This article has been accepted for inclusion in a future issue of this journal.

Content is final as presented, with the exception of pagination.

IEEE TRANSACTIONS ON POWER SYSTEMS 1

Robust ESS-Based Stabilizer Design for Damping

Inter-Area Oscillations in Multimachine
Power Systems
Linjun Shi, Member, IEEE, Kwang Y. Lee, Life Fellow, IEEE, and Feng Wu

Abstract—In this paper, a robust approach to the energy storage However, it is shown that interactions occur between stabilizers
system (ESS)-based stabilizer is proposed to select its installation in multimachine power systems, where the stabilizers being
location, feedback signals, and damping control loop, and to tune PSSs, FACTS device stabilizers or both [7]. In [8], a novel
the stabilizer parameters. Based on the linearized model of the
power system with ESS and application of damping torque anal- algorithm is presented for simultaneous coordinated design of
ysis (DTA), the maximum of all minimum damping torque indices PSSs and a thyristor-controlled series capacitor (TCSC) in a
(DTIs) under multi-operational conditions is used as the index for multimachine power system for damping power system oscilla-
selecting ESS installation locations, feedback signals and damping tions. In [9], different input–output controllability analyses are
control loops. The operational condition for the minimum DTI pro- used to assess the most appropriate input signals for the static
vides the condition for tuning the ESS-based stabilizer parame-
ters for robust operation. And a proper compensation angle is se- var compensators (SVCs), the static synchronous compensator
lected by a robust method to design the ESS-based stabilizer pa- (SSSC), and the unified power-flow controller (UPFC) for
rameters. The eigenvalue analysis and non-linear simulation re- achieving good damping of inter-area oscillations. Taking
sults of a four-machine power system and New England ten-ma- advantages of the FACTS devices depends largely on how
chine power system with ESS show that power system oscillations these devices are placed in the power system, namely, on their
are suppressed effectively and robustly by the ESS-based stabi-
lizer. location and size. In [10], power system stability is used as an
index for optimal allocation of the controllers. After placing the
Index Terms—Damping torque indices (DTIs), energy storage
SVCs based on their primary functions, the most appropriate
system (ESS), inter-area oscillations, multimachine power systems,
robust stabilizer. input signal supplementary controller is also selected.
With the advancement of energy storage technologies in re-
cent years, energy storage systems (ESSs) are applied in power
I. INTRODUCTION systems widely [11]–[14], such as to improve system stability
[15], [16], power quality [17], and reliability of supply [18].
Since the inter-area oscillation between two connected power
D AMPING inter-area low-frequency oscillations in large
scale power systems has been one of the most important
research topics for several decades [1]–[4]. Power system
systems is the oscillation of active power, damping oscillation
with ESSs by exchanging active power directly can be a more
stabilizers (PSSs) have been proven to be effective in damping efficient and effective way than PSSs or FACTS [19]. And since
local-area oscillation. While PSSs have limited effectiveness selecting installation location for ESSs is more flexible than
for inter-area damping control, inter-area oscillation still re- for PSSs, there have been a growing interest in the applica-
mains a critical challenge in today's power systems [5]. Many tion of ESSs on damping low frequency inter-area oscillations
other approaches have been proposed for damping inter-area [20]–[22]. For example, a battery energy storage system (BESS)
low-frequency oscillations. For example, a comprehensive re- is used for damping inter-area oscillations in a multimachine
view is presented in [6] on the research and developments in the power system [20], a candidate energy storage system based
power system stability enhancement using flexible alternative on ultra-capacitor technology is evaluated for damping control
current transmission systems (FACTS) damping controllers. applications in real power systems [21], and the flywheel en-
ergy storage system (FESS) is proposed for damping power
system oscillations in a single-machine infinite-bus power sys-
Manuscript received October 11, 2014; revised January 12, 2015; accepted
tems [22]. In [20], it is explained how and why the ESS-based
April 07, 2015. This work was supported in part by the National Natural Sci- stabilizer can suppress the inter-area oscillation. In [21], an ana-
ence Foundation of China (NSFC) under Grant 51422701, 51137002. Paper no. lytical method for ensuring proper stability margins is provided.
TPWRS-01399-2014.
L. Shi and F. Wu are with the College of Energy and Electrical Engineering,
In [22], the principle of the complex torque coefficient (CTC)
Hohai University, Nanjing, Jiangsu 210098, China (e-mail: [email protected]; method is applied to analyze the use of FESS for damping os-
[email protected]). cillation in a single-machine infinite-bus system. However, all
K. Y. Lee is with the Department of Electrical and Computer En-
gineering, Baylor University, Waco, TX 76798-7356 USA (e-mail:
of these ESS-based stabilizers do not consider the robustness is-
[email protected]). sues as a whole in damping power system oscillations, such as
Color versions of one or more of the figures in this paper are available online selection of the ESS installation locations, feedback signals and
at http://ieeexplore.ieee.org.
damping control loops, and tuning of the stabilizer parameters.
Digital Object Identifier 10.1109/TPWRS.2015.2422797

0885-8950 © 2015 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.
See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

