Fenz Constantinou - Artículo 2 (2007) PDF

EARTHQUAKE ENGINEERING AND STRUCTURAL DYNAMICS
Earthquake Engng Struct. Dyn. 2008; 37:185–205

Published online 20 August 2007 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/eqe.750
Spherical sliding isolation bearings with adaptive behavior:

Experimental verification
Daniel M. Fenz∗, † and Michael C. Constantinou‡

Department of Civil, Structural and Environmental Engineering, University at Buffalo, The State University
of New York, 212 Ketter Hall, Buffalo, NY 14260, U.S.A.
SUMMARY
This paper describes an experimental program to examine the force–displacement behavior of a class
of multi-spherical sliding bearings. The primary goal of the experiments is to test the validity of the
theory developed in a companion paper that describes the behavior of these devices. Experimental work
consisted of testing the three primary variations of these bearings in several configurations of different
friction and displacement capacities. Most tests were carried out at slow speeds; however, some testing
was also conducted at high speed (up to approximately 400 mm/s) to examine the behavior under dynamic
conditions. The results of experimental testing were generally found to be in very good agreement with
the analytical results. It is shown that the forces and displacements at which transitions in stiffness occur
are predictable and therefore controllable in design. Furthermore, the underlying principles of operation
were confirmed by the fact that starting and stopping of sliding on the different surfaces occurred as
expected from theory. Copyright q 2007 John Wiley & Sons, Ltd.
Received 16 March 2007; Revised 27 June 2007; Accepted 2 July 2007
KEY WORDS: adaptive seismic isolation system; Friction Pendulum bearing; bearing testing
1. INTRODUCTION
In a companion paper [1], the behavior of three variations of multi-spherical sliding bearings is
described. These devices are (a) the triple Friction Pendulum (FP) bearing, (b) the modified single
FP bearing and (c) the double FP bearing with concave surfaces of different displacement capacities.
The internal construction of these bearings permits sliding on different combinations of surfaces
∗ Correspondence to: Daniel M. Fenz, Department of Civil, Structural and Environmental Engineering, University at
Buffalo, The State University of New York, 212 Ketter Hall, Buffalo, NY 14260, U.S.A.
†
E-mail: dmfenz@buffalo.edu
‡
Professor.
Contract/grant sponsor: Multidisciplinary Center for Earthquake Engineering Research
Copyright q 2007 John Wiley & Sons, Ltd.

186 D. M. FENZ AND M. C. CONSTANTINOU
throughout the course of motion, resulting in changes in stiffness and damping. The behavior is
termed adaptive because the stiffness and damping change to predictable values at displacement
amplitudes that are calculable and controllable. This allows the design of the isolation system
to be separately optimized for multiple performance objectives and multiple levels of input. For
example, there can be high stiffness and low damping in minor events to reduce secondary system
response and improve re-centering, softening and intermediate damping for reduction of base shear
in the design earthquake and stiffening or further increase in damping to limit the isolation system
displacements in the maximum earthquake and beyond. It is shown in the companion paper that
this type of desirable behavior is theoretically possible for the three variations of bearing presented,
when properly configured.
This paper describes an experimental program of characterization testing that was performed to
assess the validity of the theory developed in [1]. The three variations of adaptive multi-spherical
sliding bearing were tested in many configurations. The primary objective of the experimental
program was to measure the force–displacement behavior for slow speed displacement-controlled
tests of various amplitudes in order to verify the analytical model and principles of operation. Slow
speed tests were favored since hysteresis loops obtained under nearly quasi-static test conditions
exhibited clear transitions in stiffness and permitted more accurate comparison to the analytical
model. A limited number of tests were also carried out at high speed to explore the behavior under
dynamic conditions and evaluate any effects of impact when the slider contacts the displacement
restrainer.
The experimental results generally are in good agreement with the analytical models from both a
quantitative and more importantly a qualitative standpoint. The latter means that the sequencing of
sliding regimes and the starting/stopping of sliding on the various surfaces are consistent with the
behavior that is predicted based on first principles. Demonstrating the predictability of the behavior
contributes toward showing that these devices combine the performance benefits of hybrid isolation
systems with the reliability of currently used technologies.
2. EXPERIMENTAL SCOPE AND OVERVIEW
2.1. Description of test specimens

