Isolation from seismic motion

The principle of isolation is simply to provide a discontinuity between two bodies in contact so that the motion of either body, in the direction of the discontinuity, cannot be fully transmitted. The discontinuity consists of a layer between the bodies which has low resistance to shear compared with the bodies themselves. Such discontinuities may be used for isolation from horizontal seismic motions of whole structures, parts of structures, or items of equipment mounted on structures. Because they are generally located at or near the base of the item concerned, such systems are often referred to as base isolation (Figure 8.13), although the generic term seismic isolation is preferable.

The layer providing the discontinuity may take various forms, ranging from very thin sliding surfaces (e.g. PTFE bearings), through rubber bearings a few centimetres thick, to flexible or lifting structural members of any height. To control the seismic deformations which occur at the discontinuity, and to provide a reasonable minimum level of damping to the structure as a whole, the discontinuity must be associated with energy-dissipating devices. The latter are also usually used for providing the required rigidity under serviceability loads, such as wind or minor earthquakes. Because substantial vertical stiffness is generally required for gravity loads, seismic isolation is only appropriate for horizontal motions.

The soft layer providing discontinuity against horizontal motions cannot completely isolate a structure. Its effect is to increase the natural periods of vibration of the structure, and to be effective the periods must be shifted so as to reduce substantially the response of the structure. For example, a three-storey building might typically have its fundamental period shifted from 0.3 s to 2.0 s by being changed from fixed base to isolated. If the structure were located on a rock site, and had a design response spectrum as for rock in Figure 3.3(a), this would reduce the elastic response of the structure by a factor of about 10. However, a similar period shift for a structure on soft soil might not achieve a worthwhile reduction in response, or could even result in an increased response, as may be inferred from the spectra for softer soil sites in Figure 3.3. Clearly, the shape of the design spectrum, the fixed base period and the period shift are the three factors which determine whether base isolation has any force-reducing effect (or, indeed, the opposite!).

The location of the isolating devices should obviously be as low as possible to protect as much of the structure as possible. However, cost and practical considerations influence the choice of location. On bridges it is generally appropriate to isolate only the deck where isolation from thermal movements is required anyway. In buildings the choice may lie between isolating at ground level, or below the basement, or at some point up a column. Each of these locations has its advantages and disadvantages relating to accessibility and to the very important design considerations of dealing with the effects of the shear displacements on building services, partitions, and cladding, as described in Figure 8.13, which was derived from Mayes et al. (1984).

Bearings located at bottom of first-storey columns


• Minimal added structural costs

• Separation at level of base isolation is simple to incorporate

• Base of columns may be connected by diaphragm

• Easy to incorporate back-up system for vertical loads


• May require cantilever elevator pit

Bearings located at top of basement columns


• No sub-basement requirement

• Minimal added structural costs

• Base of columns connected by a diaphragm at isolation level

• Backup system for vertical loads provided by columns


• May require cantilevered elevator shaft below first floor level

• Special treatment required for internal stairways below first floor level

Bearings located at mid-height of basement columns


• No sub-basement required

• Basement columns flex in double curvature and therefore may not be required to be as stiff as for bearings located at top or bottom


• Special consideration required for elevators and stairways to accommodate displacements at mid-storey

• No diaphragm provided at isolation level

• Difficult to incorporate back-up system for vertical loads

Bearings located at sub-basement


• No special detailing required for separation of internal services such as elevators and stairways

• No special cladding separation details

• Base of columns connected by diaphragm at isolation level

• Simple to incorporate back-up system for vertical loads


• Added structural costs unless sub-basement required for other purposes

• Requires a separate (independent) retaining wall

Figure 8.13 Different locations for base isolation of buildings (from Mayes et al., 1984). (Reproduced by permission of the Earthquake Engineering Research Institute)

8.5.3 Seismic isolation using flexible bearings

The most commonly used method of introducing the added flexibility for seismic isolation is to seat the item concerned on either rubber or sliding bearings. The energy dissipa-tors (dampers) that must be provided may come in various forms. For use with standard bridge-type bearings made of rubber or sliding plates, any of the energy dissipators mentioned in Section 8.5.6 may be suitable. In addition, all-in-one devices, incorporating both isolation and damping, are used, namely lead-rubber and high damping rubber bearings.

