7th International Seminar on Seismic Isolation, Passive Energy Dissipation and Active Control of Vibrations of Structures Assisi, Italy, October 2-5, 2001

SEISMIC ISOLATION USING SLIDE AND RUBBER BEARINGS: LARGE AMPLITUDE FREE VIBRATION TESTS ON "RAPOLLA RESIDENCE BUILDING" F. Braga Structural Engineering Department - University “La Sapienza” - Roma, Italy [email protected] M. Laterza, R. Gigliotti Structural Engineering Department - University of Basilicata - Potenza, Italy [email protected], [email protected]

ABSTRACT An application of a mixed seismic isolation system (slide bearings for isolation and steel rubber bearings to have a recentering force) is presented. Such technique has been applied to a residence building, built in Italy at Rapolla (PZ- Basilicata). The application has allowed to investigate, through a series of dynamic Snap-Back tests, the behaviour of base isolated structure with only elastomeric bearings or with joined system (slide bearings and elastomeric bearings). Tests were carried out by using a mechanical device purposely designed to give the initial displacement, of the order of design base drift equal to 180 mm, to the building and then produce free damped vibrations acquired by using an accelerometer station. As far as the protection level is concerned, the seismic response of the structure on the mixed isolation system has been compared with that of the structure mounted on a steel rubber isolation system, confirming the effectiveness of the slide bearing technique.

1. INTRODUCTION Two twin residence buildings (Fig. 1.2), the one (Fig. 1.3,1.4,1.5) on fixed base and the other one (Fig. 1.3,1.6,1.7) on isolated base, have been designed and realised in Rapolla (Potenza Italy - Fig.1.1). The construction of the twin buildings started in 1996 in high seismicity area (1st Italian Seismic Class) on high stiffness soil. The isolated building (building "A") was designed to use steel rubber bearings as working isolation system but a special mixed bearing, capable of using alternatively a slide or rubber bearing, was mounted for testing. In September 2000 at the end of the structural construction, in order to investigate the behaviour of the isolated building with only elastomeric bearings or with joined system (slide bearings and elastomeric bearings), a series of dynamic Snap Back tests was carried out. In the paper the main design structural characteristics of building and isolation systems, tests on devices and structures and relevant results will be discussed.

1

Fig.1.1 Location of Rapolla The building was designed in December 1995, and is composed of three apartment floors and cover, its rectangular plan shape (22,6 m x13,3m) is also regular along its height; the interstory height is 3,05m. Most of the inner beams are in-slab beams with 25 cm depth while the others inner beams, at least where it is possible, and the peripheral beams have 50÷60 cm depth. The columns are 28 with constant section along the height, 10 of the inner columns have 40x60 cm rectangular section while the last 18 peripheral columns have a 40x40 cm squared section. The seismic isolation has been adopted to one of the two buildings (building "A"), while the second (building "B") has a conventional R/C structure. The isolation system is placed between the foundations and the upper R/C structure. It is made of 28 HDRB having design damping greater than 10% of critical damping. In order to create a gap for the inspection and the eventual replacement of each bearing, the foundations of the isolated building are surrounded by reinforced concrete soil-retaining walls separate from the structure.

2

A

B

Fig.1.2: The twin buildings (Fix Base "B" and Base Isolated "A")

Fig.1.3: Layout of typical floor

3

Fig.1.4 Transverse section of fix base structure "B"

Fig.1.5 Fix base structural foundation "B"

4

Fig.1.6 Transverse section of base isolated structure "A"

Fig 1.7 Base isolated structural foundation "A"

5

2. DESIGN DATA Seismic response has been calculated with the response spectrum method, using a spatial FEM model (Fig. 2.1) with frame elements for beams and columns and shell elements for each floor and infill panels.

Fig 2.1 FEM model

Design values of periods and participating factors for the twin structures are presented as follows. FXED BASE BUILDING MODAL FREQUENCIES AND PERIODS MODE EIGENVALUE FREQUENCY NUMBER (RAD/SEC)**2 (RAD/SEC) 1 .166138E+03 .128895E+02 2 .254558E+03 .159549E+02 3 .260163E+03 .161296E+02 4 .197192E+04 .444063E+02 5 .296503E+04 .544521E+02 6 .297727E+04 .545644E+02 MASS PARTICIPATING FACTORS - (%) MODE DIR-X DIR-Y DIR-Z 1 87.091 .000 00.000 2 .003 .000 00.000 3 .000 84.596 00.000 4 10.426 .000 00.000 5 .010 .000 00.000 6 .000 12.362 00.000

