Paper

LMI-Based Mixed H2/H Controller Design with Regional Pole

Constraints for Damping Power System Oscillations
Hardiansyah Student Member
Seizo Furuya Member
Juichi Irisawa Member
This paper presents the mixed H2 /H controller design with regional pole constraints for damping power
system oscillations. The state feedback gain can be obtained by solving a linear matrix inequality (LMI)
feasibility problem that robustly assigns the closed-loop poles in a prescribed LMI region. The proposed
technique is illustrated with applications to the design of stabilizer for a typical single-machine innite-bus
(SMIB) and a multimachine power system. The LMI-based control ensures adequate damping for widely
varying system operating conditions. The simulation results illustrate the eectiveness and robustness of the
proposed stabilizer.
Keywords: mixed H2 /H control, regional pole constraints, linear matrix inequality, power system stabilizer, robust
control

uncertainty in plant parameter variations. Some of

those eorts have been contributed to design robust
controllers for PSS and/or FACTS devices using H
concept such as mixed-sensitivity (7)(10) ; -synthesis or
structured singular value (SSV) (11) and H2 norm concept such as LQG (12) . Normally, the problem is formulated as a weighted mixed-sensitivity design and solved
by a Riccati approach. In addition, robust H design
being essentially a frequency domain approach does not
provide much control over the transient behavior and
closed-loop pole location. Robust pole placement stabilizer design using linear matrix inequalities (LMIs) has
been presented in Ref. (13), where the feedback gain matrix is obtained as the solution of a linear matrix inequality expressing the pole region constraints for polytopic
plants.
Design methods based on the H norm of the closedloop transfer function have gained popularity, because
unlike H2 methods (best known as LQG), they oer a
single framework in which to deal both with performance
and robustness. On the other hand, since an H2 cost
function oers a more natural way of representing certain aspects of the system performance, improving the
robustness of H2 -based design methods against perturbations of the nominal plant is a problem of considerable
importance for practical applications (14) . In the robust
H2 approach, the controller is designed to minimize an
upper bound on the worst- case H2 norm for a range of
admissible plant perturbations. Thus, a combination of
H2 control and H control, called mixed H2 /H control that minimized the H2 norm of some closed-loop
function subject to the H norm constraint of another
closed-loop function. Khargonekar et al. (18) considered
state- and output-feedback problems of mixed H2 /H
control and gave ecient convex optimization approach

1. Introduction
Power systems are usually large nonlinear systems,
which are often subject to low frequency oscillations
when working under some adverse loading conditions.
The oscillation may sustain and grow to cause system
separation if no adequate damping is available. To enhance system damping, the generators are equipped with
power system stabilizers (PSSs) that provide supplementary feedback stabilizing signals in the excitation systems. PSSs enhance the power system stability limit by
enhancing the system damping of low frequency oscillations associated with the electromechanical modes (1) .
Many approaches are available for PSS design, most of
which are based either on classical control methods (1)(3)
or on intelligent control strategies (4)(6) .
However, as power systems are large nonlinear systems, it is impossible for the system to always run at
the preselected operating conditions. When the system
is away from the specied operating point, the performance of the PSS will degenerate. Power systems continually undergo changes in the operating condition due
to changes in the loads, generation and the transmission network resulting in accompanying changes in the
system dynamics. A well-designed PSS has to perform
satisfactorily in the presence of such variations in the
power system. In other words, the stabilizer should be
robust to changes in the power system over its entire
operating range.
In the last few years, robust control technique has
been applied to power system controller design to guarantee robust performance and robust stability, due to

Department of Electrical Engineering, Nagaoka University of

Technology
1603-1, Kamitomiokamachi, Nagaoka 940-2188

920

IEEJ Trans. PE, Vol.124, No.7, 2004

LMI-based Mixed H2 /H Controller Design

where Q(x) = Q(x)T , R(x) = R(x)T , and S(x) depend

anely on x, is equivalent to

to solve the coupled nonlinear matrix Riccati equations.