Figure 1 shows the three reduced-scale specimens that were tested as part of the experimental
program. Each could be tested with different combinations of displacement capacity and coefficient
of friction. To change the displacement capacities of the outer concave surfaces, a urethane ring
(hardness 62D) was machined that could be easily inserted and removed. As shown in Figure 2,
the ring could be press fit into the outer concave plate to reduce the displacement capacity of this
surface by 25 mm. Frictional conditions were varied by lubricating the various sliding materials
with a silicone grease. Figure 2 and Table I summarize the configurations that were tested. The
nomenclature and definitions of parameters used in Table I and throughout this paper are consistent
with that used in [1].
The triple FP specimen was tested without the urethane ring, with the urethane ring in the top
concave plate and with the urethane ring in the bottom concave plate. The nominal displacement
capacity of the bearing was 166 mm without the urethane ring and 141 mm with the ring (actual
displacement capacities, however, are slightly different due to slider rotation effects). Though
the standard configuration of triple FP bearing is outer concave surfaces of equal displacement
Copyright q 2007 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2008; 37:185–205
DOI: 10.1002/eqe
SPHERICAL SLIDING ISOLATION BEARINGS WITH ADAPTIVE BEHAVIOR 187
(a)
(b)
(c)
Figure 1. Dimensions of the (a) triple FP; (b) modified single FP; and (c) double FP specimens tested.
capacity, configurations with different displacement capacities were tested to evaluate the validity
of the analytical model in more general cases.
Two different non-metallic sliding materials denoted Materials A and B were used on the sliding
interfaces. Each surface of the rigid slider was coated with Material A that had been lubricated,
the upper slide plate was coated with dry Material A and the lower slide plate with dry Material
B. When dry, the coefficient of friction of Material A is less than that of Material B. Lubrication
combined with the increased pressure due to the smaller contact area ensured that the coefficients
of friction of the rigid slider would be lower than those of the slide plates. At each sliding interface,
the radius of curvature of the slider was manufactured larger than the mating stainless steel surface.
As a result, there was bearing over an annular area on the perimeter of each slider which created
a pocket containing the lubricant.
DOI: 10.1002/eqe
(a)
(b)
(c)
Figure 2. Overview of the various configurations of (a) triple FP; (b) modified single FP; and (c) double
FP bearing with different displacement capacities that were tested.
The modified single FP specimen was assembled using the outer concave plate and slide plate
coated with Material A from the triple FP specimen. The post-plate and articulated slider from a
small-scale single FP bearing were used for the other half. The articulated slider was coated with
a lubricated non-metallic material denoted Material C that had coefficient of friction less than that
of dry Material A.
The double FP specimen was assembled using the same outer concave plates as the triple FP
specimen but with a different slider. The urethane insert was used to achieve concave surfaces
with different displacement capacities so the nominal displacement capacity for each configuration
tested was 127 mm. The articulated slider was coated with a different non-metallic sliding material
denoted Material D. Testing was performed with both surfaces of the slider dry to achieve nearly
equal coefficients of friction as well as with one surface lubricated to achieve substantially different
coefficients of friction.
2.2. Description of test apparatus and instrumentation

All tests were conducted using the small bearing testing machine in the Structural Engineering
and Earthquake Simulation Laboratory of the University at Buffalo. A detailed description of the
DOI: 10.1002/eqe
Table I. Summary of properties of the various configurations tested.

Nominal coefficients Nominal displacement Effective radii of
Configuration of friction capacities curvature (Ri − h i )
d2 = d3 = 19 mm Reff2 = Reff3 = 53 mm
Triple 1 2 = 3 <1 <4 d1 = 64 mm Reff1 = Reff4 = 435 mm
d4 = 64 mm
d2 = d3 = 19 mm Reff2 = Reff3 = 53 mm
d4 = 64 mm
d2 = d3 = 19 mm Reff2 = Reff3 = 53 mm
d4 = 39 mm
Modified d1 = 64 mm Reff1 = 455 mm

1 >2
single 1 d2 = 25 mm Reff2 = 73 mm
d1 = 51 mm Reff1 = 441 mm
Double 1 1 ≈ 2
d2 = 76 mm Reff2 = 437 mm
d1 = 76 mm Reff1 = 441 mm
Double 2 1 <2
d2 = 51 mm Reff2 = 437 mm
d1 = 51 mm Reff1 = 441 mm
Double 3 1 <2
d2 = 76 mm Reff2 = 437 mm
design and capabilities of the testing machine is given in Reference [2]. Essentially, it consists of
two vertical actuators and one horizontal actuator that transmit loads to the bearing via a large
loading beam. The vertical actuators are force-controlled to accommodate changes in the bearing’s
height while still maintaining a predefined (constant or otherwise) vertical force. The horizontal
actuator is displacement-controlled and imposes a predefined displacement history.
Horizontal and vertical force data are collected by a five-channel load cell that is mounted
directly beneath the bearing. Horizontal displacement is measured by an LVDT that is internal to
the horizontal actuator. However, this is only the total displacement of the top part of the bearing
with respect to the bottom, u. Individual displacements of the various sliders were measured using
string potentiometers that were mounted on a stationary reference. Two potentiometers were used
for recording the horizontal displacements of the internal components of the triple FP bearing, one
attached to each slide plate. In addition to the total displacement, in this arrangement it was possible
to measure the displacement of the top slide plate with respect to the top concave surface, u 1 , the
displacement of the two slide plates with respect to each other, u 2 + u 3 , and the displacement of
the lower slide plate with respect to the bottom concave surface, u 4 . One potentiometer was used
to decompose the displacements for the modified single FP specimen and the double FP specimen.
The string was attached to the slide plate of the modified single FP bearing and at mid-height of
the articulated slider in the double FP bearing.
DOI: 10.1002/eqe
Table II. Abridged list of tests performed.

Vertical Nominal peak
Configuration load (kN) Frequency (Hz) Amplitude (mm) total velocity (mm/s)
0.10 1.2 0.75

0.04 25 6.3
Triple 1 112 0.013 75 6.3
0.0088 115 6.3
0.0072 140 6.3
0.01 115 7.2

Triple 1 0.10 115 72
(high-velocity 112 0.30 115 217
tests) 0.50 115 361
0.60 115 434
0.0099 100 6.3

Triple 2 112
0.0084 120 6.3
0.0099 100 6.3

Triple 3 112
0.0088 115 6.3
0.16 1 1.0
Modified
28 0.040 25 6.3
single 1
0.014 70 6.3
0.013 75 6.3
Double 1 94
0.0088 115 6.3
0.16 1.5 1.5

Double 2 47 0.013 75 6.3
0.0088 115 6.3
0.16 1.5 1.5

0.013 75 6.3
Double 3 47
0.011 95 6.5
0.0088 115 6.3
2.3. Testing program and procedures

