INTEGRATED PASSIVE/AcTIVE VIBRATION ABSORBER

FOR MULTISTORY BUILDINGS

By Gina J. Lee-Glauser; Goodarz Ahmadi/ and Lucas G. Horta3

Ass"':RACT: In this paper, a passive isolator, an active vibration absorber, and an integrated passive/active
(hybnd) control are studied for their effectiveness in reducing structural vibration under seismic excitations. For
the passive isolator, a laminated rubber bearing base isolator, which has been studied and used extensively by
researchers and ~e~smic. desi~n~rs, is studied. An active vibration absorber concept, which can provide guaranteed
c~osed-loop stability Wlth mmlmum knowledge of the controlled system, is used to reduce the passive isolator
Downloaded from ascelibrary.org by MISSOURI, UNIV OF/COLUMBIA on 06/03/13. Copyright ASCE. For personal use only; all rights reserved.

dlsplacement and to suppress vibration. A three-story building model is used for the numerical simulation. The
performance of an active vibration absorber and a hybrid vibration controller in reducing peak structural re-
sponses is compared with the passively isolated structural response under the NOOW component of the El Centro
1940 and N90W component of the Mexico City 1985 earthquake excitation records.

INTRODUCTION bration suppression in their applications. In this paper, the

To alleviate detrimental seismic hazards, passive and active model-independent AVA controller concept is used to design
vibration control schemes have been introduced for the pro- and evaluate the AVA's effectiveness in structural vibration
tection of buildings and lifelines. A historical review of pas- reduction under seismic excitations.
sive structural isolation devices by Kelly (1982, 1986) and The LRB base isolator has been shown by Kelly (1982,
studies by Su et al. (1989) and Fan et al. (1991) have shown 1986), Su et al. (1989), Fan et al. (1991), Derham (1986), and
that these devices have great potential to prevent earthquake Tajirian and Kelly (1987) to be highly effective in reducing
damages to buildings, nuclear power plants, and sensitive sub- structural vibration under seismic excitation. This vibration-
systems within structures. The most common passive isolation reduction capability, however, is associated with a certain
system is the laminated rubber bearing (LRB) base isolator. amount of base displacement. With a large passive element
An LRB isolator is quite flexible in the horizontal direction displacement, there are concerns about the safety and stability
and rather stiff in the vertical direction. It is manufactured by of bearings in addition to redistribution of loads among the
alternating layers of rubber and steel with the rubber vulca- bearings due to differential vertical displacements. Further-
nized to the steel plates. The horizontal flexibility of LRB more, there is a need to construct appropriate flexible connec-
isolator shifts the fundamental frequency of vibration away tions for the infrastructures such as plumbing, electrical, and
from the high-energy frequency range of earthquake ground communication conduits. Therefore, developing an active con-
motion. The bearing is designed to resist wind forces with little trol mechanism for reducing the base displacement to a man-
or no deformation. ageable level is highly desirable. Note also that addition of
Active vibration control methodologies for civil structures simple passive dampers does not solve the problem. While
have been explored extensively in recent years as reported by increase in damping of the isolator could reduce the peak base
Soong (1990), Meirovitch (1990), and Reinhorn and Manolis displacement, it is at the expense of an increase in the accel-
(1989). Chawla et al. (1994) used axial-force-rated actuators eration experienced by the structure. In addition, Su et al.
to suppress seismic vibration. For random disturbance cancel- ~1989) and Fan et al. (1991) have shown that passive base
lation of a multi-degree-of-freedom system, Nonami et al. lsolatlOn systems are ineffective for protection against earth-
(1994) designed a feedback and feedforward controller using quakes with considerable energy at low frequencies. In those
a model-dependent active vibration absorber. They designed a cases the use of active and/or hybrid systems may be recom-
feedback and feedforward controller to control the first two mended.
modes of the structure. Their numerical and experimental re- The effectiveness of an active vibration absorber and an
sults show significant vibration reduction under random dis- integrated passive/active control in reducing structural vibra-
turbance forces. Unlike Nonami's active dynamic vibration ab- tion under seismic excitations are analyzed here. The results
sorber, Lee-Glauser et al. (1995a) designed a model-independ- are compared with that of the passive LRB system and the
ent active vibration absorber (AVA) controller to evaluate the peak structural responses in the absence of control systems.
c~osed-Ioop s!a:bility and its effectiveness in vibration suppres- The NOOW component of EI Centro 1940 and the N90W com-
Slon for a flexlble space structure. Their experimental results ponent of Mexico City 1985 earthquakes are used to excite a
show that the AVA controller significantly reduces the random three-story building model in the numerical simulation. The
disturbance into the flexible space structure model. Both Non- peak acceleration responses for various controllers are evalu-
ami et al. (1994) and Lee-Glauser et al. (1995a) have experi- ated and the results are presented as response spectra curves.
mentally demonstrated the validity of AVA controllers for vi-
PROBLEM FORMULATION
'NRC Res. Assoc., NASA Langley Res. Ctr., MS 230, Hampton Va The governing equations of motion for a general multistory
23681. '
2Prof., Dept. of Mech. and Aeronautical Engrg., Clarkson Univ., Pots- shear frame structure with a passive vibration isolator are
dam, NY 13699-5725.
'Asst. Branch Head, Struct. Dyn. Branch, NASA Langley Res. Ctr.,
MX + Ci + Kx = -M(Xb + xg){l} (I)
MS 230, Hampton, VA. where {I} = a column vector whose elements are all unity.
Note. Associate Editor: Nicholas P. Jones. Discussion open until Sep- For a fixed-base structure (without a passive isolator), Xb =O.
tember 1. 1997..To extend the closing date one month. a written request
m~st be filed With ~e ASCE Manager of Journals. The manuscript for
In this study, a three-story building is the structural model.
thiS paper was submitted for review and possible publication on April 6 !he ~ass of .each floor and the floor mat on top of the base
1995. This paper is part of the Jour,",l oj Structural Engineering. Vol: lsolatlOn deVlce are assumed to be equal. The column stiff-
123, No.4, April. 1997. ©ASCE, ISSN 0733-9445/97/0004-0499-0504/ nesses and dampings are also assumed to be identical. For this
$4.00 + $.50 per page. Paper No. 10515. structural model, the damping matrix is proportional to the
JOURNAL OF STRUCTURAL ENGINEERING / APRIL 1997/499