The most effective device, the lead-rubber bearing (Robinson and Tucker, 1977), is conceptually and practically very attractive for seismic isolation, as it combines all of the required design features of flexibility and deflection control into a single component. As shown in Figure 8.14, it is similar to the laminated steel and rubber bearings used for temperature effects on bridges, but with the addition of a lead plug energy dissipator. Under cyclic shear loading the lead plug causes the bearing to have high hysteretic damping behaviour, of almost pure bilinear form (Figure 8.15). The high initial stiffness is likely to satisfy the deflection criteria for serviceability limit-state loadings, while the low post-elastic stiffness gives the potential for a large increase in period of vibration desired for the ultimate limit state design earthquake.

As shown by Tyler and Robinson (1984), hysteretic behaviour is very stable under increasing cyclic displacements. In dynamic tests on bearings 280 x 230 x 113 mm in size, shear displacements of up to ±140 mm at frequencies of 0.1-0.3 Hz were applied,

Figure 8.14 Construction of a patented lead-rubber bearing (after Robinson and Tucker, 1977)

Rubber Isolation Pads Seismic




Earthquake Isolators Deflection

Shear strain g

Figure 8.15 Typical hysteretic behaviour of a lead-rubber bearing (after Robinson and Tucker, 1977)

Shear strain g

Figure 8.15 Typical hysteretic behaviour of a lead-rubber bearing (after Robinson and Tucker, 1977)

giving shear strains in the rubber of up to ±200%. The weight of the structure on the bearings ranged from 35 to 455 kN. It was concluded that with peak strains in the rubber in excess of 100%, the bearings would continue to function satisfactorily for a sequence of very large earthquakes.

The first building in the world to be built using lead-rubber bearings for seismic isolation was the William Clayton Building (Megget, 1978) in Wellington, New Zealand, designed c. 1978. It has a four-storey ductile moment resisting frame, a section through the building being shown in Figure 8.16. The inter-storey drifts calculated for the isolated building were about 10 mm, and were uniform over the building's height. For comparison, the maximum drift for the non-isolated model was 52 mm per storey for the top two storeys. An overall deflection ductility factor of only x = 1.6 (x is defined in Section 5.4.7) was required for the isolated building, whereas x = 7.6 would have been required for the non-isolated condition. A more recent example of a structure protected by lead-rubber bearings is discussed in Section 11.3.3.

Lead-rubber bearings have also been used in a rapidly growing number of bridges in New Zealand, Italy, Japan, the USA and elsewhere. These bearings have a wide range of applications where they are likely to lead not only to less damaged structures in earthquakes, but also sometimes to cheaper construction. Design procedures are well established, including methods using design graphs (e.g. Skinner et al., 1993) and design codes (e.g. FEMA, 1997).

Further developments in seismic isolation are going on, as described, for example, by Robinson (2000). One such development is an improvement to both the rubber isolation bearing and to the lead-rubber bearing, consisting of a centre-drive to the top and bottom of the bearings. The 'centre-drive' approach has two advantages: first, it allows a greater displacement to height ratio; and second, it provides increased damping capacity at large displacements, thus providing some of the additional damping needed for resisting 'near-fault fling'.

Lead Extrusion Dampers

17.230 y Foundation pads

Figure 8.16 Section through the William Clayton Building, Wellington, New Zealand, the first building to be seismically isolated. The lead-rubber bearings are shown beneath the basement (after Megget, 1978)

17.230 y Foundation pads

Figure 8.16 Section through the William Clayton Building, Wellington, New Zealand, the first building to be seismically isolated. The lead-rubber bearings are shown beneath the basement (after Megget, 1978)

8.5.4 Seismic isolation using flexible piles and energy dissipators

An interesting alternative to the use of lead-rubber bearings is the isolation system used first for Union House (Boardman et al., 1983), a 12-storey office block in Auckland, New Zealand, completed in 1983 (Figure 8.17). As the building required end-bearing piles about 10 m long, the designers took the opportunity of making the piles flexible and separating them from lateral contact with the soft soil layer overlying bedrock by surrounding them with a hollow sleeve, thus creating the flexibility required for base isolation. Deflection control was imposed by tapered steel energy dissipators (Figure 8.18) located at ground level. The structure was built of reinforced concrete except that the superstructure was diagonally braced with steel tubes. Lateral flexibility of the piles was attained by creating hinges of low moment resistance at the top and bottom of each pile.