With the development of numerical algorithms for
solving linear matrix inequality (LMI) problems in the
last 8 years, the LMI approach have emerged as a useful
tool for solving a wide variety of control problems (16) .
One of the advantages of linear matrix inequality (LMI)
is mixing the time and frequency domain objectives.
This paper proposes a mixed H2 /H controller design
with regional pole constraints for damping power system oscillations base on linear matrix inequality. The
eciency of an LMI-based design approach as a practical design tool is illustrated with case studies, including
a typical single-machine innite-bus (SMIB) and a multimachine power system.
The paper is organized as follows. A detailed description of the proposed design procedure is given in Section 2. In Section 3, simulation results are given for a
typical single-machine innite-bus (SMIB) and a multimachine power system to demonstrate the eectiveness
of the proposed method. Conclusions are drawn in Section 4.

R(x) > 0, Q(x) S(x)R(x)1 S(x)T > 0 (3)

In other words, the set of nonlinear inequalities Eq. (3)
can be represented as the LMI Eq. (2).
Two standard LMI optimization problems are of interest:
( 1 ) LMI feasibility problem. Given an LMI F (x) >
0, the corresponding LMI feasibility problem is
to nd xfeas such that F (xfeas ) > 0 or determine
that the LMI is infeasible.
( 2 ) Semi-Denite Programming problem (SDP).
An SDP requires the minimization of a linear objective subject to LMI constraints:

Minimize cT x
(4)
Subject to F (x) > 0
where c is a real vector, and F is a symmetric matrix
that depends anely on the optimization variable x.
This is a convex optimization problem.
Both these problems can be numerically solved vary
eciently, using currently available software (16) (17) .
2.2 LMI Formulation for Mixed H2 /H Performance
Consider the linear plant P with input
u, disturbance w, performance output z and z2 , the
measurement signal x. The input is generated by state
feedback, using the controller K. The signal z is the
performance associated with the H constraint, the signal z2 is the performance associated with the H2 criterion. The state space representation of the controlled
system can be written as follows:

x = Ax + B1 w + B2 u

z = C1 x + D12 u
(5)

z2 = C2 x + D22 u

2. LMI-based Mixed H2 /H Controller

Design

Stability is a minimum requirement for control system. However, in most practical situations, a good
controller should also deliver suciently fast and welldamped time responses. A customary way to guarantee satisfactory transients (or dynamics) is to place the
closed-loop poles in a suitable region of the complex splane.
This section discusses state feedback synthesis with
a combination of mixed H2 /H performance and pole
assignment specications. Here, the closed-loop poles
are required to lie in some LMI region D contained in
the left-half plane. Unconstrained mixed H2 /H synthesis is considered in Ref. (18), where an LMI-based
synthesis procedure is proposed. Excellent background
material on LMI may be found in Ref. (15).
2.1 Introduction of Linear Matrix Inequality
A wide variety of problems in control theory and system can be reduced to a handful of standard convex
and quasi-convex optimization problems that involve linear matrix inequalities (LMIs), that is constraints of the
form (15) :
F (x)
F0 +

m

where all the matrices are constant real matrices of appropriate dimension. The illustration of the controlled
system is shown in Fig. 1.
After substitution of the state feedback controller
u = Kx into Eq. (5), the closed-loop system becomes

x = (A + B2 K)x + B1 w

z = (C1 + D12 K)x

(6)

z2 = (C2 + D22 K)x

xi Fi > 0 (1)

Let Tz w and Tz2 w be the closed-loop transfer matrices

from the generalized disturbance w to the performance

i=1

where x = [x1 , x2 , , xm ] Rm is the variable, and the

symmetric matrices Fi = FiT Rnn , i = 0, , m, are
given. The set {x|F (x) > 0} is convex, and need not
have smooth boundary.
When the matrices Fi are diagonal, the LMI F (x) > 0
is just a set of linear inequalities. Nonlinear (convex) inequalities are converted to LMI form using Schur complements. The basic idea is as follows:


Q(x) S(x)
> 0 (2)
S(x)T R(x)
B124 7 2004

Fig. 1.
921

Mixed H2 /H control

output z and z2 , respectively:


 

A + B2 K B1
Acl Bcl
Tz w (s) =
=
C1 + D12 K 0
Ccl 0
(7)
 

A + B2 K B1
Acl Bcl
Tz2 w (s) =
=
C2 + D22 K 0
Ccl2 0


(8)
The goal of mixed H2 /H control is to nd an internally stabilizing controller K that minimizes the H2
performance, Tz2 w 2 , subject to the H performance,
Tz w  < and places the closed-loop poles in some
LMI stability region D that will be explained in the
next subsection. In this subsection, pure H2 and H
synthesis are not given. For proofs and more details, see
Refs. (19), (20).
We are now ready to give tractable necessary and sucient conditions for solving the following mixed H2 /H
problem:

Fig. 2.

