Download as pdf or txt
Download as pdf or txt
You are on page 1of 12

EARTHQUAKE ENGINEERING A N D STRUCTURAL DYNAMICS, VOL.

15,379-390 (1987)

DYNAMICS OF RESILIENT-FRICTION BASE ISOLATOR (R-FBI)

N. MOSTAGHEL* AND M. KHODAVERDIAN~


Department of Civil Engineering, Uniuersity of Utah, Salt Lake City, Utah 841 12, U.S.A.

SUMMARY
The Resilient-Friction Base Isolator (R-FBI) is composed of a set of flat rings which can slide on each other with a central
rubber core and/or peripheral rubber cores. In this base isolator the interfacial friction force acts in parallel with the elastic
force in the rubber. It combines the beneficial effect of friction damping with that of the resiliency of rubber. The rubber
cores distribute the sliding displacement and velocity along the height of the R-FBI. They d o not carry any vertical loads
and are not vulcanized to the sliding rings.
The system's analytical model and the computer experimental results for both horizontal and vertical components of
recorded ground motions and various levels of friction and damping are presented. These results clearly demonstrate the R-
FBI's potential as an effective aseismic base isolator.

INTRODUCTION
The transmission of ground motion to structures can be effectively controlled through isolation of the
structure at its base. This involves a specially designed foundation system that limits the intensity of the ground
motion transmitted to the superstructures. In contrast, the traditional method of providing for the lateral load-
carrying capacity allows the entire ground motion to be transmitted to the superstructures and tries to provide
for the absorption of energy through inelastic actions which inevitably give rise to damage to structural and
non-structural elements. Base isolation is preventive in concept and strives to control the level of ground
motions to be transmitted to the superstructure. It holds the promise of protecting the structures and their
internal equipment from damaging ground motions and provides an alternative to retrofitting of existing
structures.
R-FBI is a base isolation device that can be incorporated into structures' foundations. As shown in Figures 1
and 2, the R-FBI is composed of layers of sliding elements with rubber cores. The performance of sliding
elements (with no elastic core) as a base isolator has already been studied and the results are very
en~ouraging.'-~ Of course, the concept of limiting the intensity of motions transmitted to the structures via
certain isolation schemes is not new. Besides the classical spring-mass system: which is developed based on
harmonic excitations, there are many suggestions to isolate structures from damaging effects of earthquakes.
The flexible first storey and the soft storey concept,* due to consequential instability and P-A
effects, are not practical schemes.' Also, the use of ball bearings and specially shaped rollers'O~" under the
structures has been suggested.
Considerable work has been done to show the effectiveness of steel plate laminated rubber bearings with and
without a lead as a base isolator system. To limit the shear distortion in the elastomer and to bestow
larger displacement capability, the use of a friction plate in conjunction with steel plate laminated rubber
bearings has also been considered.' 7-20 However, if the predominant frequency of excitation is low (as in the
case of a soft site), the above systems can act as amplifiers and impose large displacement demands on the base.
The effectiveness of sliding type isolators in limiting the intensity of ground motion transmitted to the
superstructure has been demonstrated both experimentally2',22 and analyti~ally.'-~~ 23-27

* Professor.
Graduate Student.

0098-8 847/87/030379-12$06.O0 Received 10 March 1986


0 1987 by John Wiley & Sons, Ltd Revised 2 June 1986
380 N. MOSTAGHEL AND M. KHODAVERDIAN

Rubber Cores
B o t t a n Connecting
Plate

Figure 1. The R-FBI system (Mostaghel 1984)

1. Teflon coated stainless steel flat ring


2. Central rubber core
3. Peripheral rubber core
4. Top flat ring (This is welded to the top connecting plate)
5. Top connecting plate

i 6. Bottom flat ring (This is welded to the bottom connecting


I. Bottom connecting plate
8. Top cover plate
9. Bottom cover plate
10. Bottom plate bolt hole
11. Top plate bolt hole
12. Rubber cover

I
I I
SECTION A-A

-
F

I
A
1 A

I
SECTION B-B DISPLACED CONFIGURATION

Figure 2. The anatomy of an R-FBI system (Mostaghel 1984)