2 IEEE TRANSACTIONS ON POWER SYSTEMS

Damping torque analysis (DTA) was introduced in [2] on the

basis of the Phillips–Heffron model [23] for a single-machine
infinite-bus power system. It was further developed in [24] into
the well-known phase compensation method for the design of
a PSS. Since the 1980s, there has been much effort to extend
the phase compensation method to design PSSs [25]–[27], and
FACTS stabilizers [28], [29] in a multimachine power system.
In our paper, both the DTA and phase compensation method
based on linearized model will be applied to design a robust
ESS-based stabilizer in damping inter-area oscillations in mul- Fig. 1. Configuration of FESS based on DFIM.
timachine power systems since a linearized model is well ac-
cepted for studying power system oscillations [3], [13].
As a candidate energy storage system, an ESS based on the
flywheel storage technology consisting of a doubly-fed induc-
tion machine (DFIM) is considered to demonstrate the robust
selection of the ESS installation locations, feedback signals and
damping control loops, and the robust tuning of the parameters
of the ESS-based stabilizer. Based on the linearized model of the
Fig. 2. ESS connected to a power system.
power system with ESS and application of DTA, the maximum
of all minimum damping torque indices (DTIs) under multi-op-
erational conditions, as far as the inter-area mode of power os- and are, respectively, the currents of the stator and the rotor
cillation is concerned, is found to be the index for the best se- of the ESS.
lection of ESS installation locations, damping control loops and
feedback signals. And the operation condition for the minimum A. ESS Nonlinear Model
change of the real parts of eigenvalues provides the condition The third-order model of DFIM is adequate for oscillations
for tuning the ESS-based stabilizer parameters for robust op- studies [31]. The ESS based on DFIM has been represented by
eration in multi-operational conditions. The eigenvalue anal- a third-order nonlinear model [32] as follows.
ysis and nonlinear simulation results of a four-machine power 1) Stator Voltage Equation:
system and New England ten-machine power system with ESS
show that power system oscillations are suppressed effectively (1)
by the robust ESS-based stabilizer.
where and are transient electric potential in the -axis,
II. FLYWHEEL ENERGY STORAGE SYSTEM , , , and are, respectively, the stator voltage and
current in the -axis, is stator resistance, and is the tran-
The FESS consisting of DFIM has several advantages. For sient reactance equal to , with
example, it can exchange relatively larger amounts of energy and , where is stator winding self-induc-
compared with a battery or ultra-capacitor and it has quick en- tance, is rotor winding self-inductance, and is the mutual
ergy release and storage capability; therefore, it is very fitting inductance between stator and rotor windings.
for large-scale power system stability control. Moreover, the re- 2) Dynamic Equation: Transient electric potential dynamic
quired capacity of a power electronics component for the DFIM equation can be obtained as
is in a range from one-fifth to one-seventh as small as the ca-
pacity of the machine [30]. Therefore, it is cheaper to install
and operate [30]. The ESS can release and absorb active power. (2)
Inter-area oscillation is the oscillation of active power between
two connected power systems. The ESS can be controlled to bal- where is the slip of DFIM, is the rotor open circuit time
ance the power oscillation, damping the inter-area modes of os- constant, and is synchronous speed. Similar to synchronous
cillations quickly. This paper’s main work is to ensure the ESS generator, the swing equation is given by
releasing and absorbing power correctly through phase compen-
sation. (3)
In the ESS based on the flywheel storage technology con-
sisting of DFIM, the energy is stored in the rotation of the where is inertia constant, is the electromagnetic torque
rotor. The configuration of a FESS based on DFIM is shown equal to , and is the mechanical torque.
in Fig. 1 [30]. Regulating the speed of the flywheel and energy For ESS, .
exchange between ESS and the power system can be realized Thus ESS based on DFIM can be represented by the third-
through the DFIM. order dynamic model in (2), (3).
The ESS connected to a power system is shown in Fig. 2. 3) Decoupled Control Loops: Based on the orientated con-
In Fig. 2, is the terminal voltage of the ESS, is the trol strategy of stator magnetic field, the ESS can be realized
negative current which is injected to the power system, and with the active and reactive power decoupled control [33], [34].
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

SHI et al.: ROBUST ESS-BASED STABILIZER DESIGN FOR DAMPING INTER-AREA OSCILLATIONS IN MULTIMACHINE POWER SYSTEMS 3

Fig. 3. Block diagram of active control loop.

Fig. 4. Block diagram of reactive control loop.

Fig. 6. Block diagram of linearized power systems with ESS.

power system [3], [13], details of deriving the linearized power

system model with ESS is given in Appendix A. Rearranging
(A15), the linearized power system with ESS is
Fig. 5. ESS-based stabilizer superimposed on the decoupled control loop.

The block diagrams of the control loops for the active and reac-
tive power controls are shown in Fig. 3 and Fig. 4, respectively. (4)
In the figures, , , and are, respectively, the rotor where is the deviation of generator power angle, is the
power, stator power, and reference power, , , and deviation of generator speed, are all state variables except
are, respectively, the rotor voltages in the -axis, and for the power angle and generator speed variables, is the
the stator voltage reference in the -axis, and , , deviation of damping signal added to ESS output from ESS-
, and are the coefficients of the respective PI based stabilizer, and is the transfer function matrix from
controllers. to state variables.
4) ESS-Based Stabilizer: In order to damp power system The output can be expressed as
low-frequency oscillations, the structure of PSSs [24] or
FACTS-based stabilizers [7], [28] can be superimposed upon (5)
the active power or reactive power control loop of ESS, called
ESS-based stabilizer. Without loss of generality, the transfer where is the coefficient matrix from state variable to .
function model of the stabilizer [2], [3], [20], [24] is given in Output of the stabilizer is given as
Fig. 5, where is either or , depending on the real or
reactive control loop. If the input signal has information of (6)
system oscillation, the output signal of the ESS-based stabilizer,
, acts on the ESS through the control loop, and the power Thus, (4)–(6) are made up of the linearized model of a multi-
system oscillation can be suppressed quickly by storing or machine power system with the ESS. The block diagram of the
releasing energy to or from the ESS. linearized system is shown in Fig. 6.
In the figure, , , , , and are time constants, is In Fig. 6, is an identity matrix, and is the differential op-
gain, , , and are intermediate variables, and is the erator.
transfer function of the ESS-based stabilizer.
III. ROBUSTNESS ISSUES OF ESS-BASED STABILIZER
B. Full System Linearized Model
According to Fig. 2, and applying the bulk current injection A. Damping Torque Index
method [35], [36], can be expressed by all ESS’s state vari- From (4) and Fig. 6, the forward path of the output control
ables and bus voltages; meanwhile, is a branch current of signal of the ESS-based stabilizer to the electromechanical os-
the power system network. Thus, ESS can be integrated into the cillation loop can be obtained as
power system. Since analysis and control of power system os-
cillation stability is mainly based on the linearized model of a (7)
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