Region S(, r, )

fD (z) = [kl + kl z + kl z]1k,lm (fD is valued in the

space of m m Hermitian matrices).
The location of the closed-loop poles of (A + B2 K)
in Eq. (6) concern with the performance of the closedloop system, i.e., the stability, the decay rate, the maximum overshoot, the rise time and settling time. Therefore, it is interesting work for control engineers to design the control gain K such that the closed-loop poles
of (A + B2 K) lie in a suitable subregion of the left half
plane. The interesting region for control purposes is the
set S(, r, ) of complex number x + jy such that

T
min trace (Ccl2 P Ccl2
) (9)

T
T
T 
Acl P + P Acl + Bcl Bcl
P Ccl
s.t.
<0
Ccl P
2 I

(10)
P = PT > 0
The optimization problem above is not yet convex because of the products KP arising in terms like Acl P . So,
dening the new variables Y = Y T = P , L = KY and
W = W T and using Schurs complement it is possible
to rewrite the problem above as the LMI problem

x < < 0, |x + jy| < r, and tan()x < |y|

(15)

min trace (W ) (11)


T 
H
Y C1T + LT D12
s.t.
< 0 (12)
C1 Y + D12 L
2 I

T 
Y
Y C2T + LT D22
> 0 (13)
C2 Y + D22 L
W

as shown in Fig. 2. Conning the closed-loop poles to

this region ensures a minimum decay rates , a minimum damping ratio = cos , and a maximum undamped natural frequency d = r sin ( in radian).
The LMI formulations for the poles of (A + B2 K) lie
in the region S(, r, ) are characterized as the following
LMIs (19) (20) : if there exists symmetric P > 0 such that

where H = AY + Y AT + B2 L + LT B2T + B1 B1T .

2.3 LMI Formulation for Regional Pole Constraints
In the synthesis of control systems, meeting
some desired transient performance objectives (to ensure fast and well-damped transient response, reasonable
feedback gain, etc.) should be considered. Generally,
H2 -norm and H synthesis design do not directly deal
with the transient response of the closed-loop system. In
contrast, a satisfactory transient response can be guaranteed by conning its poles in a prescribe region. For
many practical problems, exact pole assignment may not
be necessary; it suces to locate the closed-loop poles
in a prescribe subregion in the complex left half plane.

(A + B2 K)P + P (A + B2 K)T + 2P < 0

(16)


rP
(A + B2 K)P
< 0 (17)
P (A + B2 K)T
rP
and

sin()((A + B2 K)P + P (A + B2 K)T )

cos()(P (A + B2 K)T (A + B2 K)P )

cos()((A + B2 K)P P (A + B2 K)T )

<0
sin()((A + B2 K)P + P (A + B2 K)T )
(18)

Definition 1. LMI stability region (19) . A subset D of

the complex plane is called an LMI region if there exist
a symmetric matrix = [kl ] Rmm and a matrix
= [kl ] Rmm such that

with L = KP ; Y = P , the above LMIs are equivalent

to
AY + Y AT + B2 L + (B2 L)T + 2Y < 0 (19)


rY
AY + B2 L
< 0 (20)
Y AT + (B2 L)T
rY

D = {z C : fD (z) < 0} (14)

where the characteristic function fD (z) is given by
922

IEEJ Trans. PE, Vol.124, No.7, 2004

LMI-based Mixed H2 /H Controller Design

sin()(AY + B2 L + Y AT + (B2 L)T )

cos()(Y AT + (B2 L)T AY B2 L)

cos()(AY + B2 L Y AT (B2 L)T )

<0
sin()(AY + B2 L + Y AT + (B2 L)T )
(21)

From the analysis above, if there exists Y and L for

Eqs. (19)(21), then the poles of (A + B2 K) lie in the
region S(, r, ).
2.4 Mixed H2 /H Controller with Regional
Pole Constraints
The combination objectives of
robust mixed H2 /H control with regional pole constraints can be characterized as follows:

Fig. 3.

