appl Syst 1970;89(4):556–63.
1970.
[59] Bechert TE, Chen N. Area automatic generation control by multi-pass
dynamic programming. IEEE Trans Power Appl Syst 1977;96(5):1460–8.
[60] Das D, Nanda J, Kothari ML, Kothari DP. Automatic generation control of
hydrothermal system with new area control error considering generation
rate constraint. Electr Mach Power Syst 1990;18(6):461–71.
[61] Tacker EC, Lee CC, Reddoch TW, Tan TO, Julich PM. Optimal control of
interconnected electric energy systems: a new formulation. Proc IEEE
1972;60(10):1239–41.
[62] Bohn EV, Miniesy SM. Optimum load frequency sample data control with
randomly varying system disturbances. IEEE Trans Power Appl Syst
1972;91(5):1916–23.
[63] Yamashita K, Taniguchi T. Optimal observer design for load frequency control.
Int J Electr Power Energy Syst 1986;8(2):93–100.
[64] Feliachi A. Load frequency control using reduced order models and local
observers. Int J Energy Syst 1987;7(2):72–5.
[65] Rubaai A, Udo V. An adaptive control scheme for LFC of multi-area power
systems. Part I: Identification and functional design. Part-II: Implementation
and test results by simulation. Electr Power Syst Res 1992;24(3):183–97.
[66] Velusami S, Ramar K. Design of observer-based decentralized load-frequency
controllers for interconnected power systems. Int J Power Energy Syst
1997;17(2):152–60.
[67] Hain Y, Kulessky R, Nudelman G. Identification-based power unit model for
load-frequency control purposes. IEEE Trans Power Syst 2000;15(4):1313–21.
[68] Choi SS, Sim HK, Tan KS. Load frequency control via constant limited-state
feedback. Electr Power Syst Res 1981;4(4):265–9.
[69] Aldeen M, Trinh H. Combined modal and singular perturbation approach to
decentralized control. Int J Syst Sci 1992;23(5):741–64.
[70] Shirai G. Load frequency control using Lyapunov’s second method: bang–
bang control of speed changer position. Proc IEEE 1979;67(10):1458–9.
[71] Ross CW. Error adaptive control computer for interconnected power system.
IEEE Trans Power Appl Syst 1966;85:749.
[72] Pan CT, Liaw CM. An adaptive controller for power system and load frequency
control. IEEE Trans Power Syst 1989;4(1):122–8.
[73] Shoults RR, Ibarra JAJ. Multi-area adaptive LFC developed for a
comprehensive AGC simulator. IEEE Trans Power Appl Syst 1993;8(2):541–7.
[74] Liaw CM. Design of a reduced-order adaptive LFC for an interconnected
hydrothermal power system. Int J Control 1994;60(6):1051–63.
[75] Rubaai A, Udo V. Self-tuning LFC: multilevel adaptive approach. IEE Proc Gen
Transm Distri 1994;141(4):285–90.
[76] Talaq J, Al-Basri F. Adaptive fuzzy gain scheduling for load-frequency control.
IEEE Trans Power Syst 1999;14(1):145–50.
[77] Sivaramkrishna AY et al. Design of variable structure load frequency
controller using pole assignment technique. Int J Control 1984;40(3):487–98.
[78] Kumar A, Malik OP, Hope GS. Discrete variable-structure controller for load
frequency control of multi-area interconnected power system. IEE Proc Gen
Transm Distri 1987;134(2):116–22.
[79] Das D, Kothari ML, Kothari DP, Nanda J. Variable structure control strategy to
automatic generation control of interconnected reheat thermal systems. IEE
Proc Contr Theor Appl 1991;138(6):579–85.
[80] Ha Q, Trinh H. A variable-structure based controller with fuzzy tuning for
load frequency control. Int J Power Energy Syst 2001.
[81] Kazemi A, Andami H. A decentralized fuzzy logic enhanced variable structure
controller applied to load frequency control system. Iran J Sci Technol Trans B
2004;28(B3):295–303.
[82] Wang Y, Zhou R, Wen C. Robust load-frequency controller design for power
systems. Proc IEE Proc Gen Transm Distri 1993;140(1):111–6.