As shown in Figures 1 and 2, the R-FBI is composed of a set of flat rings which can slide on each other with a
central rubber core and/or peripheral rubber cores. Also, there is a very flexible rubber cover which protects the
sliding rings from corrosion and dust. To reduce the friction, the sliding rings are teflon coated. The rubber
cores help to distribute the lateral displacement and velocity along the height of the isolator. The R-FBI is
RESILIENT-FRICTION BASE ISOLATOR 38 1

characterized by the coefficient of friction of the sliding elements and the total lateral stiffness of the rubber
cores. The damping capability of the rubber is small, and the friction damping is the main energy dissipator of
the R-FBI. As the rubber cores are only fitted but not bonded to the sliding rings, the construction of an R-FBI
should be a relatively simple task. Of course, as shown in Figure 3, the cross section of the sliding rings or the
rubber core does not have to be circular.
It should be noted that the R-FBI is designed such that all the vertical load is carried entirely by the sliding
elements, and no vertical load is carried by the rubber cores. Therefore, the R-FBI is very rigid in the vertical
direction. Its rigidity will be close to that of the metal used to fabricate the sliding elements. Hence, the R-FBI
does not offer isolation against vertical ground motion. The response of structures to vertical ground motion
may cause non-uniform distribution of vertical load on the isolators leading to variations in frictional
resistance affecting the horizontal sliding displacements. There are two factors which tend to limit this
potential effect. First, the sliding of an R-FBI is restrained by the stiffness of the rubber cores. Second, the
variation of the vertical load is of much higher frequencies as compared to the horizontal load. This should lead
to a high degree of decoupling between horizontal and vertical responses. Since base isolated structures are
essentially subjected to much smaller lateral loads, the overturning moments are smaller. Previous studies” on
steel plate laminated rubber bearings with friction plates have shown that overturning moment and rocking do
not appreciably affect the overall seismic response of the base-isolated structures.
In order to show the isolation characteristics of the R-FBI, the response of rigid structures supported on an
R-FBI is considered. This is to avoid any interaction of the superstructure and the isolation system. To be
effective, the isolator period should be much larger than the fundamental period of the structure it is
supporting, thereby causing insignificant interaction. The interaction effects are under consideration and the
results will be reported later on. In this paper the mathematical model of a shear building supported on an R-
FBI system is formulated. The model is then reduced to that of a rigid structure founded on an R-FBI system.
The latter model is subjected to both the N-S component of the 1940El Centro earthquake and to the N-S and
vertical components of the 1940 El Centro earthquake simultaneously. Response spectra for the sliding
displacement, velocity, and the total acceleration for various coefficients of friction and levels of damping are
presented.
Although for design purposes, one should consider more examples of earthquake records, the above example
clearly demonstrates the potential of the R-FBI technology in controlling the seismic response of structures.

GOVERNING EQUATIONS
Figure 4represents a multi-storey shear structure supported on an R-FBI system. The system is subjected to the
horizontal and vertical components of ground motions x ( f )and jj(f). As long as the intensity of ground motion
is small, the friction force prevents any lateral motion. As soon as the ground motion exceeds certain levels, the
lateral load exceeds the friction force, the base starts to slide, and the rubber core deforms and builds resistance.
Applications of Newton’s law yield

s + 25nS + R2s = -x - p ( g + yIRS - r,I


N
= t
aiui

mu+cu+ku = -((S’+x)mr (2)

Figure 3. Various cross sectional shapes


382 N . MOSTAGHEL A N D M. KHODAVERDIAN

$ Y

Figure 4. A structure supported on an R-FBI system

where
a. =
K C
’ M + Emi’ Q* = M + Zm,’ 2tl-I = M + Cm,
mi

and
x = horizontal ground acceleration,
j; = vertical ground acceleration,
K = horizontal stiffness of the elastic core,
M = foundation mass over the isolator
C = damping of the elastic core
( = damping ratio of the base isolator.
p = coefficient of friction of the isolator,
g = gravitational acceleration
s = sliding displacement of the isolator,
u = relative displacement vector,
m = structure’s mass matrix,
c = structure’s damping matrix,
k = structure’s stiffness matrix,
r = a unit vector,
N = number of floors, and dot (.) represents differentiation with respect to time.
Since, in this paper, the dynamic behaviour of the R-FBI itself is of principal interest, in order to eliminate
any masking effects due to interactions, the structure is assumed to be rigid. The response of flexible structures
on an R-FBI is currently being investigated, and the results will be reported later on. The rigid structure
assumption reduces equations (1) and (2) to the following form:

S+25RS+R2S = - X - p ( g + j ) (4)
RESILIENT-FRICTION BASE ISOLATOR 383

In the digitized interval ti, t i + ] the horizontal and vertical components of ground acceleration can be
represented by

X(ti + T) = X(ti) + At
+1)
(5)

j;(ti + 7) = j;(ti)+ AY(ti


At
~
+ 1)
(6)
where At = t i + l -ti and 0 6 T 6 At. Substitution of equations ( 5 ) and (6) into equation (4) yields, for
ti < t 6 t i + 1
T
S’+21;RS+R2s = P i + Q i + ]- (7)
At
where
X(tJ + p [ g + jqt,)]

Solving equation (7) and incorporating the initial conditions yield

where z ( t i + , ) is the complementary solution of equation (7), and it is given by

, [
z(ti+ ) = e-cQAt - 1 - 2 p SinRdAt + m25
COSRdAt
1
Qi+l

+ e-cmt

where !& = RJ(1- 1;’).