A SMIB power system

min trace (W )

{Y,L}

s.t. Eqs. (12), (13), and (19)(21) (22)

From the analysis above, the most important task in
this paper is to nd the variables Y , L, and W can be
solved using standard optimization techniques. Once a
feasible solution (Y , L) satisfying Eq. (22) is found, the
required state feedback gain matrix can be computed as
Fig. 4.

K = LY 1 (23)

Static fast exciter model

Table 1.

which leads to
Tz w  , Tz2 w 2

Operating conditions


trace (W ) (24)

The Lyapunov shaping paradigm for multi-objective

design provides a greater exibility than single-objective
optimal design techniques such as H synthesis or H2 norm technique.
3. Simulation Results
To test the performance of the proposed stabilizer,
simulation studies were performed on the two systems:
a single-machine innite-bus (SMIB) and a 3-machine
9-bus power system.
3.1 A Single-machine Infinite-bus System A
typical single-machine innite-bus (SMIB) power system (21) is chosen for analysis of the proposed controller
as shown in Fig. 3. The static fast exciter is shown in
Fig. 4. The overall system is of fourth order, with the
synchronous machine represented by third order model
and exciter-voltage regulator system represented by rst
order dynamics. The state variables of this model are
, , Eq , Ef d , respectively, angular speed, rotor angle, voltage behind transient, and excitation voltage. The power input to the generator shaft is assumed
constant, the network is represented by a set of algebraic equations and the loads are modeled by constant
impedance. The machine data, exciter data, block diagram of CPSS, and CPSS constants are given in the
Appendix.
The operating condition: Pe = 1.0 pu, Qe = 0.015 pu
and vt = 1.05 pu are chosen as the nominal operating condition and other operating points are regarded
as perturbations of the nominal system. Four dierent
loading conditions representing nominal, heavy, light,
B124 7 2004

and leading power factor (PF) are considered as given

in Table 1. The eigenvalues of the nominal system are
0.295 j4.96 and 10.4 j3.28. It is observed that
the electromechanical mode (characterized by the pair
of eigenvalues 0.295 j4.96) is negatively damped and
the eigenvalues for this mode should be shifted leftward
to more desirable locations into the left half s-plane.
It is proposed to apply a full state feedback controller
using the LMI based approach described in Section 2.
There is some limitation for application of state feedback design in comparison with output feedback design
because state feedback design needs all states to be measured. However, there are many instruments that could
be used to measure the state variables accurately. For
the 4th order model used here, the states are the deviations in the rotor speed , load angle , eld voltage
Ef d and the internal voltage Eq . Of these, , and
Ef d can be directly measured using appropriate transducers. The internal voltage Eq can be computed from
the instantaneous values of the stator currents and the
equivalent circuit parameters. On the other hand, there
are some essential problems in output feedback design.
For instance, a higher-order controller resulted from output feedback design would lead to diculties in practical
implementation. Therefore, state feedback design would
923

be a reasonable choice than output feedback design, if

all states can be measured.
The feasibility problem was solved for (Y , L) and
the required state feedback matrix was obtained as
K = LY 1 , where Y is a symmetric, positive denite
matrix and L is the matrix introduced to obtain linearity.
To determine the gain of the state feedback controller,
the minimization problem Eq. (22) is solved using LMI
Control Toolbox (16) . We can solve the eigenvalue problem in Eq. (22) with regional pole constraints in the region of S(2, 16, 1.1593). In this case, the associated matrices and scalars are

0.0151 0.2730
1.4668 100.90
0.2730 10.8880 16.902
618.11

Y =
1.4668 16.902
197.26 19365
100.90
618.11 19365 2524100

For completeness and verications, all controllers were

tasted at the following disturbances and loading conditions.
(a) Nominal loading (P, Q) = (1.0, 0.015) pu with
one-line fault.
(b) Heavy loading (P, Q) = (1.2, 0.3) pu with oneline fault.
A line fault is assumed; one of the transmission lines (as

4.7418 25.4840 2486.9]


= 0.8864, trace (W ) = 1.1388
L = [0.0151

and thus the feedback gain matrix is

K = [288.65 1.9197

3.7586 0.01678]

which makes the poles of (A + B2 K) locate at 5.77

j10.2 and 12.70 j5.12. As a nal check, we can calculate
Tz w  = 0.8748,

Fig. 5.