[83] Wang Y, Zhou R, Wen C. New robust adaptive load frequency control with
system parameter uncertainties. Proc IEE Proc Gen Transm Distri
1994;141(3):184–90.
[84] Stankovi AM, Tadmor G, Sakharuk TA. On robust control analysis and design
for load frequency regulation. IEEE Trans Power Syst 1998;13(2):449–54.
[85] Azzam M. Robust automatic generation control. Energy Convers Manage
1999;40(13):1413–21.
[86] Goshaidas R, Prasad AN, Prasad GD. A new approach to the design of load
frequency controller for large scale power system. Electr Power Syst Res
1999;51:13–22.
[87] Azzam M, Mohamed YS. Robust controller design for automatic generation
control based on Q-parameterization. Energy Convers Manage
2002;43(13):1663–73.
[88] Lim KY, Wang Y, Zhou R. Robust decentralized load frequency control of
multi-area power systems. IEE Proc Gen Transm Distrib 1996;143(5):377–86.
[89] Ishii T, Shirai G, Fujita G. Decentralized load frequency control based on H1
control. Electr Eng Jpn 2001;136(3):28–38.
[90] Bevrani H, Mitani Y, Tsuji K. Sequential design of decentralized load
frequency controllers using l synthesis and analysis. Energy Convers
Manage 2004;45(6):865–81.
[91] Chaturvedi DK, Satsangi PS, Kalra PK. Load frequency control: a generalized
neural network approach. Electr Power Energy Syst 1999;21(4):405–15.
[92] Zeynelgil HL, Demiroren A, Sengor NS. The application of ANN technique to
automatic generation control for multi-area power system. Electr Power
Energy Syst 2002;24(5):345–54.
[93] Ahamed TPI, Rao PSN, Sastry PS. A reinforcement learning approach to
automatic generation control. Electr Power Syst Res 2002;63:9–26.
[94] Shayeghi H, Shayanfar HA. Application of ANN technique for interconnected
power system load frequency control. Int J Eng 2003;16(3):247–54.
[95] Indulkar CS, Raj B. Application of fuzzy controller to automatic generation
control. Electr Mach Power Syst 1995;23(2):209–20.
[96] Chang CS, Fu W. Area load frequency control using fuzzy gain scheduling of PI
controllers. Electr Power Syst Res 1997;47:145–52.
[97] Cam E, Kocaarslan I. Load frequency control in two area power system using
fuzzy logic controller. Energy Convers Manage 2005;45:233–45.
[98] Yesil E, Guzelkaya M, Eksin I. Self tuning fuzzy PID type load and frequency
controller. Energy Convers Manage 2004;45:377–90.
[99] Gegov AE, Frank PM. Decomposition of multivariable systems for distributed
fuzzy control [power system load frequency control]. Fuzzy Sets Syst
1995;73(3):329–40.
[100] Karnavas YL, Papadopoulos DP. AGC for autonomous power system using
combined intelligent techniques. Electr Power Syst Res 2002;62(3):225–39.
[101] Chang CS, Fu W, Wen F. Load frequency control using genetic-algorithm
based fuzzy gain scheduling of PI controllers. Electr Mach Power Syst
1998;26(1):39–52.
 H. Shayeghi et al. / Energy Conversion and Management 50 (2009) 344–353
[102] Al-Hamouz ZM, Al-Duwaish HN. A new load frequency variable structure
controller using genetic algorithms. Electr Power Syst Res 2000;55(1):1–6.
[103] Aditya SK, Das D. Design of load frequency controllers using genetic
algorithm for two area interconnected hydro power system. Electr Power
Compon Syst 2003;31(1):81–94.
[104] Rerkpreedapong D, Hasanovic A, Feliachi A. Robust load frequency control
using genetic algorithms and linear matrix inequalities. IEEE Trans Power
Syst 2003;18(2):855–61.
[105] Ghoshal SP. Application of GA/GA–SA based fuzzy automatic generation
control of a multi-area thermal generating system. Electr Power Syst Res
2004;70(2):115–27.
[106] Ghoshal SP. Optimizations of PID gains by particle swarm optimizations in
fuzzy based automatic generation control. Electr Power Syst Res
2004;72(3):203–12.