FREQUENCY (CYCLES/SEC) 2.051420 2.539295 2.567100 7.067484 8.666317 8.684193 X-SUM 87.091 87.094 87.094 97.520 97.530 97.530

6

PERIOD (SEC) .487467 .393810 .389545 .141493 .115389 .115152

Y-SUM .000 .000 84.596 84.596 84.596 96.959

Z-SUM 00.000 00.000 00.000 00.000 00.000 00.000

BASE ISOLATED BUILDING MODAL FREQUENCIES AND PERIODS MODE EIGENVALUE FREQUENCY NUMBER (RAD/SEC)**2 (RAD/SEC) 1 .100351E+02 .316783E+01 2 .101896E+02 .319212E+01 3 .130578E+02 .361356E+01 4 .491679E+03 .221738E+02 5 .694554E+03 .263544E+02 MASS PARTICIPATING FACTORS - (%) MODE DIR-X DIR-Y DIR-Z 1 99.916 .000 00.000 2 .000 99.961 00.000 3 .000 .000 00.000 4 .082 .000 00.000 5 .000 .038 00.000

FREQUENCY (CYCLES/SEC) .504175 .508042 .575115 3.529077 4.194433 X-SUM 99.916 99.916 99.916 99.998 99.998

PERIOD (SEC) 1.983437 1.968341 1.738782 .283360 .238411 Y-SUM .000 99.961 99.961 99.961 99.999

Z-SUM 00.000 00.000 00.000 00.000 00.000

Seismic effects reduction was obtained by means of damping and of an increased first mode period. Rubber isolators stiffness features can provide the structure with a period of about 2 seconds in the first two translational modes. Respect with the fixed-base structure, which has periods lower than 1 sec, the isolated structure can achieve an acceleration as well as an inertial force reduction factor equal to the spectral values ratio (2.0/0,8)2/3=1,84. Dissipative features of the special rubber mix of the isolators, which can result in a damping ratio of about 10% produce a further decrease in response of about 20-40%. We reach therefore a total reduction of effects that ,using more cautelative values of damping, can be equal to 1,84*1,2=2,2. This big reduction of effects can be useful to have a very cheap and light structure, respect with the fixed base one. However ,according to our goals, reduction of effects has been used only to get a greater ultimate state an damage safety. In matter of fact design seismic actions of building “A” is about equal to that of building “B”. Concrete and reinforcement design is the same for both structures.

3. DESIGN CRITERIA FOR THE ISOLATED STRUCTURE. Isolators have maximum vertical force equal to 1500 KN; design stiffness is presented in tab.3.1. TYPE (VERT.LOAD) MAX kN

TYPE (VERT.LOAD MIN kN

HORIZONTAL STIFFNESS

MINIMUM VERTICAL STIFFNESS

kN/m

kN/m

1500

350

657

1900000

Tab.3.1 isolators stiffness

7

Seismic design has been done according to the Italian Code Elastic Spectrum (D.M. January 24th 1986) for a 1st seismic class (C=0.1g) in which the town of Rapolla is located. The following acceleration spectrum was used: a α = ⋅ C ⋅ R(T ) ⋅ ε ⋅ I ⋅ δ g q where response coefficient = R(T)=1.0 for T≤0.8 sec or = 0.862/T2/3 for T>0.8 sec and α=4. According to the use of the building, an importance coefficient I=1 has been assumed. For the isolated building, in order to take into account the effects of additional energy dissipation, an action reduction coefficient equal to δ=1/1.2=0.83 (10% equivalent damping) has been used. A behaviour factor q=2 was assumed for structural strength design of the isolated building "A" and q=4 for the fixed base building "B". As a result both the buildings "A" and "B" have been equally reinforced. The isolation system was designed to obtain a fundamental period of 2.0 sec, thus, neglecting torsional effects, the maximum displacement can be approximately calculated with a 1-DOF mass model as follows: δ = α∗0.083*981*(0.8/2.0)2/3*(2.0/(2π))2 = 18 cm This value is relative to the design spectrum and is confirmed by the analysis on spatial frame model in which the isolators are modelled as elastic elements with stiffness equal to the design effective stiffness. Structural clearances quite greater than the latter value (about 30 cm) have been used between isolated and non-isolated blocks in order to avoid dangerous impacts. In this way a high security is reached against further displacements between the top floor and the isolated base and against unexpected rotational motions not included in the design due to irregular distribution of variable loads or to unexpected bearing failure.