J. Struct. Eng. 1997.123:499-504.

 AVA Controller

3 Story
Building
m
Downloaded from ascelibrary.org by MISSOURI, UNIV OF/COLUMBIA on 06/03/13. Copyright ASCE. For personal use only; all rights reserved.

~ XXbbt~mb ~Passive Controller

FIG. 1. Schematic Diagram of Three-Story Building with Passive and Active Controllers

stiffness matrix. A fundamental natural frequency of 3.33 Hz No Control

and a damping ratio of 0.02 are assumed for the structure. For
peak acceleration response analyses, a range of fundamental
natural frequencies of the building are used.
The equations of motion for passive and active vibration-
control systems are summarized in the following subsections. Passive Control

Passive Vibration Isolator

For a laminated rubber bearing base isolator, the equation
of motion of the passive system is
Active Control
+ W~Xb +
..
Xb +
2'bWbXb
Olb
-
etXt
mb
ktXt
= -x
.
g (2)

The natural frequency of the laminated rubber bearing Wb and 2.0 f-------+---------l~----__1
its effective damping ratio ~b are defined as Hybrid Control

0.0 ~...,I'V-~----------
(3a,b)
-2.0 L..- ~ ~ _J
The parameter Olb = ratio of base mass to the total mass of the
structure, i.e. 0.0 20.0 40.0 60.0
TIme (sec)
(4a,b)
FIG. 2. Sample Time Histories of Third Floor Absolute Accel-
erations for Various Vibration Controllers and Unprotected
In this study, a commonly suggested natural frequency of Case for EI Centro 1940 Earthquake
0.5 Hz is used for the LRB base isolation. The effective damp-
ing ratio of the rubber varies between 0.05 at high strain and Mex., + Dexe + KcXe = Beue, Ye = u.ex., (6)
about 0.3 at low strain according to Derham (1986) and Ta-
jirian and Kelly (1987). Since for a strong earthquake a high The controller mass, damping, and stiffness matrices are, in
level of strain in the bearing is expected, a typical effective general, symmetric and positive definite, so that the controller
damping ratio of 0.08 is used in the analysis. is asymptotically stable.
For the interconnected controller and structure with collo-
Active Vibration Absorber cated sensors/actuators, the control equation is revised to in-
Consider an n-mode structural dynamics model with only clude a direct acceleration feedback designed to form a model-
the acceleration measurement of the system masses. The gov- independent controller that guarantees the closed-loop stability
erning equations are written as regardless of any perturbations, i.e.