4 IEEE TRANSACTIONS ON POWER SYSTEMS

and the output can be restructured by rotor speed as [37] From (11b) and (11c), we know that is changed by the
ESS-based stabilizer. The range of change can be described by
DTI defined as follows:
(8)
(14)
where is restructuring function. In Appendix B, the detail
of deriving of is given. Thus The DTI is defined as the effects of the ESS-based stabilizer
on the real part of the eigenvalue. Therefore, if the value of
(9) the DTI is higher, the ESS-based stabilizer can provide more
damping torque which results in more improvement of the
From Fig. 6, (7), and (9), the damping torque [2] provided damping of the oscillation mode, that is, (14) shows the effect
by the ESS-based stabilizer to the electromechanical oscillation of the ESS-based stabilizer, which can be used for selection
loop of the th generator in the power system is of the best installation location, damping control loop, and the
feedback signal.
(10a)
(10b) B. Robust Selection Method
A conventional method for selecting ESS installation loca-
(10c) tion, control loop, and the feedback signal is to use

where is the inter-area oscillation mode of interest at an op- (15)

erating condition, is the th element of in (7), , where is the collection of all candidates of installation loca-
and is the number of generators; , tions, control loops, and the feedback signals, is the result of
and . the choice, , and is a typical operating condition.
From (10a) and (10b), we know that the damping torque con- However, the DTI will change under different operating con-
tributed by the ESS-based stabilizer is changed in the variations ditions even if in the same installation location, control loop and
of system operating conditions and it is acted on every gen- with the feedback signal. Therefore, (15) is not always the best
erator through a single forward channel . So there are choice under a typical operating condition.
total channels through which the ESS-based stabilizer pro- In this paper, an alternative method of selecting ESS instal-
vides all generators with the damping torque. However, if the lation location, control loop, and feedback signal is defined as
ESS-based stabilizer for a generator provides a damping torque follows:
which results in little improvement of the damping of an oscil-
lation mode, the generator is not really participating in damping (16)
the oscillation mode. Thus, for an oscillation mode , we can
define the sensitivity of with respect to an addition of the where is a collection of given operating conditions which
damping torque with the th generator as . often include load condition from light to heavy load according
Thus, the total improvement of the damping of the oscillation to inter-area modes of power oscillation concerned. There is
mode due to the addition of a damping torque with all gener- a minimum value of DTI under different operating conditions
ators is for every element of . Then choose an operating condition
which has maximum of all the minimum values of
DTI. And is the result of selecting the ESS installation
(11a)
location, control loop and the feedback signal under selected op-
erating condition . Equation (16) shows that the selection re-
sult is the least effective in damping the th oscillation mode
(11b) at the operating condition . Thus, if we choose as
the operating condition to design the ESS-based stabilizer, an
(11c) effective design of the ESS-based stabilizer at can ensure its
effectiveness for all other operating conditions in . There-
where . fore, (16) is a robust method for selecting the ESS installation
The transfer function of the stabilizer in Fig. 5 can be ex- location, control loop and the feedback signals.
pressed as:
C. Robust Tuning of ESS-based Stabilizer Parameters
(12) Using the phase compensation method [24], [29], the param-
eters of the ESS-based stabilizer can be tuned, which should
From (11c) and (12), we know stabilizer affects the real part of
ensure that the power systems oscillations can be suppressed
the as:
effectively under various operating conditions.
If
(13a)
(13b) (17)
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

SHI et al.: ROBUST ESS-BASED STABILIZER DESIGN FOR DAMPING INTER-AREA OSCILLATIONS IN MULTIMACHINE POWER SYSTEMS 5

where , then (17) shows that the least effective

damping of the th oscillation mode is provided at the operating
condition . If we choose as the operating condi-
tion to tune the ESS-based stabilizer parameters, more effective
damping will be obtained at all other operating conditions in
. Thus, is the robust operating condition for tuning the
ESS-based stabilizer parameters.
From (13b), if a compensation angle causes moving left- Fig. 7. Four-machine power system with an ESS.
ward, the system damping of the oscillation is increased. Thus,
the key problem is how to select a proper compensation angle TABLE I
to ensure moving leftward maximally at all operating con- VALUE OF DTI UNDER DIFFERENT OPERATING CONDITIONS
ditions.
Since the robust operating condition is selected by (17),
the process of selecting the compensation angle is given here.
1) In order to ensure is moving leftward at all operating
conditions and provide positive synchronous torque [2],
[37], the range is

(18)