3.1. Rubber bearing isolation system The isolation system consists of High Dumping Rubber Bearings (HDRB). A special neoprene compound is used to have low stiffness and an equivalent damping coefficient greater then 10%. Table 3.2 shows the dimensions of the 28 cylindrical isolators designed. Isol. Type

Dg

1500

500

Ds

Tg

Ng

Ts

Te

H

480 4 34 136 2 Tab.3.2 – Isolator dimensions.

20

242

Legend: Dg = rubber layers diameter (mm) Ds = steel layers diameter (mm) Tg = rubber layer thickness (mm) Ng = number of rubber layers

Hg = total rubber height (mm) Ts = steel sheets thickness (mm) Te = external plates thickness (mm) H = total isolator height (mm)

8

Hg

The design of a suitable number of metallic insertions allows to get the required ratio (about 3000) of vertical over horizontal stiffness. The total height of the isolator, included the upper and lover plates which lock the bearing to the structure, is equal to 312mm. The design of the isolators agrees with the CNR_UNI 10018. The maximum shear strain of the rubber, corresponding to the maximum displacement induced by the earthquake (about 18 cm), is equal to the 134%. The distribution of the isolators between the two grids of beams can allow the access for inspection and replacement at any time. The upper beam grid was designed and verified in order to replace the isolators using two jack each at 1m from the axis of the isolator.

3.2. Slide Bearings Isolation System Seismic passive isolation techniques realised by means of mixed-devices employing both friction sliders and rubber bearings have been adopted in several applications. First used only in bridge structures to absorb thermal effects, purely frictional systems have been introduced in seismic isolation during the eighties in various kinds of devices. It is possible to divide the applications into two main classes: the first one includes in-series coupled devices, the second one includes in-parallel coupled devices. An in-series isolation system standardised for the design of nuclear power plants in high seismicity zones was developed by Electricitè de France (Gueraud et al. 1985) and employed in some nuclear plants in France, South Africa and Iran. Framatome employed this system in a two-unit plant in Koeberg, South Africa, a similar system was employed in the Karun River plant in Iran; Only designed are also a plant in Laguna Verde Mexico, and a plant at Le Carnet in western France, (Tajirian 1998). An in-parallel application named ‘Alexisismon’ was patented by Ikonomou in the U.S. and other seismic countries (U.S. patent no. 4,554,767, november 26, 1985, Ikonomou). Other applications are the following: • Mackay School of Mines, Reno, Nevada, retrofit 1993, isolated using high-damping rubber bearings and sliding bearings, the building houses the new university library. • Long Beach V.A. Hospital, California, retrofit 1995, 12 story RC, isolation system 110 lead-rubber bearings 18 normal-rubber bearings 36 sliding bearings. • Cathedral of Our Lady of the Angels , Los Angeles, California, upon completion in 2002, isolation system 200 high damping rubber bearings + sliding bearings. • Public Safety building, Berkeley, California, 2000, isolation system: 27 lead-rubber bearings + 5 sliders. It is the first new concrete shear wall base isolated building to be designed and built in the U.S. • National Museum of New Zealand, Wellingon, New Zealand, 1995, 23m high 5-story 35000 m2, isolation system 145 lead-rubber bearings, 230 normal-rubber bearings, 42 sliding bearings. • Los Angeles City Hall, Los Angeles, California, retrofit 1999, 28-story, isolation system: 475 HDRB + 58 sliding bearings + 58 viscous base-dampers + 12 viscous dampers in the superstructure. • New Los Angeles Country Martin Luther King Hospital, Charles R. Drew Trauma Center, construction in progress, 70 HDRB + 12 sliders • Villa Fiorita, Ancona, Italy, (designed), isolation system high damping rubber bearings and sliders.