The majority of testing focused on evaluating each configuration at several amplitudes of dis-
placement to determine the force–displacement behavior exhibited during each sliding regime. All
tests were conducted under a constant vertical load with three fully reversed cycles of sinusoidal
horizontal displacement. Vertical loads are listed in Table II and were chosen to give pressures that
were consistent with what are typically used in practice, 13.8–55 MPa [3]. The exception to this is
the double FP in configurations 2 and 3, where the vertical load was reduced in order to increase
the differences in the coefficients of friction of the two sliding interfaces. Also in the modified
single FP specimen, due to the differences in the two areas of the articulated slider and the slide
plate it was not possible to have both pressures within this range.
DOI: 10.1002/eqe
Figure 3. Illustration of the range in the coefficient of friction for low-speed and high-speed testing.
Most tests were performed at very low velocity as this gave hysteresis loops with clearly defined
transition points. As shown in Table II which describes the tests reported, the peak total sliding
velocity was typically only 6.3 mm/s. Hysteresis loops from a limited number of tests conducted
at high velocity were more rounded due to velocity dependence of the coefficient of friction
[4], which tended to mask the transitions in stiffness. At low velocity, the coefficient of friction
is highly dependent on velocity as they are related by a negative exponential relationship. The
relative differences are substantial, but there is still small absolute difference in the range of values
compared with high speed. This is illustrated in Figure 3.
Although high-speed tests are more representative of seismic conditions, accurate extraction
of the experimental values of force and displacement at which the various sliding regimes start
and stop was necessary to verify the fundamental mechanical behavior of the bearings. After the
basic principles of operation of these devices are verified, velocity dependence of the coefficient
of friction can be incorporated into the formulation of more sophisticated dynamic models.
Prior to every test, the loading beam was raised and the various sliders were centered within
the bearing. This was to ensure that the displacement capacity of each surface was the same for
both positive and negative displacements. The stiffening behavior of adaptive bearings is highly
dependent on the displacement capacity of each surface. Therefore, if the slider starts off center
there will be asymmetry in the force–displacement loops. In addition, to limit the effects of
contamination on the coefficient of friction the stainless steel interfaces were cleaned periodically
during testing with isopropyl alcohol and a soft cloth.
3. EXPERIMENTAL RESULTS FOR THE TRIPLE FP BEARING
3.1. Data analysis and construction of analytical force–displacement loops

The analytical loops presented throughout this paper follow from the theory presented in [1]. Since
the exact values of the coefficients of friction are not known a priori, analytical loops must be
constructed using experimentally measured values of the coefficients of friction from each test. The
values of the coefficients of friction on surfaces 1 and 4 were determined from the decomposed
force–displacement loops as the three cycle average normalized force at zero relative displacement
(decomposed loops refer to the hysteresis loops that are obtained when the horizontal force is
plotted against relative displacement on a particular sliding surface). This is the typical procedure
DOI: 10.1002/eqe
for experimentally determining the coefficient of sliding friction [3]. The coefficient of friction of
surfaces 2 and 3 were determined from the decomposed force–displacement loops as half of the
drop in normalized force that occurs upon reversal of motion. This procedure had to be employed
for surfaces 2 and 3 since their force–displacement behavior is atypical.
Appendix Table AI lists the experimentally measured values of the coefficient of friction of each
surface that were used to construct the analytical loops. The variability in the friction coefficients,
most noticeably in 1 for configuration 1, is attributed chiefly to velocity dependence. For example,
the sliding velocities on surface 1 at zero relative displacement were 5.7, 6.0, 3.4 and 3.4 mm/s
in sliding regimes II, III, IV and V, respectively.
0.4
Experimental
0.3
0.2
Horizontal Force
0.1
Vertical Force
0.0
-0.1
-0.2
-0.3
-0.4
-150 -100 -50 0 50 100 150
0.4
Analytical
0.3
0.2
Horizontal Force
0.1
Ver tical Force
0.0
-0.1
-0.2 µ1 = 0.021 - 0.041

µ 2 = µ3 = 0.010 - 0.025
-0.3 µ4 = 0.099 - 0.129
d1 = d4 = 61 mm
-0.4
-150 -100 -50 0 50 100 150
Total Displacement, u (mm)
Figure 4. Comparison of analytical and experimental results for force–displacement relationship of triple
FP specimen in configuration 1. All five sliding regimes are shown.
DOI: 10.1002/eqe
In addition, the actual displacement capacities of each surface are used to construct the analytical
loops. The nominal displacement capacity is determined using only the diameters of the articulated
slider and concave plate. The actual displacement capacity is defined as the relative displacement
of the pivot point when the slider contacts the displacement restrainer. It is the nominal value
reduced by an amount proportional to the slider height and inversely proportional to the surface’s
radius of curvature. The difference between nominal and actual values is small in the typical
case where the slider height is small compared with the radius of curvature. The values of the
experimental displacement capacity reported in Appendix Table AI were determined from the
decomposed force–displacement loops as the displacement where substantial stiffening occurred.
Lastly, where sliding velocities are reported, they have been determined numerically from the
displacement data by fitting a cubic polynomial through a moving five-point window. The derivative
of the polynomial is evaluated at the mid-point of the window and the resulting data are smoothed.
3.2. Quantitative comparison of experimental and analytical results