2) For any in (18), calculate through (13b) for all operating

conditions included in . Then, there exists a minimum and reactive power controls parameters are: ,
value given by , , , and
[39].
(19)
A. Robust Parameter Tuning of ESS-Based Stabilizer
where , is the number of included in An ESS is installed at bus 7, and the oscillation power on line
(18). 7–8 is taken as the feedback signal, which is applied to the active
3) Define. Then, there exists power control loop.
(20) The value of DTI in different operating conditions is shown
in Table I.
(21) From Table I, it can be seen that DTI of
has the minimum DTI value. Thus, is the robust op-
From above, we can know that can change the max- erating condition for the design of the ESS-based stabilizer ac-
imum in all minimum changes of the real part of eigenvalues in cording to (17).
operating condition selected by (17), that is, it can The next step is to select a proper compensation angle .
provide maximum DTI to all operating conditions in at Based on (18), the range of is
when designing the stabilizer. Thus, is the proper , that is, .
compensation angle to provide maximum damping for all oper- From Table I, is from 83.091 to 87.792 . The range of
ating conditions in . is relatively narrow. According to (19) and (20), if com-
Therefore, having obtained a proper compensation angle in pensation angle deviates 180 too much, the value of
an operating condition, the robustness in the design of an ESS- will be too small to find . Thus, in order to simplify
based stabilizer to the variations of power system operating con- calculation, assume , such that a given can-
ditions is achieved. The stabilizer parameters can be tuned by didate is in
(12)–(13b) according to [24]. . Ac-
cording to (19), for any , there exists a minimum value,
IV. CASE STUDY I: FOUR-MACHINE SYSTEM . So the is
An example of a four-machine power system is shown in . The re-
Fig. 7 [38]. The collection of different operating conditions sults are also shown in Table I. When ,
is 400 MW 420 MW , where there exists . Thus,
is represented for the active power transferred along the is the proper compensation angle for providing the maximum
double-circuit line 7–8. The typical collection of different op- damping to all operating conditions in .
erating conditions is 420 MW The ESS-based stabilizer parameters are tuned by the phase
250 MW 125 MW compensation. Consider the following settings: target damping
150 MW 250 MW 400 MW . ratio is 0.1, is 5 s, and lag-lead time constant are 0.05
Parameters of an example of ESS based on flywheel are as s. Then, the robust ESS-based stabilizer parameters are tuned as
follows [30]: capacity is 70 MVA, is 0.0013 p.u., , by the phase compensation
is 2.9 p.u., is 2.6 p.u., and is 3.4 s. The active method [28]. In order to verify the damping effect, the results
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

6 IEEE TRANSACTIONS ON POWER SYSTEMS

TABLE II TABLE III

EIGENVALUES UNDER DIFFERENT OPERATING CONDITIONS VALUES OF DTI IN

Fig. 8. Damping effects under different operating conditions.

of eigenvalue and nonlinear simulation are respectively shown

in Table II and Fig. 8.
In Table II, damping ratios at all operating conditions are
better than that at . In Fig. 8, the oscillations are triggered
Fig. 9. DTI of various candidates in different operating conditions.
by a three-phase short circuit occurring at the end of line 9–10
at 0.1 s and removed after 100 ms. Nonlinear simulation results
also show that all system oscillations are suppressed effectively
is , and stabilizer control loop is a voltage control loop, i.e.,
at all other operating conditions with the ESS-based stabilizer
a reactive power control loop.
designed at the operating condition . Table II and Fig. 8 both
For the robust selection method proposed in this paper, the
show that this is a robust stabilizer. Therefore, the method pro-
values of DTI should be calculated first for all candidates in
posed provides a robust method of tuning the ESS-based stabi-
the group in different operating conditions. The results are
lizer parameters.
shown in Fig. 9.
In Fig. 9, every curve delegates the DTI values of a candidate
B. Robust Selection of Installation Location, Feedback Signal
in in different operating conditions. And the colored dot
and Control Loop
represents the minimum DTI value of the candidate for each
The candidate group is selected as follows. operating conditions. From Table III and Fig. 9, we observe the
• Candidate ESS locations: bus 7, bus 8 and bus 9. following.
• Candidate stabilizer feedback signals: between gener- 1) is the robust selection among all candidates in . Be-
ator 1 and generator 3, and power oscillations on line 7–8. cause the minimum DTI value of candidate is the max-
• Candidate stabilizer control loops: active power control imum in all minimum DTI values of candidates in .
loop and reactive power control loop. Thus, according to (16), candidate is the best choice,
The general method of selection is based on a typical oper- that is, the ESS should be installed at bus 7, the feedback
ating condition, such as . The values of DTI are shown in signal should be , and the stabilizer control loop is the
Table III under the operating condition . active power control loop.
From Table III, maximum DTI is 0.231 at . According to 2) Feedback signal is better than feedback signal .
(15), candidate is the best choice by the conventional method, 3) The ESS installed at bus 7 and bus 9 is better than that at
that is, the ESS installation location is at bus 7, feedback signal bus 8. This shows that ESS installed at a terminal of the
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