9

The main feature of the mixed isolation technique is the decoupling of stiffness and damping, this is impossible when using HDRB only. Friction interfaces can provide reliable wind restraints, energy dissipation and control of displacements together with vertical loads support, while rubber bearings can give restoring effects as well as carry vertical loads. Economic convenience can play an important key role in the case of big constructions, for the low costs of friction sliders in comparison with those required by rubber bearings and easy installation in retrofit. Rubber Bearings are not able to reach a good compromise between required horizontal stiffness and vertical stability when, as in the case, the buildings is low rised. In rubber bearings, low stiffness induces buckling (Aiken et Al. 1989, Buckle et Al. 1994, Braga et Al. 2000) if the cross section is reduced. On the other hand rubber with a low elastic shear modulus produces vulcanisation problems and compounds ageing instability. Each of the 28 bearings, used in base isolation system of building "A", is a package which includes a slider mounted on the top of the HDRB. The device can work alternatively as a rubber isolator or a slider by simply locking the upper part or the lower part respectively (Fig. 3.1 e 3.2).

Fig.3.1: Special Mixed bearing (HDRB + Slider)

10

Fig.3.2 Isolated building Foundation and special bearings

11

4. BEARINGS MECHANICAL TESTS Qualification and acceptance tests on 8 full scale elastomeric bearings and 2 slide-rubber bearings were carried out at the structural laboratory of the University of Basilicata. The tests were made to verify the behaviour, such as stability (Figg. 4.1, 4.5, 4.6) and creep (Fig. 4.2), of new and accelerated ageing (21 days at 70°C) devices under the working conditions and the agreement with the designed mechanical characteristics such as vertical and horizontal stiffness and energy dissipation for rubber bearings (Fig.4.1, 4.3, 4.4, 4.5, 4.6, 4.8, 4.9) and friction for sliders (Fig. 4.7). 2500

Compression Load (kN)

2000

1500

1000

500

0 0

0.2

0.4

0.6

0.8 1 Displacement (mm)

1.2

1.4

1.6

Fig.4.1: Compression tests on HDRB

1.4 Compression Load N=1500 kN 1.2

Displacement (mm)

1 0.8 0.6 0.4 0.2 0 0

1

2

3

4

5

6

7

8

9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 Time (h)

Fig.4.2: Creep tests on HDRB

12

200

150

Compression Load N=75 kN di=36, 90, 180, 270 mm

Shear Load (kN)

100

-300

50

0 -200

-100

0

100

200

300

100

200

300

100

200

300

-50

-100

-150

-200

Displacement (mm) 200

150

Compression Load N=750 kN di=36, 90, 180, 270 mm

Shear Load (kN)

100

-300

50

0 -200

-100

0 -50

-100

-150

-200

Displacement (mm) 200

150

Compression Load N=1500 kN di=36, 90, 180, 270 mm

Shear Load (kN)

100

-300

50

0 -200

-100

0 -50

-100

-150

-200

Displacement (mm)

Fig.4.3: Shear test on new HDRB with different compression loads

13

200

150

Compression Load N=75 kN di=36, 90, 180, 270 mm

Shear Load (kN)

100

-300

50

0 -200

-100

0

100

200

300

100

200

300

100

200

300

-50

-100

-150

-200

Displacement (mm) 200

150

Compression Load N=750 kN di=36, 90, 180, 270 mm

Shear Load (kN)

100

-300

50

0 -200

-100

0 -50

-100

-150

-200

Displacement (mm) 200

150

Compression Load N=1500 kN di=36, 90, 180, 270 mm

Shear Load (kN)

100

-300

50

0 -200

-100

0 -50

-100

-150

-200

Displacement (mm)

Fig.4.4: Shear test on aged HDRB with different compression loads

14

250

Compression Load N=2250 kN

200

Shear Load (kN)

150

100

50

0 -50

0

50

100

150

200

250

300

350

400

450

350

400

450

-50 Displacement (mm)

Fig.4.5: Shear failure test on new HDRB

250

Compression Load N=2250 kN

200

Shear Load (kN)

150

100

50

0 -50

0

50

100

150

200

250

300

-50 Displacement (mm)

Fig.4.6: Shear failure test on aged HDRB

15

200

150

Compression Load N=250 kN

Shear Load (kN)

100

-300

50

0 -200

-100

0

100

200

300

100

200

300

100

200

300

-50

-100

-150

-200

Displacement (mm) 200

150

Compression Load N=500 kN

Shear Load (kN)

100

-300

50

0 -200

-100

0 -50

-100

-150

-200

Displacement (mm) 200

150

Compression Load N=750 kN

Shear Load (kN)

100

-300

50

0 -200

-100

0 -50

-100

-150

-200

Displacement (mm)

Fig.4.7: Shear test on sliders with different compression loads

16

1.20 Compression Load N=75,750,1500 kN

Lab Tests on new devices Lab Tests on aged devices

Shear Stiffness (kN/mm)