Response with 10% step in Vref for case 1

Fig. 6.

Response with 10% step in Vref for case 2

Fig. 7.

Response with 10% step in Vref for case 3

Tz2 w 2 = 0.5616

that satises the inequality Eq. (24).

For evaluation purposes, the performance of the system with the proposed controller was compared to the
CPSS (21) and H PSS. A small disturbance of 10% step
increase in the reference voltage (Vref ) was applied to
the SMIB power system at four dierent operating conditions. The system eigenvalues and damping ratios of
electromechanical modes at various operating conditions
are given in Table 2. Note that the damping ratio as
shown in Table 2 is written in the brackets. It is clear
that the system stability is greatly enhanced with the
proposed stabilizer. It can also be seen that all eigenvalues and damping ratios with the proposed stabilizer
lie in an LMI region of S. Simulation results shown in
Figs. 58 illustrate the performance and robustness of
the proposed PSS under dierent operating conditions.
It can be seen that the proposed PSS yields the best
damping performance compared to the CPSS as well as
H PSS. It is less sensitive to changes in the system parameters and more robust against model uncertainties.
Table 2. System eigenvalues and damping ratios
at various operating conditions

924

IEEJ Trans. PE, Vol.124, No.7, 2004

LMI-based Mixed H2 /H Controller Design

Fig. 8.

Response with 10% step in Vref for case 4

Fig. 11.

Fig. 9.

Response with fault disturbance for case (a)

Fig. 10.

Response with fault disturbance for case (b)

stability of the power systems.

3.2 A 3-machine 9-bus Power System
In this
part of the study, the 3-machine 9-bus power system
shown in Fig. 11 is considered. Details of the system
data are given in Ref. (22).
Each machine has been represented by a 3rd-order
generators equipped with a static exciter. Without
power system stabilizers, the system damping is poor
and the system exhibits highly oscillatory response. It
is therefore necessary to install one or more PSSs to
improve the dynamic performance. To identify the
optimum locations of PSSs, the participation factor
method (23) was used. The results of the method indicate
that all generators (G1, G2 and G3) are the optimum
locations for installing PSSs to damp out the electromechanical modes of oscillations.
To design the proposed controller, three operation
conditions, i.e. a heavy loading condition, a nominal
loading condition, and a light loading condition, are considered as shown in Table 3. The open loop eigenvalues
(dominant eigenvalues) of the study system for three
operating conditions are given in Table 4. As each pair
of conjugate eigenvalues corresponds to an oscillation
mode, there are three modes in this study system. Mode
1, 2 and 3 are the rotor oscillation modes (the electromechanical modes). It can be seen that the damping of the
rotor oscillation modes for all the operating conditions
are poor. In the power systems, a damping ratio () of at
least 10% and the real part of eigenvalue () not greater
than 0.5 for the troublesome low frequency electromechanical mode, guarantees that the low frequency oscillations, when excited, will die down in a reasonably
short time.
In order to improve the damping of electromechanical modes, a decentralized controller was designed for
all generators at the nominal loading conditions based
on the proposed design technique in Section 2. In this
scheme, each of the generators is tted with a partial
state feedback controller so that only locally available
states are feedback at each generator. This implies that

shown in Fig. 3) met a line-fault and the circuit breaker

operated. The simulation results for cases (a) and (b) as
shown in Fig. 9 and Fig. 10, respectively. Fig. 10 shows
that both CPSS and H PSS fail to stabilize the system with disturbance (b), the proposed PSS provide
good damping characteristics and system is stable under
this disturbance. It is clear that the proposed PSS exhibits better damping properties and guarantees robust
B124 7 2004

Three-machine nine-bus power system

925

Table 3.

