Earthquake Protection Systems_12 docx

8.2.2 Building Response to Ground Motion When a rigid object is shaken, so-called inertia forces act on it which increase according to the acceleration of the object, and to its mass.. b

the point of view of building response, an even more important measure of

fre-quency content is the response spectrum, which is discussed below.

8.2.2 Building Response to Ground Motion

When a rigid object is shaken, so-called inertia forces act on it which increase according to the acceleration of the object, and to its mass If an absolutely rigid

building is firmly tied to the ground, and shakes with the ground, then the inertiaforces are transmitted from the ground into the building: the magnitude of theforce is proportional to the mass of the building and varies with time in the samemanner as the acceleration

This simple model is unfortunately inadequate, however, because no real ing is quite rigid All buildings deform to some extent as they are shaken, andthe deformation of the building substantially alters the force distribution Small,massive buildings are relatively stiff, but as buildings become taller and lighterthey tend to become more flexible When a flexible building is shaken, the forceacting on any part of it is still proportional to the mass and acceleration of thatpart, but the distribution of forces within the building depends on the way thebuilding itself deforms Depending on the mass and flexibility of the building,the accelerations within the building may be greater or less than the groundaccelerations, and thus the forces may also be greater or less than if the build-ing was a rigid body The consequences of this for building design are of greatsignificance

build-The property of a building which principally determines its dynamic response

to earthquake ground motion is its natural frequency Because all buildings are

flexible they will vibrate when jolted, and they will then sway backwards andforwards in a regular way Taller buildings have lower natural frequencies (theysway more slowly) than lower buildings A building 10 storeys high may take

about a second to sway backwards and forwards in one cycle, i.e its natural period is 1 second A building of two storeys will take about one-fifth of a

second: its natural period is 0.2 seconds (A rough guide is that each storey addsabout one-tenth of a second to its natural period.) Three- to five-storey structuresare likely to have a natural period in the order of 0.3 to 0.5 seconds High-riseframe buildings of 10 to 20 storeys have periods of between 1 and 2 seconds.And very high buildings can have period up to 4 seconds or more

If the disturbance is a short one, the swaying will continue after the disturbancehas finished, but it will gradually die away The rate at which the swaying decays

after the end of the disturbance is a measure of the damping in the building’s

structural system

If the building is shaken by regular ground oscillations (like the effect of

a rotating machine), its response will depend on the relationship between thefrequency of these oscillations and the natural frequency of the building Forground motion frequencies much less than that of the building, the building will

simply move with the ground, and deform very little; as the frequency of theground motion increases, so the deformation of the building will increase, andwhen the two frequencies are equal the building deformation will reach a peakwhich may be many times greater than that of the ground The ground motion

and the building are in resonance For frequencies of ground motion still greater

than the natural frequency of the building, the deformation of the building will

be less the further away it is from the resonant frequency The relationship isillustrated diagrammatically in Figure 8.2

When a building is shaken by a real earthquake, which has a ground motionconsisting of a mixture of frequencies all added together, its response will dependboth on the natural frequency of the building and on the frequency content ofthe earthquake A 10-storey building, with a natural frequency of 1.0 cycles persecond, will be particularly affected by the component of the ground motionwith this frequency, but much less by the components with higher and lowerfrequencies

The effect of a particular earthquake ground motion on a range of

build-ings is shown by the response spectrum The response spectrum for a particular

ground motion shows what the maximum response would be to that groundmotion for buildings5 of different natural frequencies Its shape depends on the

Figure 8.2 Diagrammatic representation of the response to a 10-storey building to the frequency of ground motion vibration

5 Or, more strictly, to damped mass–spring systems, which are a useful mathematical idealisation

of building structures.

Figure 8.3 Typical response spectra and the building types they affect

frequency content of the earthquake and on the degree of damping of the ing Figure 8.3 shows some typical examples of response spectra Example A iswhat the response spectrum might look like for a site close to the epicentre of

build-an earthquake on firm soil or rock It has a peak value of around 3 cycles persecond It would therefore be most damaging to low-rise buildings, but less so

to taller structures, which would experience smaller forces Example B shows atypical shape for a site at some distance from the epicentre and on a soft soil,with a peak value at about 1.0 cycles per second This event would be especiallydamaging to the taller structures, but would be felt much less strongly by thelow-rise structures

An example of this second type of behaviour was the 1985 Mexico Cityearthquake which caused ground motion in the lake bed area of the city with

a period strongly concentrated around 2 seconds; the earthquake caused ularly serious damage to recently constructed 10–20-storey apartment blocks,while leaving much of the older, weaker, low-rise masonry much less severelydamaged

partic-The response spectrum is commonly used in building design codes to definethe design earthquake which buildings should be able to resist without damage.Codes are discussed further in Section 8.6

8.3 How Buildings Resist Earthquakes

Many small buildings are so stiff that they can be assumed to be rigid in a firstestimate of earthquake forces If a horizontal shaking occurs, the forces on eachelement of the building can be found by assuming that it is static, but has ahorizontal force acting on it (through its centre of gravity) proportional to theground acceleration and to the mass of the element, but in the opposite direction