1.00

0.80

0.60

0.40

0.20

0.00 0%

20%

40%

60%

80%

100%

120%

140%

160%

180%

200%

Shear rubber strain γ

Fig.4.8: Shear Secant Stiffness of HDRB during static Lab tests 30% Lab Tests on new devices

Compression Load N=75,750,1500 kN

25%

equivalent dumping ξ

Lab Tests on aged devices 20%

15%

10%

5%

0% 0%

20%

40%

60%

80%

100%

120%

140%

160%

180%

200%

Shear rubber strain γ

Fig.4.9: Equivalent dumping of HDRB during static Lab tests

5. DYNAMIC SNAP-BACK TESTS AND FREE VIBRATIONS To compare different type of isolation systems, about 20 tests of dynamic “snap-back” were carried out on the isolated building "A" at Rapolla. The same kind of test, in Italy, was already carried out on the TELECOM building isolated on Rubber-Bearings in Ancona using a set of hydraulic jacks to push the structure (10 cm) and dynamite to quickly release (Bettinali et al. 1991, Forni et Al. 1991). During the Rapolla's tests the initial displacement was applied by moving the structure just above the isolation system using the same mechanical device employed during the release

17

tests (≈2 cm of initial displacement) on one of the isolated buildings of the University of Basilicata at Potenza - Italy (Braga et Al. 2001). The release device (Fig. 5.1-5.3) was redesigned to apply an initial displacement ranging from 0 to 20 cm and thrust against a 200 tons reaction wall placed on 5 piles (150 cm diameter and 15 m length) of foundation designed at this aim (Fig.5.4). The release device is completely reusable and allows the tests to be repeated in a short time. It consists of a three hinged arch (two sets of trusses and three cylindrical shafts) vertically pulled on the middle hinge by means of an hydraulic jack equipped with a load cell. The release takes place when the three hinges reach the horizontal unstable alignment and the trusses detach from the middle hinge (central shaft).

Fig.5.1. Release device

Fig.5.2. Plane view and transverse sections of release device

18

T = Fv ⋅ cotang α T = Fv ⋅ cotang α T = Fv ⋅ cotang α

α Fh

Fv Fh =

Fv cotang α 2

Fig. 5.3. Push mechanism of release device

Fig.5.4: Reaction Wall

19

The device allows to perform in-situ cyclic tests by simply controlling the vertical displacement of the central hinge. Figures 5.5 and 5.6 show cyclic test performed before the release test n.4. 12 HDRB + 16 Sliders - Displacement time history before the release

Applied horizontal displacement (mm)

160 140 120 100 80 60 40 20 0 0

200

400

600

800

1000

1200

1400

1600

Time (sec)

Fig.5.5 Cyclic displacement history applied to the structure before releasing (test n.4)

12 HDRB + 16 Sliders - Hysteresis Loops 1800

Horizontal Force (KN)

1600 1400 1200 1000 800 600 400 200 0 0

20

40

60

80

100

120

140

160

Displacements (mm)

Fig.5.6 Hysteresis loops of 12 HDRB in-parallel with 16 Sliders due to cyclic displacement history applied to the structure before releasing (test n.4)

20

The maximum horizontal force achieved just before the release tests was 3000 kN while maximum vertical load registered by the jack during the load path was 1000 kN (Fig.5.7). Figure 5.8 compares the horizontal loads applied before the release tests n.6 and horizontal forces applied in the lab tests on the devices.

2500

Fh 2000

1500

Fv

1000

500

0 0

20

40

60 80 100 Applied Horizontal Displacement (mm)

120

140

160

Fig.5.7: Applied Loads before the release by test devices (Test n.2) 200

Lab Test Snap Back Test

150

100

Shear Load (kN)

Applied Vertical and Horizontal Loads (kN)

3000

50

0

-200

-150

-100

-50

0

50

100

150

-50

-100

-150

-200

Displacement (mm)

Fig.5.8: Applied Shear Loads, before the release, to single HDRB (Test n.6)

21

200

In table 5.1 the main tests are listed in terms of effective initial displacement equal to the maximum applied displacement minus residual. Isolation system indicates the plane pattern of sliders and HDRB (Fig. 5.9) in witch circle indicate unlocked sliders (HDRB locked) in tests n. 3-4-5.