SHI et al.: ROBUST ESS-BASED STABILIZER DESIGN FOR DAMPING INTER-AREA OSCILLATIONS IN MULTIMACHINE POWER SYSTEMS 7

TABLE IV TABLE VI
VALUE OF DTI UNDER DIFFERENT OPERATING CONDITIONS AT EIGENVALUES IN OPERATIONS AND

TABLE V
PARAMETERS OF TWO STABILIZERS

the design requirement only under operation condition and

; while in , damping ratio meets the design requirement
under all operating conditions. Obviously, higher stabilizer gain
transmission line is better than one installed in the middle is needed in to meet the requirement under all other operating
of the line. conditions. Thus, control cost will be increased at . Thus,
Obviously, the selection result is different for the conven- is a better choice.
tional selection method (15) and the robust method proposed Nonlinear simulation also is used to compare the two selec-
(16). The result of the robust method is while the result of the tion methods. Simulation condition is a three-phase short cir-
conventional method is . Then, the question is: “which selec- cuit occurring at the end of line 9–10 at 0.1 s and removed after
tion result is better?” These can be validated through eigenvalue 100 ms. The results of the simulation are shown in Figs. 10 and
analysis and nonlinear simulation. 11.
The robust tuning method was discussed in Section III to tune From Figs. 10 and 11, damping effect of is better than
the ESS-based stabilizer parameters for two selection results that of , and thus is a better choice. From Table VI and
and . In order to compare in the same condition, there must be Figs. 10 and 11, it is validated that the selection method pro-
the same gain of the stabilizer in (12) for the two selection posed in this paper is effective and robust.
results, and . In the result of in Section III, the value of
DTI in different operating conditions has been shown in Table I, V. CASE STUDY II: TEN-MACHINE SYSTEM
and operating condition is the robust operating condition for In order to further verify the validity and generality of the
tuning the ESS-based stabilizer parameters, where proposed robust method, a larger sample of the New England
is the proper compensation angle. ten-machine system is given. The system connection diagram is
In the result of , the values of DTI under different operating shown in Fig. 12. Parameters of the system are given in [40].
conditions are shown in Table IV. The parameters of FESS are same as those in Section IV.
From Table IV it can be seen that DTI of The New England ten-machine system is a typical intercon-
has the minimum value. So is the robust operation condi- nected system and has been widely used for studying the low
tion for the design of ESS-based stabilizer according to (17). frequency oscillation [4]. The governor dynamics are neglected
Also, from Table IV, it can be seen that when while each generator is equipped with an IEEE DC1A excita-
, there exists the . Thus tion system except generator 10 which is an equivalent gener-
is the proper compensation angle in providing ator, and the loads are modeled as constant impedances. The
maximum damping for all operating conditions in . initial load condition is 120 MW, and there exists an
The two ESS-based stabilizers parameters are tuned by the inter-area oscillation mode [4]. The
phase compensation method. And they have the same . If typical collection of different operating conditions is
is 1.407 and 0.05 s, 5 s, the results of 120 MW
other parameters of the two stabilizers are shown in Table V. 160 MW 200 MW according to different load
Eigenvalues are shown in Table VI under different operating conditions from light to heavy load.
conditions.
From Table VI, although the two stabilizers have the same A. Robust Parameter Tuning of ESS-based Stabilizer
gain, the damping ratio of the selection result by the conven- For suppressing the poor inter-area power oscillation mode
tional method is worse than the selection result by the ro- , an
bust method proposed in this paper. In , damping ratio meets ESS-based stabilizer is installed at bus 2. And the oscillation
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

8 IEEE TRANSACTIONS ON POWER SYSTEMS

TABLE VII
VALUE OF DTI UNDER DIFFERENT OPERATING CONDITIONS

TABLE VIII
EIGENVALUES UNDER DIFFERENT OPERATING CONDITIONS

Fig. 10. Nonlinear simulation of and in .

power on line 2 to 1 is taken as the feedback signal, which is

applied to the active power control loop.
The value of DTI in different operating conditions is shown
in Table VII (other oscillation modes are not listed).
From Table VII, we know is the robust operating con-
dition for the design of the ESS-based stabilizer according to
(17), and is the proper compensation angle for
providing the maximum damping to all operating conditions
in based on (18). Then, the robust ESS-based stabilizer
parameters are tuned as , ,
10 s, 0.5 s for target damping ratio to 0.1
by the phase compensation method [24]. Modal analysis of the
system shows that damping ratios at all operating conditions are
suppressed and better than the one at in Table VIII.
General method of tuning parameters is based on an oper-
Fig. 11. Nonlinear simulation of and in .
ating condition. For example, in , the compensation angle
is , the ESS-based stabilizer parameters are tuned as
, , 10 s,
0.5 s for target damping ratio to 0.1 by the phase compensa-
tion method [24]. Eigenvalue analysis in vary operating condi-
tions with the ESS-based stabilizer shows in Table VIII. From
Table VIII, we know that damping ratio meets the design re-
quirement only at operating condition .
Therefore, the method proposed provides a robust method of
tuning the ESS-based stabilizer parameters.

B. Robust Selection of Installation Location, Feedback Signal

and Control Loop
For simplicity, the candidate group is selected as
follows.
• Candidate ESS locations: bus 2, bus 9.
• Candidate stabilizer feedback signals: between gener-
ator 2 and generator 10, and power oscillations on line 2-1.
• Candidate stabilizer control loops: active power control
loop and reactive power control loop.
Application of the robust selection method proposed in this
Fig. 12. Ten-machine power system with an ESS.
section, the minimum DTI values of the candidate for each op-
erating conditions are shown in Table IX.
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

SHI et al.: ROBUST ESS-BASED STABILIZER DESIGN FOR DAMPING INTER-AREA OSCILLATIONS IN MULTIMACHINE POWER SYSTEMS 9

TABLE IX TABLE XII

VALUES OF MINIMUM DTI UNDER DIFFERENT OPERATING CONDITIONS EIGENVALUES IN OPERATIONS AND

TABLE X
VALUE OF DTI UNDER DIFFERENT OPERATING CONDITIONS AT

TABLE XI
PARAMETERS OF TWO STABILIZERS

Fig. 13. Nonlinear simulation of and in and .