Table 5.

Loading condition (in pu)

CPSS parameters

Table 6. Closed-loop eigenvalues and damping ratios at

various operating conditions

Table 4.

Open loop eigenvalues of the study system

were used method in Ref. (24). In this paper, a CPSS

with transfer function
G(s) = Kc

Fig. 12.

sTw (1 + sT1 )2
(25)
1 + sTw (1 + sT2 )2

was used and the parameters of stabilizer have been

tuned to provide an adequate amount of damping for
mode of oscillation. The CPSS data for a three-machine
system is given in Table 5. The output of all CPSS is
limited to 0.1 pu.
With the proposed PSS, the closed-loop eigenvalues
and damping ratios at various operating conditions are
given in Table 6. Note that the damping ratio as shown
in Table 6 is written in the brackets. It is quite clear
that the system stability is greatly enhanced with the
proposed stabilizer. It can also be seen that all eigenvalues and damping ratios associated with the electromechanical modes have been successfully shifted to lie in
an LMI region of S with the proposed PSS.
To demonstrate the capability of the proposed PSS to
enhance system damping over a wide range of operating
conditions, three dierent loading conditions were considered. A 10% step change in the reference voltage was
applied at generator 2 (G2) as follows.
( 1 ) Nominal loading condition
The system responses at the nominal loading condition are shown in
Figs. 13(a)(c), respectively, speed deviation of generator 1, speed deviation of generator 2, and speed deviation of generator 3. It is obvious that the system performance with the proposed PSS is better than CPSS.
( 2 ) Heavy loading condition
The system responses at the heavy loading condition are shown in
Figs. 14(a)(c), respectively, speed deviation of generator 1, speed deviation of generator 2, and speed deviation of generator 3. Regarding to the control efforts, the voltage response of generator following the
disturbance under heavy loading condition as shown in
Figs. 15(a)(c), respectively. The results here show the
superiority of the proposed PSS to the CPSS. It can
be concluded that the proposed PSS provides very good

Schematic of decentralized state feedback controller

the state feedback matrix K of the overall system is

block diagonal. This is schematically shown in Fig. 12,
where the submatrices K1 , K2 , and K3 are the feedback gain of each of the three generators. The locally
measured states: , , Eq , and Ef d are feedback at the AVR reference input of each machine after multiplication by suitable feedback gains. The LMI
problem was constructed by writing the LMI formulation Eq. (22). The feasibility problem was solved for
(Y , L) with regional pole constraints in the region of
S (2,121, 1.369) and the required state feedback matrix
was obtained as K = LY 1 . If the matrices L and Y
used in LMI formulation are restricted to be block diagonal then the product LY 1 will also have a block
diagonal structure. Such a requirement on the L and
Y matrices merely means restricting certain o diagonal elements to be zero which is an easily implemented
additional constraint in the optimization problem.
In order to facilitate comparison with CPSS, the design and tuning of CPSS for this multimachine system
926

IEEJ Trans. PE, Vol.124, No.7, 2004

LMI-based Mixed H2 /H Controller Design

(a)

(a)

(b)

(b)

(c)

(c)

Fig. 13. Generator response under nominal loading

condition

Fig. 14. Generator response under heavy loading

condition

damping over a wide range of operating conditions.

( 3 ) Light loading condition
The system responses at the light loading condition are shown in
Figs. 16(a)(c), respectively, speed deviation of generator 1, speed deviation of generator 2, and speed deviation of generator 3. It is clear that the proposed PSS
provide good damping characteristics to low-frequency
B124 7 2004

oscillations and greatly enhance the dynamic stability of

power system.
Studies have shown that a well-tuned PSS can improve power system dynamic stability eectively. Conventional power system stabilizer (CPSS) of the leadlag compensation type has been adopted by most utility
companies because of their simple structure, exibility
927

(a)

(a)

(b)
(b)

(c)

Fig. 15. Voltage response under heavy loading

condition

(c)