[107] Ross CW, Green TA. Dynamic performance evaluation of a computer
controlled electric power system. IEEE Trans Power Appl Syst
1972;91:1156–65.
[108] Taylor CW, Cresap RL. Real-time power system simulations for automatic
generation control. IEEE Trans Power Appl Syst 1976;95:375–84.
[109] Kumar A. Discrete load frequency control of interconnected power system.
Int J Energy Syst 1989;9(2):73–7.
[110] Demello FP, Mills RJ, B’Rells WF. Automatic generation control. Part I –
Process modeling. IEEE Trans Power Appl Syst 1973;92:710–5.
[111] Kothari ML, Nanda J, Kothari DP, Das D. Discrete mode automatic generation
control of a two area reheat thermal system with new area control error. IEEE
Trans Power Appl Syst 1989;4(2):730–8.
[112] Hari L, Kothari ML, Nanda. Optimum selection of speed regulation parameters
for automatic generation control in discrete mode considering generation
rate constraints. IEE Proc Gen Transm Distrib 1991;138(5):401–6.
[113] Yoshida Y, Machida T. Study of the effect of the DC link on frequency control
in interconnected AC systems. IEEE Trans Power Appl Syst
1969;88(7):1036–42.
[114] Sanpei M, Kakehi A, Takeda H. Application of multi-variable control for
automatic frequency controller of HVDC transmission system. IEEE Trans
Power Deliver 1994;9(2):1063–8.
[115] Lim KY, Wang Y, Zhou R. Decentralized robust load-frequency control in
coordination with frequency-controllable HVDC links. Electr Power Energy
Syst 1997;19(7):423–31.
[116] Kumar P, Ibraheem. Dynamic performance evaluation of 2-area
interconnected power systems: a comparative study. J Inst Eng
1998;78:199–208.
[117] Ibraheem, Kumar P. Dynamic performance enhancement of hydropower
systems with asynchronous tie-lines. J Electr Power Compon Syst
2003;31(7):605–26.
[118] Kunish HJ, Kramer KG, Dominik H. Battery energy storage, another option for
load-frequency control and instantaneous reserve. IEEE Trans Energy Convers
1986;1(3):46–51.
[119] Aditya SK, Das D. Battery energy storage for load frequency control of an
interconnected power system. Electr Power Syst Res 2001;58(3):179–85.
[120] Lu CF, Liu CC. Effect of battery energy storage system on load frequency
control considering governor dead-band and generation rate constraint. IEEE
Trans Energy Convers 1995;10(3):555–61.
[121] Banerjee S, Chatterjee JK, Tripathy SC. Application of magnetic energy storage
unit as load frequency stabilizer. IEEE Trans Energy Convers
1990;5(1):46–51.
[122] Tripathy SC, Balasubramanian R, Nair PSC. Adaptive automatic generation
control with superconducting magnetic energy storage in power systems.
IEEE Trans Energy Convers 1992;7(3):434–41.
[123] Tripathy SC, Balasubramanian R, Nair PSC. Effect of superconducting
magnetic energy storage on automatic generation control considering
governor dead-band and boiler dynamics. IEEE Trans Power Syst
1992;7(3):1266–73.
[124] Tripathy SC, Juengst KP. Sampled data automatic generation control with
superconducting magnetic energy storage in power systems. IEEE Trans
Energy Convers 1997;12(2):187–92.
[125] Demiroren A. Automatic generation control using ANN technique for multi-
area power system with SMES units. Electr Power Compon Syst
2004;32(2):193–213.
353
[126] Demiroren A, Yesil E. Automatic generation control with fuzzy logic
controllers in the power system including SMES units. Electr Power Energy
Syst 2004;26(4):291–305.
[127] Abraham RJ, Das D, Patra A. Automatic generation control of an
interconnected hydrothermal power system considering superconducting
magnetic energy storage. Electr Power Energy Syst 2007;29:271–579.
[128] Tripathy SC. Improved load frequency control with capacitive energy storage.
Energy Convers Manage 1997;38(6):551–62.
[129] Nagmroo I, Mitani Y, Tsui K. Application of solid-state phase shifter to
stabilization of system frequency in an interconnected power system. Electr
Eng Jpn 2000;132(2):19–28.