It can be shown that

%+A = -zcRz(t,+,, -R2Z(ti+,) (15)

Equation (7) represents the behaviour of the system in sliding phases. In non-sliding phases s(ti) = s(ti) = 0.
384 N. MOSTAGHEL AND M. KHODAVERDIAN

START OF NON-SLIDING AND SLIDING PHASES


From equation (l), for rigid structures it can be shown that

In a non-sliding phase S = s’ = 0, therefore,


p(g+jj)-IX+CPs( > 0
At the start of a sliding phase inequality (17) changes to
p(g+ji)-IX+R2sI =o
In a sliding phase
S#O
p(g+jj)-IX+R2s.l <0 (19)
In the above relations it is assumed that there is no separation, i.e. Ijl < g. By marching in time, the values of s
and S are computed from equations (10) and (11) and checks of expressions (17) and (19) show whether the
system is in a non-sliding or sliding phase.

DISCUSSION OF RESULTS
To evaluate numerical results, a double precision Fortran routine was developed. For the purpose of this study
sliding velocities less than 0.001 mm/sec are assumed to be zero. This implies that even if the system was sliding
in one direction for all the 50 sec duration of a given ground motion, the error in sliding displacement would be

12

11

10

I
I I I I
1.5 2.5 3.5 4.5
PERIOD ( S C )

Figure 5. R-FBI displacement spectrum. El Centro 1940 NS & VT components


RESILIENT-FRICTION BASE ISOLATOR 385

of the order of less than 0 5 mm. The response of the system is very sensitive to the starting times of sliding and
non-sliding phases, implying that the digitized time interval Ar should be very small. After many trials, it was
decided that a time interval of At = OW02 sec would yield accurate results for all cases considered. The N-S
and the vertical components of the El Centro 1940 earthquake are considered simultaneously. Response
spectra for displacement, velocity and total acceleration for a damping of ( = 5 per cent in rubber, and five
values of friction coefficients p = OOOO,p = 0.04, p = 006, p = 0-08 and p = 0.10 for system periods of 1.5 to
5 sec at intervals of 0.10 sec are computed and presented in Figures 5 to 7. The range of friction utilized covers
the expected range of friction for teflon under the operating conditions of bearing pressure and velocity for the
R-FBI. To reduce interaction with the superstructure, 1.5 sec has been considered to be the lowest R-FBI
period in this study.
Figure 5 clearly shows the effect of friction in limiting the displacement response. The fact that the spectral
displacement response does not vary much with period shows that R-FBI limits the effects of frequency
content of the excitation on the spectral displacement response. Similar observations hold for the velocity
spectrum shown in Figure 6. The total acceleration response spectrum is presented in Figure 7. As seen from
this figure, the level of response depends on the value of the friction coefficient. As expected, larger coefficients
of friction yield larger responses. For design purposes the level of friction should be chosen to keep the
maximum displacement response (see Figure 5 ) less than a specified limit. It is of interest that for periods larger
than 2.5 sec the spectral velocity is of the order of 12 in/sec for the excitations and the range of friction
considered. This is a very high sliding velocity. If there were only one sliding surface, such a velocity would
cause the development of much larger friction for teflon than the range 0*04-0-10considered in this study.
Fortunately in the design of an R-FBI, one can incorporate as many sliding surfaces as necessary to reduce the
sliding velocity (and consequently the friction coefficient) to a range compatible with the desired response.
Figures 5-7 were developed based on the simultaneous application of the horizontal (N-S) and the vertical
components of the 1940 El Centro earthquake. To see the contribution of the inclusion of the vertical
excitation to the horizontal response, spectral responses for the horizontal component (N-S) of the El Centro
386 N . MOSTAGHEL A N D M. KHODAVERDIAN

0.2
0.19
0.18
0.17
0.16
0.15
h
0 0.14
v
0.13
0.12
0.11
0.1
cl
hl
uu 0.09

4 0.08
0.07
0.06
0.05
0.04
0.03
0.02 1
1.5 2.5 3.5 4.5

PERIOD (SEC)