Fig. 16. Generator response under light loading

condition

and easy of implementation. Unfortunately, the major

disadvantage of the CPSS design method is that it does
not guarantee system stability under varying operating
conditions. The main objective of mixed H2 /H controller design is that the controller must ensure adequate damping at dierent operating conditions. The
adequacy of the damping is governed by the to settle the oscillations following a disturbance, including a

short circuit with clearance. It can, therefore, be said

that the LMI-based robust control will always guarantee
adequate robust damping performance whereas conventional PSS may fail to provide this over a range of operating conditions. The LMI method is very systematic
and methodical and should therefore be the preferred
option.
928

IEEJ Trans. PE, Vol.124, No.7, 2004

LMI-based Mixed H2 /H Controller Design

(17)

4. Conclusions
This paper has presented the design of mixed H2 /H
control with regional pole constraints for damping power
system oscillations. The required state feedback gain
has been obtained by solving a linear matrix inequality (LMI) feasibility problem that robustly assigns the
closed-loop poles in a prescribed LMI region. The performance of the proposed stabilizer on a SMIB and a
multimachine power system are seen to be robust over
a wide range of operating conditions. Finally, simulation results show the eectiveness and robustness of the
proposed stabilizer to enhance the damping of low frequency oscillations.
(Manuscript received Sep. 2, 2003,
revised Jan. 20, 2004)

(18)

(19)

(20)

(21)
(22)
(23)

(24)

References

L.E. Ghaoui, F.Delebecque, and R. Nikoukhah: LMITOOL: A

User-friendly Interface for LMI Optimization, ENSTA/INRIA
(1995)
P.P. Khargonekar and A.R. Mario: Mixed H2 /H Control:
A Convex Optimization Approach, IEEE Trans. Automatic
Contr., Vol.36, No.7, pp.824837 (1991)
M. Chilali and P. Gahinet: H Design with Pole Placement
Constraints: An LMI Approach, IEEE Trans. Automatic
Contr., Vol.41, No.3, pp.358367 (1996)
C. Scherer, P. Gahinet, and M. Chilali: Multiobjective
Output-Feedback Control via LMI Optimization, IEEE
Trans. Automatic Contr., Vol.42, No.7, pp.896911 (1997)
Y.N. Yu: Electric Power System Dynamics, Academic Press,
New York (1983)
P.M. Anderson and A.A. Fouad: Power System Control and
Stability, Iowa State University Press, Iowa (1977)
Y.Y. Hsu and C.L. Chen: Identication of Optimum Location for Stabilizer Applications Using Participation Factors,
IEE Proc. Part C, Vol.134, No.3, pp.238244 (1987)
M.E. Aboul-Ela, A.A. Sallam, J.D. McCalley, and A.A. Fouad:
Damping Controller Design for Power System Oscillation Using Global Signals, IEEE Trans. Power Syst., Vol.11, No.2,
pp.767773 (1996)

Appendix
(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

(9)

(10)

(11)

(12)

(13)

(14)

(15)

(16)