N. Test 1 2 3 4 5 6

Date 27.07.2000 28.07.2000 29.08.2000 30.08.2000 30.08.2000 01.09.2000

Tab. 5.1 Main tests Isolation system 28 HDRB 28 HDRB 12 HDRB + 16 Sliders 12 HDRB + 16 Sliders 12 HDRB + 16 Sliders 28 HDRB

Effective Initial disp. (mm) 87 136 150 150 150 150

Fig.5.9 Mixed System base pattern During the free vibrations acceleration and displacements have been recorded by using the following instruments 1. 0.02 mg accuracy and ±0.1 g measuring range accelerometers 2. 0.2 mg accuracy and ±1.0g measuring range accelerometers 3. 0.01 mm accuracy and ±250 mm measuring range linear transducers Three horizontal accelerometers per level were located at the position shown by figure 5.10 for both base floor and top-floor. Three more accelerometers were placed on the foundation

22

frame just underneath the isolation level. Vertical accelerations were recorded by two accelerometers placed on the peripheral columns of the central frame at the top floor. Accelerometers 1 and 3, are used to record accelerations in direction of the short plane dimension of the building as well as rotational accelerations; once the rotational accelerations are known, also the accelerations directed as the long side of the building are obtained. Instruments 4 and 5 can also detect 'rocking' effects by measuring vertical accelerations. Two more accelerometers were also placed on the upper corner of the reaction wall.

3 B u ildig "A " 1 2 Fh

5 4

Reaction W a ll

Fig.5.10 Plane pattern of the accelerometers

Acceleration

Fast Fourier Transform is performed scanning acceleration data with the windowing technique (Onsay T. et Al. 1993) in order to investigate non-linearity during the free vibration of structure and isolation systems (Fig. 5.11).

i - FFT time

i-mode Amplitude / 1 st mode Amplitude .

1 0.9

T1 in = 1.47 sec

Fast Fourier Transform- Windowing Technique Test n.6 (28 HDRB)

0.8 0.7 0.6

∆twindow = 0.5 sec

0.5 T1 fin = 1.27 sec

0.4 0.3

T3 = 0.03 sec

T2 = 0.16 sec

0.2 0.1 0 0.01

0.1

1 Period (sec)

window

Fig. 5.11 Windowed Time Fourier Transform Six displacement transducers were located at the position shown by figure 5.12. Transducers 1 and 2 measure the displacements of the base floor relative to the ground, 3 and 4 measure the displacements of the base floor relative to Reaction Wall, 5 and 6 are displacements of the

23

Reaction Wall relative to the ground. One more linear transducer records the vertical displacement of the central hinge of the release device during the push phase.

B u ilding "A "

2 6

4

F

3

1 5

Reaction W a ll Fig.5.12 Plane pattern of the displacement transducers

6. MAIN RESULTS Free vibration tests carried out on the system relevant to maximum displacement impressed (dmax=150 mm) and comparison between the behaviour of the mixed system (HDRB + Sliders) and the simple system (HDRB only) are shown. 6.1. Displacements A comparison between the base displacements histories recorded during the main tests relative to the two isolation systems is shown in figure 6.1. Free vibration of the structure with the two systems of isolation 200 Test n.6 (28 HDRB) 150

Test n.3 (12HDRB+16 Sliders)

Displacements (mm)

Test n.4 (12HDRB+16 Sliders) 100

Test n.5 (12HDRB+16 Sliders)

50

0

-50

-100 0

1

2

3

4

5

6

Time (sec)

Figure 6.1. Free vibrations tests, comparison between the base displacements of the two systems

24

Equivalent damping ratio has been roughly calculated considering the first peaks, since the following peak values are not significant, being these extremely close to the residual displacement. Equivalent damping has been calculated on the differences between two consecutive peaks, by using the relation: ξ eq =

1  v(t )   ln 2π  v(t + T ) 

The tests on the mixed system gave an equivalent damping ratio ξ of about 30% for the first and second couple of peaks. The HDRB system made more cycles than the Mixed system, allowing a more accurate evaluation of ξ, as reported by figure 6.2. 35%

Equivalent dumping ξ

30%

Dynamic Snap Back test n.4 (HDRB+Sliders)

25% 20%

Dynamic Snap Back Tests (HDRB Only)

15% 10%

Static Lab Tests on new HDRB Static Lab Tests on aged HDRB

5% 0% 0%

20%

40%

60%

80%

100%

120%

140%

160%

180%

200%

Rubber Shear Strain γ

Figure 6.2 Experimental evaluation of equivalent damping ξ Figure 6.2 outlines how HDRB exhibits higher energy dissipation when the strain rate increase as in the case of a dynamic test.