In order to compare, all DTI values are also calculated under

the operating condition . The results are shown in Table IX.
From Table IX, the result (location: bus 2, feedback signal:
, control loop: active power control loop) is selected based
on the robust selection method proposed in the paper, while the
result (location: bus 2, feedback signal: , control loop:
voltage control loop) is selected in operating condition , and
eigenvalue analysis and nonlinear simulation are used to vali-
date which selection result is better.
In order to compare, the ESS-based stabilizers for two selec-
tion results and are both tuned in the same condition by
the robust tuning method proposed in the paper. Both ESS-based
stabilizers have the same gain in (12). In the selection result
, the ESS-based stabilizer tuned by the robust tuning method
has .
In the selection result , the values of DTI in different oper-
ating conditions are shown in Table X. Fig. 14. Nonlinear simulation of and in and .
From Table X, we know is the robust operating condition
for the design of the ESS-based stabilizer at according to
(17). And is the proper compensation angle for From Table XII, although the two stabilizers have the same
providing the maximum damping torque to all operating con- gain, the damping ratio of the selection result by the conven-
ditions in based on (18). The two ESS-based stabilizers tional method is worse than the selection result by the robust
parameters are tuned by the phase compensation method. And method proposed in this paper.
they have the same . If is 1.36 and 10 s, the results Nonlinear simulation also is used to compare the two selec-
of other parameters of the two stabilizers are shown at Table V. tion results. Simulation condition is a three-phase short circuit
The results of other parameters of the two stabilizers are shown occurring at Bus 16 and removed after 100 ms. The results of
in Table XI. the simulation are shown in Figs. 13 and 14.
Eigenvalues are shown in Table XII under different operating From Figs. 13 and 14, the damping effect of is better than
conditions at and . that of , and thus is a better choice.
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

10 IEEE TRANSACTIONS ON POWER SYSTEMS

Thus, ,
, where and are the current, and and
are the voltage in the -axis. Then, from the linearized
equation (1), we can have

(A3)

where is deviation symbol, ,

, is the ESS state variable equal to
, and and are the coefficient matrices.
From Fig. 2, if PWM is ideal and power factor is unity, then
has the same phase with in the -axis. Thus, can
Fig. 15. Orientated control strategy of stator magnetic field.
be expressed as

From Table XII and Figs. 13and 14, it is demonstrated that the
selection method proposed in this paper is effective and robust (A4)
for a large-scale power system. where is rotor active power, is stator active power, and
.
VI. CONCLUSION Linearizing (A4) in -axis yields
This paper presents a robust design procedure of the ESS- (A5)
based stabilizer for damping power systems inter-area oscilla-
tions. It includes a robust selection procedure of the installation where , and , are the currents in
location, feedback signal, and control loop of the EES-based -axis, and and are the coefficient matrices.
stabilizer and a robust tuning procedure of the ESS-based sta- From (A1), (A3), and (A5), the linearized current of the ESS
bilizer parameters. The paper presents the concept of DTI as connected to power system is
a measure of robust selection criteria for selecting installation
location, feedback signal, and control loop of the EES-based (A6)
stabilizer. The minimum DTI under different operating condi-
tions combined with the improved phase compensation method where and are the coefficient matrices, with
for proper compensation angle is presented as a robust tuning , . Equation (A6) shows that the current
method. injection of the ESS to power system can be expressed in terms
The robust design procedure for the EES-based stabilizer is of the ESS state variables and the port voltage of the ESS.
demonstrated in two example power systems and its effective- According to Figs. 3, 4, and 6, the dynamic equations of ESS,
ness and robustness are verified by eigenvalue calculation and (2) and (3), are linearized as
nonlinear simulation, for both small and large-scale power sys-
tems. (A7)
Further studies will focus on coordinating design and tuning
of ESS-based stabilizer and PSS for damping multiple inter-area where and are coefficient matrices.
and local-area oscillation modes. For the whole power system, the network equation is

APPENDIX (A8)

A. Full System Linearized Model where is admittance matrix, is the nodal current injection,
and is the nodal voltage. Rearranging (A8) yields
According to Fig. 2, the current injection of the ESS to the
power system is
(A9)
(A1)

Ignoring the dynamics of the stator magnetic field and the where is the current of synchronous generators into the
ESS loss and based on the orientated control strategy of the system, is the port voltage of the synchronous generators,
stator magnetic field, the -axis of the ESS is located at (see and is the nodal load voltage.
Fig. 15), where the angle between the -axis of the ESS and the The injection current of synchronous generators is linearized
-axis of the system is [34]. as
From Fig. 15, the transformation matrix from the
axis to the axis can be obtained as (A10)

where is synchronous generator state variables equal

(A2)
to , with as the other state variables of
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

SHI et al.: ROBUST ESS-BASED STABILIZER DESIGN FOR DAMPING INTER-AREA OSCILLATIONS IN MULTIMACHINE POWER SYSTEMS 11

synchronous generators, and and are coefficient ma- From (5), (B3), and (B4), it can be seen that
trices. From (A6), (A10), and (A9), it can be seen that

(B5)

(B6)
(A11)
where , and .