[130] Ngamroo I, Tippayachai J, Dechanupaprittha S. Robust decentralized
frequency stabilizers design of static synchronous series compensators by
taking system uncertainties into consideration. Electr Power Energy Syst
2006;28(8):513–24.
[131] Curtice DH, Reddoch TW. An assessment of load frequency control impacts
caused by small wind turbines. IEEE Trans Power Syst 1983;102(1):162–70.
[132] Amenedo JLR, Arnalte S, Burgos JC. Automatic generation control of a wind
farm with variable speed wind turbines. IEEE Trans Energy Convers
2002;17(2):279–84.
[133] Bhatti TS, Kothari DP. Variable structure load-frequency control of isolated
wind-diesel-microhydro hybrid power systems. J Inst Eng 2002;83:52–6.
[134] El-Emary AA, El-Shibina MA. Application of static VAR compensation for load
frequency control. Electr Mach Power Syst 1997;25(9):1009–22.
[135] Asano H, Yajima K, Kaya Y. Influence of photovoltaic power generation on
required capacity for load frequency control. IEEE Trans Energy Convers
1996;11(1):188–93.
[136] Paradkar A, Davari A, Feliachi A, Biswas T. Integration of a fuel cell into the
power system using an optimal controller based on disturbance
accommodation control theory. J Power Sources 2004;128(2):218–30.
[137] Vanslyck L, Jaleeli N, Kelley WR. Implications of frequency bias settings on
interconnected system operation and inadvertent energy accounting. IEEE
Trans Power Syst 1989;4(2):712–23.
[138] Bakken BH, Grande OS. Automatic generation control in a deregulated power
system. IEEE Trans Power Syst 1998;13(4):1401–6.
[139] Kumar J, Ng KH, Sheble G. AGC simulator for price-based operation – Part II:
Case study results. IEEE Trans Power Syst 1997;12(2):533–8.
[140] Delfino B, Fornari F, Massucco S. Load-frequency control and inadvertent
interchange evaluation in restructured power systems. IEE Proc Gen Transm
Distrib 2002;149(5):607–14.
[141] Shayeghi H, Shayanfar HA. Decentralized robust load frequency control using
linear matrix inequalities in a deregulated multi-area power system. J Power
Eng Problems 2005;2:30–8.
[142] Bevrani H, Mitani Y, Tsuji K. Robust decentralized LFC in a restructured power
system. Energy Convers Manage 2004;45:2297–312.
[143] Bevrani H, Mitani Y, Tsuji K, Bevrani H. Bilateral based robust load frequency
control. Energy Convers Manage 2005;46(7–8):1129–46.
[144] Bhowmik S, Tomsovic K, Bose A. Communication models for third party load
frequency control. IEEE Trans Power Syst 2004;19(1):543–8.
[145] Shayeghi H, Shayanfar HA. Decentralized robust AGC based on structured
singular values. J Electron Eng 2006;57:258–63.
[146] Shayanfar HA, Shayeghi H. Decentralized load frequency control of a
deregulated electric power system using ANN technique. WSEAS Trans
Circuits Syst 2005;4(1):38–47.
[147] Shayeghi H, Shayanfar HA, Jalili A. Multi stage fuzzy PID power system
automatic generation controller in the deregulated environments. Energy
Convers Manage 2006;47:2829–45.
[148] Shayeghi H, Shayanfar HA, Jalili A. Multi stage fuzzy PID load frequency
controller in a restructured power system. J Electron Eng 2007;58(2):61–70.
[149] Shayeghi H, Jalili A, Shayanfar HA. Robust modified GA-based multi-stage
fuzzy LFC. Energy Convers Manage 2007;48(5):1656–70.
[150] Glavitsch H, Stoffel J. Automatic generation control: a survey. Electr Power
Energy Syst 1980;2(1):21–8.
[151] Kumar P, Ibraheem. AGC strategies: a comprehensive review. Int J Power
Energy Syst 1996;16(1):371–6.
[152] Ibraheem, Kumar P, Kothari DP. Recent philosophies of automatic generation
control strategies in power systems. IEEE Trans Power Syst
2005;20(1):346–57.