The experimental data presented in Figure 4 show one test typical of each sliding regime for
the triple FP bearing in configuration 1. In general, there is very good agreement between the
experimental data and the analytical loops constructed using the actual values of friction and
displacement capacity. As summarized in Table III, the analytical model matches quite well in
terms of both the values of normalized force and the displacements at which transitions in stiffness
occur. It is emphasized that the increase in strength exhibited by the bearing with increasing
displacement is a result of its internal mechanical behavior, not of the variability in the coefficients
of friction due to velocity effects. The increase is a result of sliding occurring on surface 4, which
is of substantially higher friction.
The experimentally recorded force–displacement loops also show very little variability from
cycle to cycle. Provided the coefficients of friction remain consistent, the transitions in stiffness
will occur at the same forces and displacements from cycle to cycle. There is no ratcheting or
accumulation of displacement offsets with repeated cycling.
The largest discrepancy between theory and experiment occurs in the 140 mm amplitude test,
where there is slight stiffening prior to the slider contacting the displacement restrainer on surface 4.
This is not captured in the analytical model. Also, the stiffness exhibited for the final phase of sliding
regime V is somewhat higher than what is predicted by theory. The experimental value of stiffness
determined by linear regression is W/(82 mm), however, the theoretical value is W/(106 mm).
This leads to an underestimation of the force at maximum displacement by approximately 15%,
despite very close prediction of the displacement at which the transition in stiffness occurs, u dr4 .
This discrepancy is due to a combination of the greater error in the small angle assumption (at the
peak displacement of 140 mm the rigid slider is inclined at 16◦ ) and errors introduced in the test
apparatus and load cell at large displacements.
The agreement between the experimental and analytical behavior is also very good for the other
configurations of triple FP bearing. Experimental and analytical data for configurations 2 and 3
are shown in Figures 5 and 6, respectively. Only regimes IV and V were tested since prior to
contacting the displacement restrainer, the force–displacement behavior of configurations 2 and
3 is the same as configuration 1. Referring to Table III, the forces and displacements at which
transitions in stiffness occur are predicted well.
In configurations 2 and 3, the stiffness during the final phase of sliding regime V (sliding on
surfaces 2 and 3) is slightly overpredicted by the theory. The analytical model predicts a stiffness
DOI: 10.1002/eqe
194
Copyright q
Table III. Comparison of analytical and experimental results for triple FP bearing.
Regime II Regime III Regime IV Regime V
Analytical Experimental Analytical Experimental Analytical Experimental Analytical Experimental
Configuration 1
u ∗ (mm)†,‡ 2.9 2 2.5 3 1.3 2 0.1 2
u ∗∗ (mm)‡ — — 45.5 51 48.6 41 38.4 42
u dr1 (mm)§ — — — — 86.2 85 92.1 90
u dr4 (mm)§ — — — — — — 130.4 130
Fdr1 § — — — — 0.162W 0.172W 0.161W 0.173W
2007 John Wiley & Sons, Ltd.

Fdr4 § — — — — — — 0.240W 0.272W
Configuration 2
u ∗ (mm)‡ — — — — 1.6 2.8 0.9 1.9
u ∗∗ (mm)‡ — — — — 29.0 29.3 18.4 27.6
u dr1 (mm)§ — — — — 54.2 52.6 63.3 60.9
u dr4 (mm)§ — — — — — — 106.5 108.0
Fdr1 § — — — — 0.113W 0.111W 0.113W 0.111W
Fdr4 § — — — — — — 0.201W 0.213W
Configuration 3
u ∗ (mm)‡ — — — — 1.1 1.8 1.3 2.6
D. M. FENZ AND M. C. CONSTANTINOU
u ∗∗ (mm)‡ — — — — 27.9 36.7 22.0 34.6

u dr1 (mm)§ — — — — 102.1 99.3 102.2 102.2
u dr4 (mm)§ — — — — — — 96.0 93.7
Fdr1 § — — — — 0.166W 0.170W 0.169W 0.175W
Fdr4 § — — — — — — 0.156W 0.158W
†
For definitions of parameters and equations used to determine theoretical values, see [1].
‡
Experimental value is first cycle value.
§ Experimental value is three cycle average value.
DOI: 10.1002/eqe
Earthquake Engng Struct. Dyn. 2008; 37:185–205
0.4 0.4
Experimental Analytical
0.3 0.3
Horizontal Force 0.2 0.2
Vertical Force
0.1 0.1
0.0 0.0
-0.1 -0.1 1 = 0.025 - 0.028
-0.2 -0.2 2 = 3 = 0.013 - 0.017
-0.3 -0.3 = 0.061 - 0.084

4
d1 = 37 mm, d4 = 61 mm
-0.4 -0.4
-150 -100 -50 0 50 100 150 -150 -100 -50 0 50 100 150
Total Displacement , u (mm) Total Displacement , u (mm)
FP specimen in configuration 2. Sliding regimes IV and V are shown.
0.3 0.3
0.2 0.2
Horizontal Force
Vertical Force
0.1 0.1
0.0 0.0
-0.1 -0.1 1 = 0.027 - 0.029

2 = 3 = 0.016 - 0.017
-0.2 -0.2 = 0.071 - 0.081
4
d1 = 61 mm, d4 = 37 mm
-0.3 -0.3
-150 -100 -50 0 50 100 150 -150 -100 -50 0 50 100 150
FP specimen in configuration 3. Sliding regimes IV and V are shown.
of W/(106 mm) and the experimental values determined by linear regression range between
W/(116 mm) and W/(125 mm). The difference is due to the assumption of a rigid displace-
ment restrainer in the analytical model. This assumption is reasonable when the slider is bearing
on the steel displacement restrainer that is typically used. In the test specimen, however, the ure-
thane insert is quite stiff but not completely rigid (Durometer 62D is approximately the hardness
of a golf ball).
3.3. Experimental verification of principles of operation