MX + Dx + Kx =B u, Y = Hax (5) U = Ye - GaY = Haex., - GaHax (7)

The second-order AVA control is governed by where G a = a gain matrix defined as

500 / JOURNAL OF STRUCTURAL ENGINEERING / APRIL 1997

J. Struct. Eng. 1997.123:499-504.

 (a)
(b)

~ ~ 3rd FL

, i 3rd FL

~
.......
..J L. 2nd FL

II
I
Downloaded from ascelibrary.org by MISSOURI, UNIV OF/COLUMBIA on 06/03/13. Copyright ASCE. For personal use only; all rights reserved.

2nd FL

0.16
L 11t FL 0.16

Ground tit FL
0.00 0.00 . I - -.....~IlloolI,o----_-_---l
0.0 5.0 10.0 15.0 0.0 6.0 10.0 16.0
Frequency (Hz) Frequency (Hz)

(e) (d)

.... 3rd FL 3rd FL

, .. ,
i
~
..."

2nd FL
~ 2nd FL

I 0.Q.4
.... 11t FL I 0.Q.4 ~ tit FL

0.00
0.0
.. 6.0 10.0
Ba..
16.0
0.00
0.0
A...
5.0 10.0
Boee

15.0
Frequency (Hz) Frequency (Hz)

FIG. 3. Fourier Decompositions of Acceleration Responses at Various Floors for Different Vibration-Control Systems for EI Centro
1940 Earthquake: (a) No Control; (b) Active Control; (c) Passive Control; (d) Hybrid Control

G. = H.eM;IBe (8) which is positive definite as long as M and Me are positive

definite. With the collocated sensors/actuators, it is assumed
Since the sensors/actuators are collocated that Be = H. = 1. This AVA design has been used and verified
B = H; and U.e = B~ (9a,b) experimentally and numerically by Lee-Glauser et al. (1995a).
In this study, the second-order controller is assumed to be
Let, Be be defined as attached to the third floor as shown in Fig. 1. The AVA control
law is
(lOa,b)

then the closed-loop mass matrix becomes (13)

M, = M + H.B
T-T -
5 MeBeH.
T-T
-H.BeMe]
The Xc is computed from
[ -MeBeH. Me
(14)
which is symmetric. To assure positive definiteness
x~M,x, >0 (11)
where me, de, and ke = controller parameters. The optimum
AVA controller parameters are obtained by using the frequency
must be true for all real closed-loop displacement vectors x, matching method described by Lee-Glauser et al. (1995a).
except the null vector. Substituting M, into (11) yields Here, the active vibration absorber is designed to enhance the
passive isolator. For seismic application, the actuator mass is
x~M,x, = xTMx + (M;IBeu.x - xe)Mc(M;IBeH.x - Xc) (12) approximately 1-2% of the structural mass. Therefore, the im-
JOURNAL OF STRUCTURAL ENGINEERING 1 APRIL 1997/501

J. Struct. Eng. 1997.123:499-504.

 portance of the actuator dynamics as reported by Inman (1990) for reference. This figure clearly shows that the LRB and hy-
is not included in the numerical simulation. brid systems are highly effective in reducing the peak accel-
eration responses, with the hybrid system reduced the most
NUMERICAL STUDIES throughout the frequency range examined. The active vibration
absorber system is not as effective as the passive isolator and
In this section, performance of passive, active, and hybrid the hybrid system. Once again, this is due to the power con-
vibration control systems in protecting the structure subjected straint on the active controller.
to the NOOW component of EI Centro 1940 and N90W com- Fig. 5 depicts the peak base displacements of the passive
ponent of Mexico City 1985 earthquake records are studied. and hybrid systems for the fundamental mode frequencies of
The normal modes expansion technique is used to analyze the 1-10 Hz. It is observed that the hybrid system leads to base
dynamic response of the structure. displacement that is 20% less than that for the passive system.
This base movement reduction can be a significant factor for
EI Centro 1940 Earthquake the safety and stability of bearings to protect the lifeline in-
Downloaded from ascelibrary.org by MISSOURI, UNIV OF/COLUMBIA on 06/03/13. Copyright ASCE. For personal use only; all rights reserved.