The earthquake analysis was carried out using non-linear dynamic analysis. Under the design earthquake loading the horizontal deflection of the first floor relative to the ground (i.e. at the dissipators) was calculated to be ±60 mm. The response of the building was also checked under a 'maximum credible earthquake' to ensure that adequate clearance was provided at the energy dissipators, and that no significant yielding would occur in the superstructure. In this survivability state the horizontal deflection at the dissipators was ±130 mm and a provision for ±150 mm was made. Because of the structural discontinuity at ground level, the lift shaft and the bottom story facade had to be supported from the first floor above ground level.

Another building protected in the same way is the 10-storey Wellington Central Police Station (Charleson et al., 1987), built in the 1980s. The isolation system enables achievement of the design aim of making it fully operational after the expected next magnitude 7.5 earthquake on the Wellington fault located about a kilometre away. Cost and time comparisons of the isolated and non-isolated equivalent structure estimated a capital cost saving of $300,000 and a construction time saving of three months, representing $150,000. Together these equal a substantial saving of nearly 7% in the total construction cost.

8.5.5 Rocking structures

As well as the methods described in the preceding sections, the flexibility required to reduce seismic response may be obtained by allowing part of the structure to lift during large horizontal motions. This mechanism is referred to variously as uplift, rocking, or stepping, and involves a discontinuity of contact between part of the foundations and the soil beneath, or between a vertical member and its base. The good performance of many ordinary structures in very strong ground shaking can only be explained by rocking having beneficially occurred during the earthquake. For example, this appears to have been the case for some pre-code low-rise brittle concrete buildings in the near-fault region of the Hawke's Bay, New Zealand, earthquake of 1931 (van de Vorstenbosch et al., 2002).

As early as the 1970s, various studies (e.g. Meek, 1975) were made of 'natural' rocking systems, i.e. those not incorporating displacement-controlling energy dissipators relying solely on uplifting columns or rocking of raft or local pad foundations to produce the desired effects. However, despite apparently favourable results, such structures have not yet been enthusiastically adopted in practice. This is probably due to continuing design uncertainties regarding factors such as soil behaviour under rocking foundations in the design earthquake, the possible overturning of slender structures in survivability events,

Figure 8.17 Section through Union House, Auckland, New Zealand, showing isolating piles and energy dissipators (after Boardman et al, 1983)
Seismic Isolation Pier Mobile Home

Cross-section through pier at dissipator

Figure 8.18 Cantilever steel plate dampers of the type used in Union House and Dunedin Motorway Bridge (Skinner et al., 1980)

Cross-section through pier at dissipator

Figure 8.18 Cantilever steel plate dampers of the type used in Union House and Dunedin Motorway Bridge (Skinner et al., 1980)

or possible impact effects when the separated interfaces slam back together (Meek, 1975). However, based on field experience like that of the Hawke's Bay earthquake cited above (van de Vorstenbosch et al., 2002), more boldness in allowing rocking of squat structures seems to be justified.

With the addition of energy absorbers the above hazards are lessened, and utilization of the advantageous flexibility of uplift has been put to practical effect in completed constructions, a bridge and chimney being discussed below.