There is also good qualitative agreement between the theory and experiment since starting and
stopping of motion on the different surfaces occur as predicted. Five different sliding regimes are
expected for the standard configuration of triple FP bearing over the course of motion. They are,
DOI: 10.1002/eqe
in order of increasing amplitude of maximum displacement:

1. Initial sliding on surfaces 2 and 3, the two surfaces of least and equal friction.
2. Sliding on surfaces 1 and 3 with constant nonzero displacement on surface 2 (the instant
sliding starts on surface 1 it stops on surface 2).
3. Sliding on surfaces 1 and 4 with constant nonzero displacements on surfaces 2 and 3 (the
instant sliding starts on surface 4 it stops on surface 3).
4. Sliding on surfaces 2 and 4 when the slide plate contacts the displacement restrainer on
surface 1.
5. Sliding on surfaces 2 and 3 when the slide plate contacts the displacement restrainer on
surface 4 in addition to surface 1.
In addition to real-time observation and test video, this behavior is demonstrated by the decom-
posed force–displacement loops in Figure 7 and the histories of displacement and velocity for the
140 mm amplitude test in Figure 8. Both are for configuration 1.
The decomposed loops confirm that for very small amplitudes of displacement there is sliding
only on surfaces 2 and 3. When the friction force exceeds Ff1 , motion starts on surface 1 and stops
on surface 2. The velocity history shows that as soon as motion starts on surface 1 the combined
velocity of surfaces 2 and 3 drops substantially (but not quite to zero), indicating that motion has
stopped on surface 2 but continues on surface 3. The velocity history gives similar results when
motion starts on surface 4. Immediately after motion starts on surface 4, the combined velocity
of surfaces 2 and 3 goes to zero and remains zero while sliding is occurring only on surfaces 1
and 4.
The instant motion starts on surface 4, the combined displacement of surfaces 2 and 3 should be
6.4 mm and remain constant at this value. The experimentally measured value of this displacement
is approximately 7 mm. Theory predicts that this displacement should remain constant as long as
sliding is occurring on both surfaces 1 and 4. However, the experimental data in Figure 7 for
surfaces 2 and 3 during sliding regimes III–V show a slight reduction in this value as the force
increases. This reduction (marked by the segments with negative slope) is due to error induced
in the string pot measurements by the rotation of the slide plates. Though the slide plates do
not translate relative to one another, when they rotate the lengths of the string potentiometers
change since they are not attached at the pivot point. One string will lengthen and the other will
shorten giving the impression that the relative displacement between the two is actually decreasing.
This occurs slowly compared with translation and can be distinguished from true relative sliding
between the plates using the velocity history.
During sliding regime IV, motion resumes on surface 2 as the slider contacts the displacement
restrainer of surface 1. This is evidenced both by the hysteresis loop of surfaces 2 and 3 during
regime IV and the velocity history of Figure 8. When the velocity on surface 1 drops to zero,
there is a slight increase in the combined velocities of surfaces 2 and 3. Furthermore, the de-
composed hysteresis loop of surface 1 agrees well with the assumed behavior for a slider that
contacts the displacement restrainer. It demonstrates that for surface i whose slider is bearing
on the displacement restrainer, upon reversal of motion sliding will resume when the bearing
has unloaded by Fri + 2Ffi to Fdri − 2Ffi . The very large stiffness after contacting the displace-
ment restrainer compared with the stiffness exhibited during sliding also confirms the rigidity
assumption.
At the start of sliding regime V, both slide plates have contacted the displacement restrainers and
there is sliding only on surfaces 2 and 3. The sliding velocities on surfaces 1 and 4 both are zero
DOI: 10.1002/eqe
0.4 0.4 0.4

Regime I Regime I Regime I
Horizontal Force
0.2 Surface 1 0.2 Surfaces 2 & 3 0.2 Surface 4
Vertical Force
0.0 0.0 0.0
-0.2 -0.2 -0.2
-0.4 -0.4 -0.4

-75 -50 -25 0 25 50 75 -20 -10 0 10 20 -75 -50 -25 0 25 50 75
0.4 0.4 0.4
Regime II Regime II Regime II
Surface 1 Surfaces 2 & 3 Surface 4
0.2 0.2 0.2
Horizontal Force
Vertical Force
0.0 0.0 0.0
-0.2 -0.2 -0.2
-0.4 -0.4 -0.4

-75 -50 -25 0 25 50 75 -20 -10 0 10 20 -75 -50 -25 0 25 50 75
0.4 0.4 0.4
Regime III Regime III Regime III
0.2 0.2 0.2
Horizontal Force
Vertical Force
0.0 0.0 0.0
-0.2 -0.2 -0.2
-0.4 -0.4 -0.4

-75 -50 -25 0 25 50 75 -20 -10 0 10 20 -75 -50 -25 0 25 50 75
0.4 0.4 0.4
Regime IV Regime IV Regime IV
0.2 0.2 0.2
Horizontal Force
Vertical Force
0.0 0.0 0.0
-0.2 -0.2 -0.2
-0.4 -0.4 -0.4

-75 -50 -25 0 25 50 75 -20 -10 0 10 20 -75 -50 -25 0 25 50 75
0.4 0.4 0.4
Regime V Regime V Regime V
0.2 0.2 0.2
Horizontal Force
Vertical Force
0.0 0.0 0.0
-0.2 -0.2 -0.2
-0.4 -0.4 -0.4