frastructure (such as plumbing and electrical and communi-

The EI Centro 1940 earthquake that has the features of cation conduits) connections at the building foundation. The
many common earthquakes is used in this section as the results show that the passive system alone can provide protec-
ground excitation. Fig. 2 shows sample time histories of the tion for the structure against earthquake. However, the accom-
absolute third floor lateral accelerations for each of the vibra-
tion controllers and the uncontrolled structure with a natural
frequency of 3.3 Hz. It is observed that all three controllers
are highly effective in reducing the peak absolute acceleration. -- no control
However, the vibration reduction of about 75-85% is noticed --- passive
with the passive and hybrid systems. The resulting response ...... - - - - hybrid
for the LRB isolator is also in good agreement with that of .....,
C' - .-. active
c:
Fan et al. (1991). Although, the AVA controller is tuned to the 0
:;::;
passive isolator using only small input force, Fig. 2 shows that E
lP
it still is an effective vibration suppression alone. Gi
u
Fourier decompositions of the acceleration responses at each ~ 1.0
floor for various vibration-control systems are shown in Fig.
3, which depicts the frequency content of the acceleration at
..a
0

different floors for a structure with and without vibration-con- u:

trol devices. The Fourier spectrum of the ground shown in 'E
I")

Fig. 3(a) is that of the accelerogram of the EI Centro 1940 ~

C
lP
earthquake. Fig. 3(a) shows the sharpening of the spectrum Do
near the natural frequency of 3.3 Hz for the building model
used. That is, the peak spectrum amplitude at 3.3 Hz increases
sharply as the acceleration propagates to the higher floors. It
is observed that the uncontrolled three-story building model 0.0 L..-_L..-_L..-_L..----''-----''-----''--_'--_L..----'

behaves like a shear beam structure. This can be seen by the 1 2 3 4 5 6 7 8 9 10

increasing absolute acceleration amplitude as the floor level
increases. The Fourier spectra of the structure with the AVA Building Fundamental Frequency (HZ)
controller is shown in Fig. 3(b). The spectra contains approx- FIG. 4. Peak Third Floor Acceleration Responses for Various
imately the same frequency components as those of the ground Vibration Controllers for EI Centro 1940 Earthquake
acceleration but with reduced amplitudes. That is, the AVA
controller does not filter out the broadband excitation, but does 0.16 r-----.----.---,---.--.,.--.-----.---.---,
significantly reduce the peak resonance observed in the un-
controlled case. Fig. 3(c) shows the floor acceleration Fourier - - passive
spectra for a structure with an LRB isolation system. The rigid - - - hybrid
body motion of the superstructure is observed to be identical
responses at different floors. The dominant frequency observed ~~--------j
in this figure is about 0.5 Hz, which corresponds to the LRB lj 0.14
system natural frequency. The base isolation filters out the E
high-frequency energy content of the earthquake acceleration, 8
but amplifies the energy at its natural frequency. The Fourier oS!
C.
III
spectra of the hybrid controller are shown in Fig. 3(d). While i5
the general trend of the spectra is similar to that of Fig. 3(c),
the sharp peaks observed in the passive controller responses Jl
~ 0.12
have been significantly reduced by use of the hybrid AVA con- ~

troller in conjunction with the LRB isolator. Here, the AVA 2

controller is tuned at the fundamental frequency of the passive
Do ...- ~-----------~---------
I
system. Figs. 3(c) and 3(d) also show that the Fourier spectra
amplitudes of different floors are roughly the same. This im-
plies that when passive and/or hybrid control systems are used,
the superstructure vibrates more like a rigid body and does not 0.1 0 L----'-_--'--_-'-_.L----l._~_-'-_-'--~

amplify the ground excitation. 1 2 3 4 5 6 7 8 9 10

The peak third floor absolute acceleration responses shown Building Fundamental Frequency (Hz)
in Fig. 4 are calculated for the building fundamental mode
frequency of 1-10 Hz for various vibration-control systems. FIG. 5. Peak Base Displacement Responses for LRB and Hy-
The uncontrolled response spectra is also shown in this figure brid Systems for EI Centro 1940 Earthquake

5021 JOURNAL OF STRUCTURAL ENGINEERING 1 APRIL 1997

J. Struct. Eng. 1997.123:499-504.

 1.5 hybrid controllers actually amplify the structural vibration by
an order of magnitude. However, the active controller reduces
•• - . active the peak absolute acceleration by about 40%. Frequency con-
1.0 - - hybrid tent at various floors for different vibration-control systems

-
.......
gI

Gl
~
0.5
(similar to that in Fig. 3) are not shown here but are reported
in Lee-Glauser et aI. (l995b).
Fig. 8 shows the peak third floor acceleration responses for
~
the building fundamental mode in the frequency range of
.....;, 1-10 Hz for various vibration-control devices. This figure also
D- 0.0 shows that the passive and hybrid controllers are not effective
E
... for this earthquake. The active AVA controller, however, pro-
.£ vides effective vibration suppression for the Mexico City 1985
e.....c::
Downloaded from ascelibrary.org by MISSOURI, UNIV OF/COLUMBIA on 06/03/13. Copyright ASCE. For personal use only; all rights reserved.