F.P. De Mello and C. Concordia: Concepts of Synchronous

Machine Stability as Aected by Excitation Control, IEEE
Trans. Power Apparatus Syst., Vol.88, No.4, pp.316325
(1969)
P. Kundur, M. Klein, G. Rogers, and M. Zywno: Application
of Power System Stabilizers for Enhancement of Overall System Stability, IEEE Trans. PWRS, Vol.4, No.2, pp.614626
(1989)
S. Lefebvre: Tuning of Stabilizers in Multimachine Power
Systems, IEEE Trans. Power Apparatus Syst., Vol.102,
No.2, pp.290299 (1983)
T. Hiyama: Rule-Based Stabilizers for Multi-machine Power
System, IEEE Trans. Power Syst., Vol.5, No.2, pp.403411
(1990)
Y.M. Park, S. Hyun, and J.H. Lee: Power System Stabilizer
Based on Inverse Dynamics Using an Articial Neural Network, Int. J. Electrical Power & Energy Systems, Vol.18,
No.5, pp.297305 (1996)
T. Hiyama: Robustness of Fuzzy Logic Power System Stabilizers Applied to Multimachine Power Systems, IEEE Trans.
Energy Conversion, Vol.9, No.3, pp.451459 (1994)
S. Chen and O.P. Malik: H Optimisation-Based Power
System Stabilizer Design, IEE Proc. Part C, Vol.142, No.2,
pp.179184 (1995)
S. Chen and O.P. Malik: Power System Stabilizer Design Using Mu Synthesis, IEEE Trans. Energy Conversion, Vol.10,
No.1, pp.175181 (1995)
R. Asgharian: A robust H -Power System Stabilizer with
No Adverse Eect on Shaft Torsional Modes, IEEE Trans.
Energy Conversion, Vol.9, No.3, pp.475481 (1994)
S.S. Ahmed, L. Chen, and A. Petroianu: Design of SubOptimal H Excitation Controllers, IEEE Trans. Power
Syst., Vol.11, No.1, pp.312318 (1996)
M. Rios, N. Hadjsaid, R. Feuillet, and A. Torres: Power System Stability Robustness Evaluation by Analysis, IEEE
Trans. Power Syst., Vol.14, No.2, pp.648653 (1999)
K.M. Son and J.K. Park: On the Robust LQG Control of
TCSC for Damping Power System Oscillations, IEEE Trans.
Power Syst., Vol.15, No.4, pp.13061312 (2000)
P.S. Rao and I. Sen: Robust Pole Placement Stabilizer Design Using Linear Matrix Inequality, IEEE Trans. Power
Syst., Vol.15, No.1, pp.313319 (2000)
J.C. Doyle, K. Glover, P.P. Khargonekar, and B.A. Francis:
State-Space Solutions to Standard H2 and H Control Problems, IEEE Trans. Automatic Contr., Vol.34, No.8, pp.831
847 (1989)
S. Boyd, L.E. Ghaoui, E. Feron, and V. Balakrishnan: Linear
Matrix Inequalities in Systems and Control Theory, SIAM,
Vol.15, Philadelphia, PA (1994)
P. Gahinet, A. Nemirovski, A. Laub, and M. Chilali: The LMI
Control Toolbox, The Mathworks Inc. (1995)

B124 7 2004

System equations and machine parameters

Generator:
1
(Tm Te + D( 1)) (A1)
M
= b ( 1) (A2)

1 
E q =  Ef d (xd xd )id Eq (A3)
Td0
=

Te = Eq iq + (xq xd )id iq (A4)

Exciter:
1
E f d =
{KA (vref vt + ue ) Ef d } (A5)
TA

app. Fig. 1. Block diagram of conventional PSS

used for comparison
app. Table 1.

Machine data

app. Table 2.

Exciter data

app. Table 3.

929

CPSS constants

Hardiansyah (Student Member) was born on February 27,

1967 in Mempawah, Indonesia. He received
the B.S. degree in Electrical Engineering from
the University of Tanjungpura in 1992 and
the M.S. degree in Electrical Engineering from
Bandung Institute of Technology (ITB), Indonesia in 1996. Since 1992, He has been with
Department of Electrical Engineering University of Tanjungpura, Pontianak, Indonesia,
where he is a lecturer. Currently he is pursuing a Ph.D. degree in Electrical Engineering Nagaoka Univiversity
of Technology. His current research interests include power system
dynamics and controls, robust control, and LMI approach. He is
a student member of the IEE of Japan.

Juichi Irisawa (Member) was born on February 5, 1941. He

received the Ph.D. in 1969 in electrical engineering from Tokyo Institute of Technology.
He was an assistant at Tokyo Institute of Technology from 1969 to 1978. He was an assistant
professor at Nagaoka University of Technology
from 1978 to 1992. He has been a professor at
Nagaoka University of Technology since 1992.
He has been engaged in the research of electrical power system and pulsed power technology.
He is a member of IEEJ and the Japan Society of Plasma Science
and Nuclear Fusion Research.

Seizo Furuya (Member) was born on July 30, 1968. He

received the M.E. degree in 1993 from Tokyo
Institute of Technology. After he quit the doctor course of Tokyo Institute of Technology in
1997, he has been an assistant at Nagaoka University of Technology. He has been engaged
in the research of electrical power system and
pulsed power technology. He is a member of
IEEJ, the Japan Society of Applied Physics,
the Physical Society of Japan and the Japan
Society of Plasma Science and Nuclear Fusion Research.

930

IEEJ Trans. PE, Vol.124, No.7, 2004