-75 -50 -25 0 25 50 75 -20 -10 0 10 20 -75 -50 -25 0 25 50 75
u1 (mm) u2 + u3 (mm) u4 (mm)
Figure 7. Decomposed experimental force–displacement loops for triple FP specimen in configuration 1.
DOI: 10.1002/eqe
Total
Surface 1
Surfaces 2 & 3
Surface 4
150
100
Displacement (mm)
50
-50
-100
-150
0 50 100 150 200 250 300 350 400 450
Time (sec)
8
6
4
Velocity (mm/s)
2
0
-2
-4
-6
-8
0 50 100 150 200 250 300 350 400 450
Time (sec)
Figure 8. Experimental histories of displacement and velocity for triple FP bearing in configuration 1
during cyclic testing of 140 mm amplitude.
and the combined velocity of surfaces 2 and 3 overlays exactly the total velocity. The hysteresis
loops of surfaces 2 and 3 demonstrate rigid linear behavior after the displacement restrainer is
contacted on surface 4.
Lastly, the phenomenon of slider offset can be seen in the histories of displacement and velocity
of Figure 8. When the total displacement is zero, the individual displacements on each of the
sliding surfaces are all nonzero (though they must sum to zero). For the 140 mm amplitude test,
the experimental values of the offsets are −13.4 mm for surface 1, −6.4 mm for surfaces 2 and 3
combined and 19.8 mm for surface 4 (three cycle average values). These are in good agreement
with the values obtained by tracking the analytical force–displacement loops: −14.9, −4.3 and
19.2 mm, respectively. The offset results from the displacement controlled nature of the testing
and is different from the permanent total displacements after earthquake excitation.
DOI: 10.1002/eqe
0.3 0.3
0.01 Hz Frequency 0.10 Hz Frequency
Horizontal Force 0.2 Peak Velocity = 7.2 mm/s 0.2 Peak Velocity = 72 mm/s
0.1 0.1
Vertical Force
0.0 0.0
-0.1 -0.1
-0.2 -0.2
-0.3 -0.3
-150 -100 -50 0 50 100 150 -150 -100 -50 0 50 100 150
0.3 0.3
0.30 Hz Frequency 0.50 Hz Frequency
0.2 Peak Velocity = 216 mm/s 0.2 Peak Velocity = 359 mm/s
Horizontal Force
Vertical Force
0.1 0.1
0.0 0.0
-0.1 -0.1
-0.2 -0.2
-0.3 -0.3
-150 -100 -50 0 50 100 150 -150 -100 -50 0 50 100 150
0.3
0.60 Hz Frequency
0.2 Peak Velocity = 431 mm/s
Horizontal Force
Vertical Force
0.1
0.0
-0.1
-0.2
-0.3
-150 -100 -50 0 50 100 150
Figure 9. Force–displacement loops for triple FP specimen in configuration 1 during testing at 115 mm
amplitude and increasing frequency of sinusoidal motion. Peak velocities are experimentally measured
values and differ slightly from the nominal or target values.
DOI: 10.1002/eqe
3.4. Comments on tests at high speed

The results of the tests carried out at high speed are shown in Figure 9 and indicate that the
behavior under dynamic conditions is essentially the same as the behavior exhibited under quasi-
static conditions. However, due to the wider range in values of the coefficients of friction, the
transitions in stiffness between sliding regimes are less clear. Gradual transitions are actually
beneficial when considering the response of secondary systems as abrupt changes in stiffness may
lead to excitation of higher modes. Also, the consistency of behavior from the 0.10 Hz frequency
test to the 0.60 Hz frequency test should be noted. Although the peak sliding velocity on each
surface increases with increasing frequency, the velocities are large enough so that the coefficient
of friction remains constant with increasing velocity [4].
Furthermore, the high-speed tests do not exhibit any spikes or substantial fluctuation in horizontal
force when the slider contacts the displacement restrainer. The sliding velocity when the slider
contacts the displacement restrainer is much less than the peak total velocity. This is due to the fact
that (a) the slider typically contacts the displacement restrainer at larger displacements when the
total velocity is typically much less than the peak value and (b) sliding is shared among two sliding
interfaces. For example, in the 115 mm amplitude and 0.50 Hz test, the peak total velocity achieved
was 359 mm/s, however, the sliding velocity on surface 1 just prior to the slider contacting the
displacement restrainer was approximately 25 mm/s.
4. EXPERIMENTAL RESULTS FOR THE MODIFIED SINGLE FP BEARING
Data analysis and construction of analytical loops for the modified single FP bearing were carried
out in the same manner as for the triple FP bearing. Experimental and analytical loops are pre-
sented in Figure 10. The figures show that the experimental force–displacement data are in very
good agreement with the analytical results. Moreover, the experimental histories of displacement
and velocity for the 70 mm amplitude test demonstrate behavior consistent with theory. Figure 11
0.4 0.4
0.2 0.2
Horizontal Force
Vertical Force
0.0 0.0
-0.2 -0.2 µ 1 = 0.061 - 0.067

µ 2 = 0.021 - 0.045
d1 = 61 mm
-0.4 -0.4
-75 -50 -25 0 25 50 75 -75 -50 -25 0 25 50 75
Figure 10. Comparison of analytical and experimental results for force–displacement relationship of
modified single FP bearing in configuration 1. All three sliding regimes are shown.
DOI: 10.1002/eqe
Total
Surface 1
Surface 2
75
50
Displacement (mm)
25
-25
-50
-75
0 50 100 150 200 250
Time (sec)
8
6
4
Velocity (mm/s)
2
0
-2
-4
-6
-8
0 50 100 150 200 250
Time (sec)
Figure 11. Experimental histories of displacement and velocity of modified single FP bearing in
configuration 1 for the 70 mm amplitude test.
shows that there is no simultaneous sliding on both surfaces (except for very brief instances due
to the inertia of the moving parts). Due to their similarities in construction, this helps to confirm
that simultaneous sliding on surfaces 1 and 2 (or 3 and 4) for the triple FP bearing does not occur.
5. EXPERIMENTAL RESULTS FOR DOUBLE FP BEARING WITH CONCAVE SURFACES