6.2. Accelerations In figures 6.3 and 6.4 are represented base and top-floor accelerations of the HDRB system in the test n.6 and mixed isolated structure in the test n.4 respectively.

25

500 400

Test n.6 (28 HDRB)

300 200 Acceleration (mg)

100 0 -100 -200 -300 Base Acceleration (Experimental)

-400 -500

-536

Top Floor Acceleration (Experimental)

-600 -700 -800 -880

-900 0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

5

Time (sec)

Figure 6.3 Floor accelerations history during free vibrations (Test n.6)

200

Test n.4 (12 HDRB + 16 Sliders)

Acceleration (mg)

100

0

-100 Base Acceleration (Experimental) -200 Top Floor Acceleration (Experimental)

-231 -272 -300 0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

5

Time (sec)

Figure 6.4 Floor accelerations history during free vibrations (Test n.4)

Comparison between the HDRB system (28 HDRB, test n.6) and the Mixed system (12 HDRB+16 Sliders, test n.4) reveals greater values of the accelerations for the first one.

26

6.3. Vibration frequencies The first three frequencies (Fig. 6.5, 6.6) of the structure have been obtained by using the Fast Fourier Transform (FFT). It is clearly shown that the instantaneous release excites in the structure three different vibration frequencies. The first of these is due to the isolated mode, for which the structure moves as a 1-DOF system, while the other two are related to the structural modes excited.

i-mode Amplitude / 1 st mode Amplitude .

1

T1 in = 1.47 sec

Fast Fourier Transform - Windowing Technique

0.9

Test n.6 (28 HDRB)

0.8 0.7

∆twindow = 0.5 sec

0.6 0.5

T1 fin = 1.27 sec

0.4 T3 = 0.03 sec

0.3

T2 = 0.16 sec

0.2 0.1 0 0.01

0.1

1

10

Period (sec)

Figure 6.5 Normalised FFT of base accelerations (Test n.6) 1

T 1 in = 2.05 sec

Fast Fourier Transform - Windowing Technique Test n.4 (12 HDRB + 16 Sliders)

0.8 0.7 0.6

∆t window = 0.5 sec

i-mode Amplitude / 1

st

mode Amplitude

0.9

0.5 T 2 = 0.16 sec

0.4

T 1 fin = 1.7 sec

0.3 0.2

T 3 = 0.03 sec

0.1 0 0.01

0.1

Period (sec)

1

Figure 6.6 Normalised FFT of base accelerations (Test n.4)

27

10

First mode period in the test direction (Fig. 6.7-1) produced by release is about 1.47 sec for the HDRB system and 2.05 sec for the Mixed system. Design period (1.97 sec - HDRB only) is greater than experimental period, because the masses in the release test were 80% of the the design masses. Concerning vibration periods, since the actual stiffness of the isolation system is about the same as it is assumed in the design, the ratio Td/Te, where Td=design period and Te=experimental period, has to be as follows:

Td Md = = 1,34 Te Me

Second and third mode periods equal to 0.16 sec and 0.03 sec, shown by the FFT, are relative to 2nd isolated mode in the direction of the test (Fig. 6.7-2) and to vibration mode in the plane of the 1st isolated floor (Fig. 6.7-3) respectively

1) 1st Isolated Mode (T = 1.47 sec)

2) 2nd Isolated Mode (T = 0.16 sec)

3) Mode of 1st Floor (T = 0.035 sec)

Figure 6.7 Structural modes excited by release

The FFT carried on the top-floor acceleration history reveal for the Mixed system an isolated period 1,7 times greater than that of the HDRB system (T1Mix=2,05sec T1HDRB =1.47sec), while the values of the higher modes periods remain almost the same (T2=0,16 sec). FFT shows also an increase in the relative importance of higher mode contribution respect to first mode, in fact the rate of second-to-first mode contribution in terms of top-floor acceleration passes from 0.29 (HDRB) to 0.37 (Mixed). This can be due to the non-linearity introduced by sliding devices, that, while providing the structure with higher dissipation that results in a generally decreased response, also amplified superstructure modal response as shown by relative top-floor-base accelerations time history (Fig.6.8, 6.9).