Figure 7. R-FBI acceleration spectrum El Centro 1940 N S & VT components

1.01
1.009
1.008
1.007
1.006
1.005
2 1.004
2 1.003
o: 1.002
I-
1.001
z;E 1
0.999
2 0.998
0.997
0.996
0.995
0.994
0.993
0.992
0.991
0.99
1.5 2.5 3.5 4.5

PERIOD (SEC)

Figure 8. R-FBI displacement ratio spectrum. El Centro 1940 (NS & VT)/NS
RESILIENT-FRICTION BASE ISOLATOR 387

1940 earthquake are also calculated. Figures 8-10 represent the ratios of the spectral response quantities
involving both horizontal and vertical excitation to their corresponding values involving only horizontal
excitation. These figures demonstrate that for the case of the El Centro 1940 earthquake, and the range of

1.5 2.5 3.5 4.5


PERIOD (SEC)
Figure 9. R-FBI velocity ratio spectrum. El Centro 1940 (NS & VT)/NS

1.12

1.1

1.08

1.06

1.04

1.02

0.98

0.96

0.94

0.92

0.9

0.88
1.5 2.5 3.5 4.5
PERIOD ( X C )
Figure 10. R-FBI acceleration ratio spectrum. El Centro 1940 (NS & VT)/NS
388 N. MOSTAGHEL AND M. KHODAVERDIAN

parameters considered, the vertical excitation’s contributions to the horizontal response quantities are in
general of the order of less than 1 per cent.
To study the effect of the damping of the rubber core on the R-FBI response, for a friction coefficient of 0.06
and simultaneous excitations of both the horizontal (N-S) and the vertical components of the El Centro 1940
earthquake, the spectral responses for equivalent viscous damping ratios of 0~00,0~02,005,0~07 and 0.10 in
rubber are calculated and presented in Figures 11-13. From these figures, it may be concluded that for the
range of parameters considered, damping in the rubber core does not have a noticeable effect on velocity and
acceleration spectral amplitudes, and its effect on the displacement spectral amplitude is small. Considering
these observations together with Figures 5-7, which display the effect of friction on spectral amplitudes, one
infers that the main energy absorption mechanism in an R-FBI is friction.
Although the R-FBI can be designed to remain stable for any imposed displacement, in practice, it is prudent
to build a block adjacent to each R-FBI unit. This block, which is almost as tall as the R-FBI, will act as a fail-
safe mechanism. If, for any reason, the R-FBI system fails, the superstructure will sit on these blocks.

CONCLUSIONS
The response spectra presented in Figures 5-7 demonstrate the R-FBI’s potential in both controlling the level
of forces transmitted to the superstructures and limiting the maximum sliding displacements to levels which
can easily be considered and provided for in design. The interfacial friction in an R-FBI acts both as energy
absorber and as structural fuse. It may be chosen such that it would prevent any sliding under low level
excitations such as wind or small earthquakes.
The simplicity of the concept, the ease of construction, the effectiveness and the cost should lead to quick
acceptance of the R-FBI technology in the design community and will provide a viable alternative to the
traditional aseismic design practice.

h
Y

a
v1
i;
RESILIENT-FRICTION BASE ISOLATOR 389

20

19

18

17

16

15

14

13

12

11

10
1.5 2.5 3.5 4.5
PERIOD (SEC)

Figure 12. R-FBI velocity spectrum. El Centro 1940 NS & VT components

0.15

0.14

0.13

0.12

0.11

0.1

0*09

0.08

0.07

0.06

0.05
1.5 2.5 3.5 4.5
PERIOD (SEC)
Figure 13. R-FBI acceleration spectrum. El Centro 1940 NS & VT components
390 N. MOSTAGHEL AND M. KHODAVERDIAN

ACKNOWLEDGEMENT
The support of the U.S. National Science Foundation under grant number CEE-8414504 is gratefully
acknowledged.

REFERENCES
1. N. Mostaghel, M. Hejazi and J. Tanbakuchi, ‘Responseof sliding structures to harmonic support motion’, Earthquake eng. struct. dyn.
11, 355-366 (1983).
2. N. Mostaghel and J. Tanbakuchi, ‘Response of sliding structures to earthquake support motion’, Earthquake eng. struct. dyn. 11,
729-748 (1983).
3. N. Mostaghel and J. Tanbakuchi, Response of structures on slinky type supports’, ASCE nat. conu. Philadelphia. PA (1983).
4. C. E. Crede and J. E. Ruzicka.’Theorv ofvibration isolation’. in Shock and Vibration Handbook. Vol. 2. McGraw-Hill. New York. 1961.