The first such structure to be built was the South Rangitikei Railway Bridge in New Zealand, the design of which was carried out c. 1971. The bridge deck is 320 m long, comprising six prestressed concrete spans, about half of which is at a height of 70 m above the riverbed. The piers consist of hollow reinforced concrete twin shafts 10.7 m apart coupled together with cross beams at three levels, so that they act as a kind of portal frame lateral to the line of the bridge. At their base, these shafts are seated on an elastomeric bearing (Figure 8.19) and lateral rocking of the portals is possible under the control of a steel torsion-beam energy dissipator of the type shown in Figure 8.19. As with other forms of base isolation, substantial reductions in earthquake stresses are possible, as

South Rangitikei Railway Bridge
Figure 8.19 South Rangitikei railway bridge, New Zealand, showing locations of bearings and torsion-beam energy dissipators (Skinner et al., 1980)

described for an early investigation of this bridge by Beck and Skinner (1974), the final configuration differing in detail but not in principle.

A further example of the use of rocking in seismic design is given in Section 8.6.

8.5.6 Energy dissipators for seismically isolated structures

In the preceding sections on isolation methods, we have discussed a number of energy dissipators (dampers) that have been used with seismically isolated structures, namely:

(1) lead plugs, in lead-rubber bearings (Figure 8.14);

(2) tapered steel plate cantilevers (Figure 8.18);

(3) steel torsion-beams (Figure 8.20).

A variety of other devices have been investigated which are also suitable in this situation:

(4) lead extrusion devices;

(5) flexural beam dampers.

A general overview of some of these energy dissipators is given by Skinner et al. (1993) and Hanson and Soong (2001).

Edge Wise Bus Bar

Figure 8.20 A torsion beam hysteretic damper. The arrows show the opposing actions of the structure and its support (Kelly et al., 1972)

Torsional beam

Figure 8.20 A torsion beam hysteretic damper. The arrows show the opposing actions of the structure and its support (Kelly et al., 1972)

8.5.7 Energy dissipators for non-isolated structures

As discussed above, energy-dissipating devices are an essential component of seismic isolation systems, and they also may be used to reduce seismic stresses in non-isolated structures. Various forms of energy-dissipating devices have been developed for such structures, some of which are discussed below. In addition to the discussion below, the reader is referred to the book by Skinner et al. (1993), and one specifically on energy-absorbing devices by Hanson and Soong (2001).

Energy dissipators in diagonal bracing

Diagonal bracings incorporating energy dissipators control the horizontal deflections of the frame and also the locations of damage, thus protecting both the main structure and non-structure. A practical example is provided by a six-storey government office building constructed in Wanganui, New Zealand, in 1980. This building obtains its lateral load resistance from diagonally braced precast concrete cladding panels, thus minimizing the amount of internal structure to ease architectural planning. Each diagonal brace contains a steel insert consisting of a sleeve housing a specially fabricated steel tube 90 mm diameter and 1.4 m long, which was designed to yield axially at a given load level. A movement gap was provided through the surrounding structure, and buckling was prevented by the surrounding sleeve and concrete.

A number of devices to be connected to diagonal steel bracing show high energy-absorbing capabilities, such as the lead extrusion damper (Skinner et al., 1993) also used in base isolated structures (Section 8.5.6). Pall and Marsh (1982) developed friction damped devices to suit both X- and K-bracing (but the reliability of the friction forces needs to be established).

A device developed in the late 1990s (Monti et al., 1998) is a lead-shear damper which, unlike most other seismic dampers, is suitable for damper vibrations from other sources such as wind. For example, one such device has design working displacements in the range of 50 |m to 10 mm. The effectiveness of its hysteretic damping at three levels of displacement is shown in Figure 8.21.

Steel Frame Hysteretic
Figure 8.21 Hysteresis loops of the Penguin Vibration Damper and its bilinear-spring model (shaded) (from Monti et al., 1998): (a) dmax = 0.1 mm; (b) dmax = 1.0 mm; (c) dmax = 4.9 mm

Energy dissipation in bolted beam-column joints

If bolted joint interfaces are designed to permit controlled sliding rotations in earthquake motion, not only is energy dissipated, but also damage is limited or eliminated. The development and behaviour of promising devices of this type are described by Clifton (2005) and Clifton et al. (2007).

Was this article helpful?

0 -1


  • sonja
    How are earthquake dissipaters used in buildings?
    8 years ago
  • Adelbert
    When did japan start investing in seismic dampers?
    11 months ago
  • robert
    What isolation device were used in union house auckland?
    2 months ago

Post a comment