28

800 Relative Accelerations (Top floor-Base) Test n.6 (28 HDRB)

600

Acceleration (mg)

400 200 0 -200 -400 -600 -800 0.0

1.0

2.0

Time (sec)

3.0

4.0

5.0

Figure 6.8 Relative Top floor-base accelerations (Test n.6)

300

Relative Accelerations (Top floor-Base) Test n.4 (12 HDRB + 16 Sliders)

Acceleration (mg)

200

100

0

-100

-200

-300 0.0

1.0

2.0

Time (sec)

3.0

4.0

5.0

Figure 6.9 Relative Top floor-base accelerations (Test n.4)

FFT carried on base acceleration history, reveals the presence of a period of about 0,035 sec, due to free planar vibration of the base floor, caused by the impulse given to the base by the release device. This mode does not influence top-floor vibrations as shown in figure 6.10. FFT also reveals that, for all the modes excited, the energy in the Mixed system is quite below the energy in the HDRB system.

29

70000 T 1 HRDB = 1.47 sec

60000

Fast Fourier Transform Comparison Test 4 (12 HDRB+16 Sliders) Test 6 (28 HDRB)

Amplitude

50000 40000 30000

T 2 = 0.16 sec

20000

T 1 Mixed = 2.05 sec

10000 0 0.01

0.1

1

10

Period (sec)

Figure 6.10 Fast Fourier Transform - Top-floor acceleration

7. CONCLUSIONS Dynamic release tests carried out produced a huge number of acceleration and displacement time histories. Analysis of acquired experimental data are now in progress, but first results confirm good behaviour of Mixed system and its capability to solve some design problems due to HDRB systems. The registered data are giving many suggestion for the design of Mixed Isolation systems not taken into account by several codes as in the case of the Italian Guidelines for Seismic Isolation. In the future, the two equally reinforced buildings, traditional fixed base structure "B" and base isolated structure "A", could provide interesting details on their seismic behaviour.

ACKNOWLEGMENTS The assistance of Marco Faggella, undergraduate engineering student, for the tests data processing is gratefully acknowledged.

REFERENCES Guèraud R., Noel-Leroux J.-P., Livolant M., Michalopoulos A.P. 1985. Seismic Isolation Using Sliding-elastomer Bearing Pads. Nuclear Engineering and Design 84 pp. 363-377

30

Tajirian F.F. 1998. Base Isolation Design for Civil Components and Civil Structures. Proceedings, Structural Engineers World Congress, San Francisco, California, July 1998. Ikonomou A.S. 1985. Alexisismon Isolation for Nuclear Power Plants. Nuclear Engineering and Design 85 pp. 201-216. Aiken I. D., Kelly J. M., Tajirian F. F. 1989. Mechanics of Low Shape Factor Elastomeric Seismic Isolation Bearings, Report No. UCB/EERC-89/13 1989. Buckle I.G., Liu H. 1994. Experimental Determination of Critical Loads of Elastomeric Isolators at High Shear Strain, NCEER Bulletin, Vol. 8, N. 3, July 1994. Braga F., Laterza M., Masi A. 2000 (in Italian). Valutazione numerica del carico critico di isolatori in elastomero armato ”; Ingegneria Sismica Anno XVII – N.1/2000 – pagg. 17-26. Braga F., Laterza M., Masi A. & Nigro D. 2000 (in Italian). “ Valutazione sperimentale del carico critico di isolatori cilindrici in elastomero armato ”; Ingegneria Sismica Anno XVII – N.1/2000 – pagg. 27-39. Bettinali F., Forni M., Indirli M., Martelli A., Masoni P., Bonacina G., Pucci G., Serino G., Venturuzzo, M. & Giuliani G.C. 1991. On site dynamic tests of a large seismically isolated building, International Meeting on Earthquake Protection of Buildings. Ancona Forni M., Martelli A., Spadoni B., Casalini E., Bonacina G., Pucci G. & Serino G. 1991. Dynamic tests on seismically isolated structure mock – ups and validation of numerical models. International Meet-ing on Earthquake Protection of Buildings. Ancona. Braga F., Nicoletti M., Bixio A. R., Dolce M., Nigro D. & Ponzo F. C. 2001. Repeatable Dynamic Release Tests on a Base-Isolated Building. Journal of Earthquake Engineering, Vol. 5, No. 3 (2001). Imperial College Press. Onsay T. & Haddow A.G.1993. Comparison of STFT, Gabor, and Wavelet transforms in transient vibration analysis of mechanical systems, ASA 125th Meeting. Ottawa.

31