5. R. R. Martel, ‘The effect of LrthqAkes on buildings with a flexible first story’, Bull. seism: soc. Am. 19, 167-178 (1929).
6. N. B. Green, ‘Flexible first story construction for earthquake resistance’, Trans. ASCE, 100, 644-674 (1935).
7. L. S. Jacobsen, ‘Effect of flexible first story in a building located on vibrating ground’, in S. Timoshenko 60th Anniversary Volume,
MacMillan, New York 1938.
8. M. Fintel and R. R. Khan, ‘Shock-absorbing soft-story concept for multistory earthquake structures’, J. Am. concrete inst., No. 66-29,
318-390 (19691.
~I

9. A. K. Chopra, D. P. Clough and R. W.Clough, ‘Earthquake resistance of building with a soft first story’,Earthquakeeng. strucr. dyn. 1,
347-355 (1973).
10. M. S. Caspe, ‘Earthquake isolation of multistory concrete structures’, J. Am. concrete inst. 67, 923-933 (1970).
11. K. Matsushita and M. Izumi, ‘Studies on mechanisms to decrease earthquake forces applied to buildings’, Proc. 4th world con&
earthquake eng. Santiago de Chile, I (1969).
12. J. M. Kelly, J. M. Eidinger and C. J. Derham, ‘A practical soft story isolation system’, Report No. UCB/EERC-77/27, Earthquake
Engineering Research Center, University of California, Berkeley, CA, 1977.
13. C. J. Derham, A. G. Thomas and J. M. Kelly, ‘A rubber bearing system for seismic protection of structures’, Proc. con& eng. designfor
earthquake enuironrnents, London, 1. Mech. E. Conf. Publ. 1978-12, pp. 53--58 (1978).
14. R. I. Skinner, J. M. Kelly and A. J. Heine, ‘Hysteretic dampers for earthquake resistant structures’, Earrhquake eng. struct. dyn. 3,
287-296 (1975).
15. J. M. Kelly and S. B. Hodder, ‘Experimental study of lead and elastomeric dampers for base isolation systems’, Report No.
UCBIEERC-8 1/16. Earthquake Engineering Research Center, University of California, Berkeley, CA, 1981.
16. J. M. Kelly, ‘The economic feasibility of seismic rehabilitation of buildings by base isolation’, Report No. UCB/EERC-L3/01,
Earthquake Engineering Research Center, University of California, Berkeley, CA, 1983.
17. C. Plichon and F. Jolivet, ‘Aseismicfoundation systems for nuclear power plants’, Proc. S.M.I.R.T. con$ London, England, Paper No.
C190/1978 (1978).
18. B. Pavot and E. Polus, ‘Aseismic bearing pads’, Tribol. int. 107-111, June (1974).
19. C. Plichon, R. Guerand, M. H. Richill and J. F. Casagrande, ‘Protection of nuclear power plants against seism’, Nucl. tech. 49,245-306
(1980).
20. N. R. Vaidya and A. J. Eggenberger, ‘Feasibility evaluation of base isolation for the aseismic design of structures’, DAppolonia,
Consulting Engineers, Project No. 82-1365, 1984.
21. M. Wakabayashi, T. Fujiwara, T. Nakamura and T. Bastov, ‘Experimental study on the dynamic characteristics of isolated structures’,
Bull. disaster preuention inst. Kyoto Univ. 31, Part 3, No. 282, 151-169 (1981).
22. J. M. Kelly and K. E. Beucke, ‘A friction damped base isolation system with fail-safe characteristics’, Earthquake eng. struct. dyn. 11,
33-56 (1 983).
23. S. H. Crandall, S. S. Lee and J. H. Williams, Jr., ‘Accumulatedslip of a frictioncontrolled mass excited by earthquake motions’, J. appl.
mech. ASME 41, 10941098 (1974).
24. I. G. Tadjbakhsh and J. J. Ma, ‘Rigid body response of base-isolated structures’, J. strucf. div. ASCE 108, 18W1814 (1982).
25. C. J. Younisand I. G. Tadjbakhsh, ‘Responseof sliding rigid structures to base excitation’,J. eng. mech. diu. ASCE 110,417431 (1983).
26. G. Ahmadi, ‘Stochastic earthquake response of structures on sliding foundation’, Int. j. eng. sci. 21, 93-102 (1983).
27. D. Chen and R. W. Clough. ‘Earthquake response of structures with friction sliding motion’, Earthquake Engineering Research
Center, University of California, Berkeley, CA, 1981.

You might also like