-0.5 earthquake.
u
0 Fig. 9 displays the peak base displacement of the LRB and
hybrid systems for various fundamental mode frequencies. It
-1.0 is observed that the base displacements for the LRB system is
too large for any structure to be standing after the earthquake.

1.2 r----.-----.---.----,---r--..-----.---,-----,
-1.5
0.0 20.0 40.0 60.0 - - no control
- - - passive
Time (sec) - - - - hybrid
- - - . active
FIG. 6. AVA Controller Input Force for EI Centro 1940 Earth- :§0.9
quake c::
o
~Gl
t-, ......
...... _- ------------------
No Control '8
.li! 0.6
8
ti:

Passive Control

" " ,-----

Active Control 0.0
... _--- -
1 2 3 4 5 6 7 8 9 10
...
g Building Fundamental Frequency (Hz)
ti:
FIG. 8. Peak Third Floor Acceleration Responses for Various
~ 1.0 Vibration Controllers for Mexico City 1985 Earthquake
Hybrid Control

0.0 f--~~iV'tMrtM"..".,...------- 1.0 ,----.-----.---..---,---r--..-----.---,----,

v
-1.0 L-_~_-,-- --,--_~_ .......... .....J

o 50 100 150 200 0.8

Time (sec)
E
.......
1:Gl
FIG. 7. Sample Time Histories of Third Floor Absolute Accel-
eratlona for Various Vibration Controllers and Unprotected ~ 0.6
o
Case for Mexico City 1985 Earthquake c
Q.
~
---------------------
panying large base displacement requires special connection
devices for the lifelines, whereas the hybrid controller can pro- :l 0.4
c
lJ)
tect the structure even more effectively while reducing the ~
peak base displacement. - - passive
Q.
il - - - hybrid
The AVA controller input force for active and hybrid sys- 0.2
tems are shown in Fig. 6, in which the hybrid system input
force requirement is only 1/4 of that for the active system.

Mexico City 1985 Earthquake 0.0 l...-----L_--'-_-'-_-'----"_---'-_-'-_"""'-----l

1 2 3 4 5 6 7 8 9 10
In this section the structural responses to the Mexico City
1985 earthquake is studied. Fig. 7 shows sample time histories Building Fundamental Frequency (Hz)
of the absolute accelerations at the third floor for various con- FIG. 9. Peak Base Displacement Responses for LRB and Hy-
trollers. For this earthquake, it is observed that the passive and brid Systems for Mexico City 1985 Earthquake

JOURNAL OF STRUCTURAL ENGINEERING / APRIL 1997/503

J. Struct. Eng. 1997.123:499-504.

 Use of the hybrid system reduces the peak displacement by 94-1748-CP; Proc., AlAAIASMEIASCEIAHS 35th Struct., Struct. Dyn.
half; however, it is still too large for any structure to withstand and Mat. Conf., Adaptive Struct. Forum, 143-150.
Derham, C. J. (1986). "Nonlinear natural rubber bearings for seismic
that much of a shift. isolation." Proc., ATC-I7 Seminar on Base Isolation and Passive En-
ergy Dissipation.
CONCLUSIONS Fan, F. G., Ahmadi, G., Mostaghel, N., and Tadjbakhsh, 1. G. (1991).
"Performance analysis of seismic base isolation systems for a multi-
The performance of a passive (LRB) isolator, active vibra- story building." Soil Dyn. and Earthquake Engrg., 10, 152-171.
tion absorber, and integrated passive and active control (hy- Inman, D. J. (1990). "ControVstructure interaction: effects of actuator
brid) systems in reducing structural vibration under seismic dynamics." AlAA-90-I223-CP; Proc., AIAA Dyn. Specialists Conf.,
excitations was evaluated. The EI Centro 1940 and Mexico 311-321.
Kelly, J. M. (1982). "Aseismic base isolation." Shock and Vibration Dig.,
City 1985 earthquake accelerograms were used as base exci- 14,17-25.
tations. Based on the presented results, the following conclu- Kelly, J. M. (1986). "Aseismic base isolation: review and bibliography."
sions may be drawn: Soil Dyn. Earthquake Engrg., 5, 202-216.
Downloaded from ascelibrary.org by MISSOURI, UNIV OF/COLUMBIA on 06/03/13. Copyright ASCE. For personal use only; all rights reserved.