Rearranging (A11), we have REFERENCES

[1] J. Ma, T. Wang, S. Wang, X. Gao, X. Zhu, Z. Wang, and J. S. Thorp,
(A12) “Application of dual Youla parameterization based adaptive wide-area
damping control for power system oscillations,” IEEE Trans. Power
Syst., vol. 29, no. 4, pp. 1602–1610, Nov. 2014.
where , , and [2] F. P. deMello and C. Concordia, “Concepts of synchronous machine
stability as affected by excitation control,” IEEE Trans. Power App.
Syst., vol. PAS-88, no. 4, pp. 316–329, Apr. 1969.
[3] G. Rogers, Power System Oscillations. Boston, MA, USA: Kluwer
Academic, 2000.
[4] Y. Zhang and A. Bose, “Design of wide-area damping controllers for
interarea oscillations,” IEEE Trans. Power Syst., vol. 23, no. 3, pp.
1136–1143, Aug. 2008.
[5] D. Molina, G. K. Venayagamoorthy, J. Liang, and R. G. Harley, “Intel-
ligent local area signals based damping of power system oscillations
using virtual generators and approximate dynamic programming,”
Synchronous generator dynamic equations [36] are linearized IEEE Trans. Smart Grid, vol. 4, pp. 498–508, 2013.
as [6] M. A. Abido, “Power system stability enhancement using FACTS con-
trollers: A review,” Arabian J. Sci. Eng., vol. 34, pp. 153–172, 2009.
[7] M. J. Gibbard, D. J. Vowles, and P. Pourbeik, “Interactions between, ef-
(A13) fectiveness of, power system stabilizers and FACTS device stabilizers
in multimachine systems,” IEEE Trans. Power Syst., vol. 15, no. 2, pp.
748–755, May 2000.
where and are coefficient matrices. From (A7) and [8] E. S. Ali and S. M. Abd-Elazim, “Coordinated design of PSSs and
(A13), it can be seen that TCSC via bacterial swarm optimization algorithm in a multimachine
power system,” Int. J. Electr. Power Energy Syst., vol. 36, pp. 84–92,
2012.
(A14) [9] M. M. Farsangi, Y. H. Song, and K. Y. Lee, “Choice of FACTS device
control inputs for damping interarea oscillations,” IEEE Trans. Power
where and are coefficient matrices. From (A12) and Syst., vol. 19, no. 2, pp. 1135–1143, May 2004.
[10] M. M. Farsangi, H. Nezamabadi-pour, S. Yong-hua, and K. Y. Lee,
(A14), we conclude that “Placement of SVCs and selection of stabilizing signals in power sys-
tems,” IEEE Trans. Power Syst., vol. 22, no. 3, pp. 1061–1071, Aug.
(A15) 2007.
[11] P. F. Ribeiro, B. K. Johnson, M. L. Crow, A. Arsoy, and Y. Liu, “En-
ergy storage systems for advanced power applications,” Proc. IEEE,
where . vol. 89, no. 12, pp. 1744–1756, Dec. 2001.
[12] V. A. Boicea, “Energy storage technologies: The past and the present,”
Proc. IEEE, vol. 102, no. 12, pp. 1777–1794, Dec. 2014.
B. Output Restructuring Method [13] W. Du, S. Gao, H. F. Wang, R. Dunn, and S. J. Cheng, “Power system
oscillations and energy storage systems,” Power Syst. Technol., vol. 14,
From (4) and Fig. 6, and not considering damping signal , pp. 113–114, 2008.
it can be seen that [14] R. Sebastián and R. Peña Alzola, “Flywheel energy storage systems:
Review and simulation for an isolated wind power system,” Renewable
(B1) Sustain. Energy Rev., vol. 16, pp. 6803–6813, 2012.
[15] P. Mercier, R. Cherkaoui, and A. Oudalov, “Optimizing a battery
energy storage system for frequency control application in an iso-
where is differential operator. Thus, can be obtained as lated power system,” IEEE Trans. Power Syst., vol. 24, no. 4, pp.
1469–1477, Nov. 2009.
(B2) [16] M. H. Ali, T. Murata, and J. Tamura, “Transient stability enhancement
by fuzzy logic-controlled SMES considering coordination with optimal
reclosing of circuit breakers,” IEEE Trans. Power Syst., vol. 23, pp.
Because of 631–640, 2008.
[17] L. Kai, X. Hongbing, M. Qian, and Z. Ji, “Hierarchy control of power
(B3) quality for wind-battery energy storage system,” IET Power Electron.,
vol. 7, pp. 2123–2132, 2014.
[18] Y. Xu and C. Singh, “Power system reliability impact of energy
(B2) becomes Storage integration with intelligent operation strategy,” IEEE Trans.
Smart Grid, vol. 5, pp. 1129–1137, 2014.
[19] X. Sui, Y. Tang, H. He, and J. Wen, “Energy-storage-based low-fre-
(B4) quency oscillation damping control using particle swarm optimization
and heuristic dynamic programming,” IEEE Trans. Power Syst., vol.
where . 29, no. 4, pp. 2539–2548, Nov. 2014.
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