OF DIFFERENT DISPLACEMENT CAPACITIES
Experimental and analytical force–displacement data are shown in Figures 12–15 for the three
configurations of double FP tested. The experimental data agree well with the theoretical results
and demonstrate that desirable changes in stiffness and damping occur with increasing amplitude
of displacement. It should be noted that in configuration 3, the double FP bearing exhibits increases
in both stiffness and damping at large displacements.
DOI: 10.1002/eqe
0.2 0.2
0.1 0.1
Horizontal Force
Vertical Force
0.0 0.0
-0.1 -0.1 µ1 = 0.018 - 0.021

µ2 = 0.024 - 0.029
d1 = 48 mm
-0.2 -0.2
-150 -100 -50 0 50 100 150 -150 -100 -50 0 50 100 150
Total Displacement, u (mm) Total Displacement, u (mm)
Figure 12. Comparison of analytical and experimental results for force–displacement relationship of double
FP bearing in configuration 1. Both sliding regimes are shown.
0.2 0.2
0.1 0.1
Horizontal Force
Vertical Force
0.0 0.0
-0.1 -0.1 µ1 = 0.015 - 0.016

µ 2 = 0.022 - 0.024
d2 = 50 mm
-0.2 -0.2
-150 -100 -50 0 50 100 150 -150 -100 -50 0 50 100 150
FP bearing in configuration 2. All three sliding regimes are shown.
Figure 15 shows the velocity histories of the bearing in configuration 3 for the 95 mm ampli-
tude test (sliding regime III(a)) and for the 115 mm amplitude test (sliding regime III(b)). For
both tests, the loading behavior is identical and consistent with theory. Upon reversal of motion,
the two plots demonstrate the different sequences in which sliding can resume. In the 95 mm
amplitude test, Equation (40) of [1] is not satisfied (u max = 95 mm<98.2 mm) and motion re-
sumes on surface 1 prior to surface 2. In the 115 mm amplitude test, the equation is satisfied
(u max = 115 mm>102.4 mm) and motion resumes on surface 2 prior to surface 1. It should also be
noted that in configuration 2, the condition expressed by Equation (40) of [1] was always satisfied
since 1 <2 . This phenomenon admittedly may have a small impact on global structure response;
however, the fact that the analytical model captures such nuanced behavior gives added confidence
in its overall validity.
DOI: 10.1002/eqe
0.2 0.2
Horizontal Force 0.1 0.1

Vertical Force
0.0 0.0
-0.1 -0.1 µ1 = 0.011 - 0.016

µ2 = 0.022 - 0.029
d1 = 48 mm
-0.2 -0.2
-150 -100 -50 0 50 100 150 -150 -100 -50 0 50 100 150
Total Displacement, u (mm) Total Displacement, u (mm)
FP bearing in configuration 3. All three sliding regimes are shown.
Total
Surface 1
Surface 2
8
Regime III
6 95 mm Amplitude
4
Velocity (mm/s)
2
0
-2
-4
-6
-8
0 50 100 150 200 250 300
(a) Time (sec)
8
Regime III
6 115 mm Amplitude
4
Velocity (mm/s)
2
0
-2
-4
-6
-8
0 50 100 150 200 250 300 350
(b) Time (sec)
Figure 15. Histories of velocity for the double FP bearing in configuration 3 during the 95 and 115 mm
amplitude tests demonstrating the different possible orders of unloading.
DOI: 10.1002/eqe
6. CONCLUSION
This and the companion paper have described and experimentally verified the principles of opera-
tion and force–displacement relationships of a class of multi-spherical sliding bearings that exhibit
adaptive behavior. These bearings are very appealing due to the desirable changes and stiffness
and damping that they exhibit with increasing input. Furthermore, they are simple passive devices
derived from currently used technologies and constructed from basic materials of proven reliabil-
ity. The theory that has been developed follows directly from consideration of equilibrium and
geometry. The devices have been tested in a variety of configurations with different displacement
capacities and coefficients of friction and in each case the analytical model closely matches the
experimental results. There is good agreement both in a numerical sense and in the prediction of
initiation and stoppage of sliding on the various concave surfaces. This work confirms that the
basic theory presented in Reference [1] is valid.
APPENDIX
Appendix Table AI lists the experimentally measured values of the coefficient of friction and actual
displacement capacity of each surface that were used to construct the analytical loops.
Table AI. Coefficients of friction and displacement capacities measured during experiments used to
construct analytical force–displacement loops.
Peak Coefficients Displacement
Configuration displacement (mm) Regime of friction capacities
1.2 I 1 = −, 2 = 3 = 0.022, 4 = −
25 II 1 = 0.038, 2 = 3 = 0.010, 4 = −
d1 = 61 mm
Triple 1 75 III 1 = 0.041, 2 = 3 = 0.025, 4 = 0.129
d4 = 61 mm
115 IV 1 = 0.022, 2 = 3 = 0.010, 4 = 0.119
140 V 1 = 0.021, 2 = 3 = 0.019, 4 = 0.099
100 IV 1 = 0.028, 2 = 3 = 0.013, 4 = 0.084 d1 = 37 mm