Lee-Glauser, G., Juang, J.-N., and Sulla, J. I. (1995a). "Optimal active

1. The passive (LRB) isolator, the integrated passive/active vibration absorber: design and experimental results." J. Vibration and
Acoustics, 117 (April), 165-171.
(hybrid) system, and the active AVA controller could be Lee-Glauser, G., Ahmadi, G., and Horta, G. L. (1995b). "Integrated pas-
designed to be highly effective in suppressing the struc- sive/active vibration absorber for multi-story buildings." NASA TM-
tural vibrations. The degree of effectiveness of the vi- 109178.
bration-suppression systems depends on the earthquake Meirovitch, I. (1990). Dynamics and control of structure. Wiley-Intersci-
(i.e., its expected frequency content). ence, New York, N.Y.
2. For a broadband earthquake (such as the EI Centro 1940 Nonami, K., Nishimura, H., and Cui, W. (1994). "Disturbance cancel-
lation control for vibration of multi-degree-of-freedom systems."
earthquake), the passive and hybrid controllers are highly JSME Int. J., 37(0, 86-93.
effective in reducing the transmitted acceleration. Reinhom, A. M., and Manolis, G. D. (1989). "Recent advances in struc-
3. For a narrow-band earthquake (such as the Mexico City tural control." Shock and Vibration Dig., 21(1), 3-8.
1985 earthquake), the active AVA controller is the most Soong, T. T. (1990). Active structural control: theory and practice. Long-
effective in reducing the structural vibration. For earth- man's, London, England.
quakes with considerable energy at low frequencies, the Su, L., Ahmadi, G., and Tadjbakhsh, I. G. (1989). "A comparative study
of performances of various base isolation systems." Earthquake Engrg.
passive and hybrid controllers have an adverse effect and and Struct. Dyn., 18, 11-32.
amplify the structural vibration. Tajirian, F. F., and Kelly, J. M. (1987). "Seismic and shock isolation
4. The structure with a passive and/or hybrid control sys- system for modular power plants." Proc., 1987 Pressure Vessels and
tems vibrates roughly in its rigid body mode. The un- Piping Conf.
controlled structure and the one with an active AVA sys-
tems behaves as a shear beam and amplifies the APPENDIX II. NOTATION
transmitted acceleration along its floor. The following symbols are used in this paper:
5. The tuned hybrid system reduces the sharp resonant peak
at the fundamental frequency of the passive isolator. B :: n X p actuator force distribution matrix;
6. The hybrid system reduces the base isolation displace- Be :: ne X m controller influence matrix;
ment from the passive isolator displacement. C :: damping matrix;
Cb :: rubber bearing damping;
De :: controller damping matrix;
The results of this study show that knowledge of the fre- H m X n acceleration influence matrix;
::
Q
quency content of the expected earthquake is necessary for the Hoc :: p X no controller acceleration influence matrix;
proper selection of a vibration-control device. For earthquakes K :: stiffness matrix;
in cities on bedrock, the passive and/or the integrated passive! K., :: controller stiffness matrix;
active controller is most effective in protecting structures. For kb :: rubber bearing horizontal stiffness;
cities with soft soil, the AVA active vibration controller is most M :: mass matrix;
appropriate. Me :: controller mass matrix;
mb = effective passive element mass;
ACKNOWLEDGMENT m/ :: ith floor mass;
m, :: total mass of the structure;
This work is supported by the National Aeronautics and Space Ad- u :: p X I control vector;
ministration (NASA), National Research Council, at Langley Research
Ue :: m X I input force vector;
Center, Grant No. 44.50.38.03.
x :: floor displacement relative to the passive element;
Xb :: passive element displacement relative to the ground;
APPENDIX I. REFERENCES Xc :: Xc X I controller displacement vector;
Chawla, V., Sener, M., Utku, S., and Wada, B. K. (1994). "Seismic vi- y = m X I measurement vector; and
bration suppression by the axial-force-rated actuators." AIAA Paper y c :: p X I controller measurement vector.

504/ JOURNAL OF STRUCTURAL ENGINEERING / APRIL 1997

J. Struct. Eng. 1997.123:499-504.