12 IEEE TRANSACTIONS ON POWER SYSTEMS

[20] W. Du, H. F. Wang, J. Cao, and L. Y. Xiao, “Robustness of an energy [37] H. F. Wang, “Selection of operating conditions for the co-ordinated
storage system-based stabiliser to suppress inter-area oscillations in a setting of robust fixed-parameter stabilisers,” Proc. Inst. Electr.
multi-machine power system,” IET Gener., Transm. Distrib., vol. 6, Eng.—Gener., Transm. Distrib., vol. 145, pp. 111–116, 1998.
pp. 339–351, 2012. [38] P. Kundur, Power System Stability and Control. New York, NY,
[21] J. C. Neely, R. H. Byrne, R. T. Elliott, C. A. Silva-Monroy, D. A. USA: McGraw-Hill, 1994.
Schoenwald, D. J. Trudnowski, and M. K. Donnelly, “Damping of [39] L. Shi, G. Tang, and L. Zhang, “Parameter optimization of FESS PI
inter-area oscillations using energy storage,” in Proc. IEEE PES Gen. controllers,” Electric Power Autom. Equipment, vol. 31, no. 10, pp.
Meeting, Vancouver, BC, Canada, 2013, pp. 1–5. 65–69, 2011.
[22] Z. Chen, X. Zou, S. Duan, J. Wen, and S. Cheng, “Application of fly- [40] K. R. Padiyar, Power System Dynamics: Stability and Control, 2nd
wheel energy storage to damp power system oscillations,” Przegląd ed. Andhra Pradesh, India: BS Publications, 2008.
Elektrotechniczny (Electr. Rev.), vol. R, no. 87, pp. 333–337, 2011,
NR: 3.
[23] W. G. Heffron and R. A. Phillips, “Effect of a modern amplidyne Linjun Shi (M'11) received the B.S. and M.S. de-
voltage regulator on underexcited operation of large turbine genera- grees from Hohai University, Nanjing, China, in 1999
tors,” AIEE Trans. Power App. Syst., vol. 71, 1952. and 2003, respectively, and the Ph.D. degree from
[24] E. V. Larsen and D. A. Swann, “Applying power system stabilizers Southeast University, Nanjing, China, in 2010, all in
Part I–III,” IEEE Trans. Power App. Syst., vol. PAS-100, no. 6, pp. electrical engineering.
3017–3046, 1981. He is currently an Associate Professor with the
[25] M. J. Gibbard and D. J. Vowles, “Reconciliation of methods of compen- College of Energy and Electrical Engineering, Hohai
sation for PSSs in multimachine systems,” IEEE Trans. Power Syst., University, Nanjing, China. He is also a Visiting
vol. 19, no. 3, pp. 463–472, Aug. 2004. Scholar with the Department of Electrical and Com-
[26] C. Liu, R. Yokoyama, K. Koyanagi, and K. Y. Lee, “PSS design for puter Engineering, Baylor University, Waco, TX,
damping of inter-area power oscillations by coherency-based equiva- USA. His research interests include power system
lent model,” Int. J. Electr. Power Energy Syst., vol. 26, pp. 535–544, analysis and control, new energy and energy storages applications to power
2004. systems.
[27] M. Ishimaru, R. Yokoyama, O. Neto, and K. Y. Lee, “Allocation and
design of power system stabilizers for mitigating low-frequency oscil-
lations in the eastern interconnected power system in Japan,” Int. J.
Electr. Power Energy Syst., vol. 26, pp. 607–618, 2004.
[28] H. F. Wang and Swift, “Multiple stabilizer setting in multi-machine Kwang Y. Lee (F'01–LF’08) received the B.S.
power systems by the phase compensation method,” Int. J. Electr. degree in electrical engineering from Seoul National
Power Energy Syst., vol. 20, pp. 241–246, 1998. University, Seoul, Korea, in 1964, the M.S. degree
[29] W. Du, H. F. Wang, J. Cao, H. F. Li, and L. Xiao, “Application of in electrical engineering from North Dakota State
the phase compensation method for the design of a DC/AC converter- University, Fargo, ND, USA, in 1968, and the
based stabilizer to damp inter-area power oscillations,” IEEE Trans. Ph.D. degree in system science from Michigan State
Power Syst., vol. 27, no. 4, pp. 1302–1310, Nov. 2012. University, East Lansing, MI, USA, in 1971.
[30] H. Akagi and H. Sato, “Control and performance of a doubly-fed in- He has been on the faculties of Michigan State,
duction machine intended for a flywheel energy storage system,” IEEE Oregon State, Houston, the Pennsylvania State
Trans. Power Electron., vol. 17, no. 1, pp. 109–116, Jan. 2002. University, and Baylor University, Waco, TX, USA,
[31] J. B. Ekanayake, L. Holdsworth, and N. Jenkins, “Comparison of 5th where he is currently a Professor and Chair of the
order and 3rd order machine models for doubly fed induction generator Electrical and Computer Engineering and Director of the Power and Energy
(DFIG) wind turbines,” Electric Power Syst. Res., vol. 67, pp. 207–215, Systems Laboratory. His research interests are power systems control, oper-
2003. ation and planning, and intelligent systems applications to power plant and
[32] F. Wu, X. P. Zhang, P. Ju, and M. J. H. Sterling, “Decentralized non- power systems control.
linear control of wind turbine with doubly fed induction generator,” Dr. Lee has served as an editor of the IEEE TRANSACTIONS ON ENERGY
IEEE Trans. Power Syst., vol. 23, no. 3, pp. 613–621, Aug. 2008. CONVERSION and an associate editor of the IEEE TRANSACTIONS ON NEURAL
[33] F. M. Hughes, O. Anaya-Lara, N. Jenkins, and G. Strbac, “A power NETWORKS. He is currently serving as Chair of the IEEE Working Group on
system stabilizer for DFIG-based wind generation,” IEEE Trans. Modern Heuristic Optimization and Chair of the IFAC Technical Committee,
Power Syst., vol. 21, no. 3, pp. 763–772, Aug. 2006. TC 6.3—Power and Energy Systems.
[34] F. Wu, X. P. Zhang, K. Godfrey, and P. Ju, “Small signal stability anal-
ysis and optimal control of a wind turbine with doubly fed induction
generator,” IET Gener. Transm. Distrib., vol. 1, no. 5, pp. 751–760,
2007. Feng Wu received the B.Eng. and M.Sc. degrees from Hohai University, Nan-
[35] L. Shi, Z. Chen, H. F. Wang, and G. Tang, “Damping of power system jing, China, in 1998 and 2002, respectively, and the Ph.D. degree from the Uni-
low-frequency oscillations with FESS,” Autom. Electric Power Syst., versity of Birmingham, Birmingham, U.K., in 2008.
vol. 34, no. 8, pp. 29–33, 2010. He is currently a Professor with the College of Energy and Electrical Engi-
[36] X. F. Wang, Modern Power System Analysis. Beijing, China: Sci- neering, Hohai University, Nanjing, China. His research interest is power system
ence, 2003. modeling and control.