Triple 2
120 V 1 = 0.025, 2 = 3 = 0.017, 4 = 0.061 d4 = 61 mm
100 IV 1 = 0.027, 2 = 3 = 0.016, 4 = 0.081 d1 = 61 mm
Triple 3
115 V 1 = 0.029, 2 = 3 = 0.017, 4 = 0.071 d4 = 37 mm
1 I 1 = −, 2 = 0.035
Modified
25 II 1 = 0.067, 2 = 0.021 d1 = 61 mm
single 1
70 III 1 = 0.061, 2 = 0.045
75 I 1 = 0.021, 2 = 0.024
Double 1 d1 = 48 mm
115 II 1 = 0.018, 2 = 0.029
1.5 I 1 = 0.015, 2 = −
Double 2 75 II 1 = 0.016, 2 = 0.022 d2 = 50 mm
115 III 1 = 0.015, 2 = 0.024
1.5 I 1 = 0.016, 2 = −
75 II 1 = 0.015, 2 = 0.022
Double 3 d1 = 48 mm
95 III(a) 1 = 0.016, 2 = 0.025
115 III(b) 1 = 0.011, 2 = 0.029
DOI: 10.1002/eqe
ACKNOWLEDGEMENTS
Financial support for this project was provided by the Multidisciplinary Center for Earthquake Engineering
Research (Thrust Area 2). Earthquake Protection Systems, Inc. provided the specimens that were tested.
This support is gratefully acknowledged.
REFERENCES
1. Fenz DM, Constantinou MC. Spherical sliding isolation bearings with adaptive behavior: Theory. Earthquake
Engineering and Structural Dynamics 2007; DOI: 10.1002/eqe.751.
2. Kasalanati A, Constantinou MC. Experimental study of bridge elastomeric and other isolation and energy dissipation
systems with emphasis on uplift prevention and high velocity near source seismic excitation. Technical Report
MCEER-99-0004, Multidisciplinary Center for Earthquake Engineering Research, State University of New York
at Buffalo, Buffalo, NY, 1999.
3. Constantinou MC, Whittaker AS, Fenz DM, Apostolakis G. Seismic isolation of bridges: Version 2. Report to
Sponsor, California Department of Transportation, June 2007.
4. Mokha A, Constantinou MC, Reinhorn AM. Teflon bearings in base isolation. I: Testing. Journal of Structural
Engineering (ASCE) 1990; 116(2):438–454.
DOI: 10.1002/eqe

Fenz Constantinou - Artículo 2 (2007) PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Fenz Constantinou - Artículo 2 (2007) PDF

Uploaded by

Copyright:

Available Formats

EARTHQUAKE ENGINEERING AND STRUCTURAL DYNAMICS

Earthquake Engng Struct. Dyn. 2008; 37:185–205

Spherical sliding isolation bearings with adaptive behavior:

Daniel M. Fenz∗, † and Michael C. Constantinou‡

Received 16 March 2007; Revised 27 June 2007; Accepted 2 July 2007

Contract/grant sponsor: Multidisciplinary Center for Earthquake Engineering Research

Copyright q 2007 John Wiley & Sons, Ltd.

2. EXPERIMENTAL SCOPE AND OVERVIEW

2.1. Description of test specimens

2.2. Description of test apparatus and instrumentation

Table I. Summary of properties of the various configurations tested.

Modified d1 = 64 mm Reff1 = 455 mm

Table II. Abridged list of tests performed.

0.10 1.2 0.75

0.01 115 7.2

0.0099 100 6.3

0.0099 100 6.3

0.16 1.5 1.5

0.16 1.5 1.5

2.3. Testing program and procedures

3. EXPERIMENTAL RESULTS FOR THE TRIPLE FP BEARING

3.1. Data analysis and construction of analytical force–displacement loops

-0.2 µ1 = 0.021 - 0.041

3.2. Quantitative comparison of experimental and analytical results

Analytical Experimental Analytical Experimental Analytical Experimental Analytical Experimental

2007 John Wiley & Sons, Ltd.

u ∗∗ (mm)‡ — — — — 27.9 36.7 22.0 34.6

-0.3 -0.3 = 0.061 - 0.084

-0.1 -0.1 1 = 0.027 - 0.029

3.3. Experimental verification of principles of operation

in order of increasing amplitude of maximum displacement:

0.4 0.4 0.4

0.0 0.0 0.0

-0.2 -0.2 -0.2

-0.4 -0.4 -0.4

0.0 0.0 0.0

-0.2 -0.2 -0.2

-0.4 -0.4 -0.4

0.0 0.0 0.0

-0.2 -0.2 -0.2

-0.4 -0.4 -0.4

0.0 0.0 0.0

-0.2 -0.2 -0.2

-0.4 -0.4 -0.4

0.0 0.0 0.0

-0.2 -0.2 -0.2

-0.4 -0.4 -0.4

Figure 7. Decomposed experimental force–displacement loops for triple FP specimen in configuration 1.

3.4. Comments on tests at high speed

4. EXPERIMENTAL RESULTS FOR THE MODIFIED SINGLE FP BEARING

-0.2 -0.2 µ 1 = 0.061 - 0.067

5. EXPERIMENTAL RESULTS FOR DOUBLE FP BEARING WITH CONCAVE SURFACES

-0.1 -0.1 µ1 = 0.018 - 0.021

-0.1 -0.1 µ1 = 0.015 - 0.016

Horizontal Force 0.1 0.1

-0.1 -0.1 µ1 = 0.011 - 0.016

100 IV 1 = 0.028, 2 = 3 = 0.013, 4 = 0.084 d1 = 37 mm

You might also like