This is what is referred to as the inertia force The effect of vertical shaking

is similar The resistance of this stiff building is principally determined by theability of the structure to transmit these large and rapidly varying inertia forces

to the ground without failure

Consider first a single-storey building, consisting of four walls (with windowand door openings) and a flexible roof which sits on two of the walls, but doesnot tie them together, see diagram A in Figure 8.4 The effect of a primarilyvertical ground shaking will be to increase or decrease the vertical forces, but asthe structure is capable of carrying substantial vertical gravitational forces undernormal conditions, it can usually accept extra vertical forces without difficulty.The effect of a horizontal shaking parallel to two of the walls will be to set uphorizontal inertia forces on each wall in proportion to their mass: the forces onthe walls parallel to the direction of the shaking (the in-plane walls) will be alongtheir length, while those on the perpendicular walls (the out-of-plane walls) will

be at right angles to them The force on the roof will also cause an additionalhorizontal force to be transmitted on to whichever wall supports it The principal

effect of out-of-plane forces is to cause the walls to bend (i.e deform out of

Figure 8.4 Response of single-storey masonry building to earthquake ground shaking

their plane), which can cause damage to brittle masonry structures even under

low levels of loading Wall elements tend to be stronger under in-plane forces:

these cause in-plane shear forces which are easier to resist in a solid wall or can

be provided for by bracing or other means

In the same building the effect of a horizontal force in the direction dicular to that just described would be to exchange the responses of the walls,those previously out-of-plane becoming in-plane and vice versa Thus under areal earthquake shaking, with horizontal shaking in all directions, all walls aresubjected to both out-of-plane bending and in-plane shear simultaneously Thistype of building tends to have little resistance to earthquake forces

perpen-If instead the roof is constructed in such a way as to tie the tops of the walls

together as a rigid diaphragm, the behaviour will be different, as in diagram B in

Figure 8.4 The unresisted out-of-plane bending of diagram A will be prevented,

as the out-of-plane wall will be connected to the roof diaphragm member, which

is then able to transfer the forces involved to the tops of the stiffer in-plane walls,and then to the ground In addition the continuity of the roof will also tie thecorners together, inhibiting corner cracking Under shaking in the other plane,the behaviour is the same in reverse

Thus these elements – the stiff vertical shear wall in each direction, to carry the loads to the ground, and the stiff horizontal diaphragm, to transfer the earth-

quake forces at this level to the appropriate wall – form the basis of an effectiveearthquake-resistant structural system The same system can be used as effectively

in multi-storey construction, in which case the horizontal loads to be transmitted

by the shear walls increase (as do the vertical gravitational loads) from top tobottom of the building, so that the ground floor walls are required to transmit tothe ground the horizontal forces acting on the whole building

However, the use of extensive shear walls can often create serious limitations

on the planning of a building, and the equivalent shear strength can also, in some

cases, be achieved by means of alternative vertical elements such as braced frames and moment-resisting frames (Figure 8.5).

In the braced frame, the bracing members transmit the horizontal forces in

tension and compression; such frames can be very stiff but are often appropriate

only on the external walls of a building In the moment-resisting frame, the

horizontal forces are transmitted by bending moments in the columns and intheir framing beams The moment-resisting frame can be designed (using steel

or reinforced concrete) to be as strong as required, but frame structures will tend

to be rather more flexible than braced or shear wall structures

Similarly it is not always necessary (especially in a small building) for a fullyrigid diaphragm to be provided at each level Cross-bracing of a framed floor(steel or timber or trusses), along with the provision of a ringbeam in concrete

or even timber, may in some cases be an adequate alternative

Where a flexible, moment-resisting frame is to be used, care also needs to

be taken with the additional bending moments in the columns which arise from

Figure 8.5 Alternative earthquake-resistant structural forms: shear wall structures, moment-resisting frames and braced frames

the relative displacement of their ends This so-called P-delta effect can be the

cause of rapid material breakdown and collapse if adequate provision has notbeen made for it.6

8.4 Structural Form and Earthquake Resistance

The simple elements of an earthquake-resisting structure described in the vious section can be provided in a great variety of ways But simply providingthese elements is unfortunately not sufficient to guarantee good performance in anearthquake The static force analogy presented above fails to explain the complexbehaviour of real structures subjected to the unpredictable, large and rapidly vary-ing forces of real earthquakes In addition, there are certain principles of overall

pre-structural design which need to be observed Structures should be symmetrical, continuous, small in plan, not elongated in plan or elevation.

Experience has repeatedly shown that simple structures symmetrical in planperform much better in earthquakes than complex and unsymmetrical ones Theforce distribution in complex and unsymmetrical structures under earthquakeloading is extremely difficult to predict; torsional forces are liable to be set up

if the centre of mass is not coincident with the centre of resistance, and this cancause local failures

Adequate design of members and details with complex arrangements and undercomplex force systems is much more difficult than for simple cases The sameapplies to re-entrant plan shapes even if they are symmetrical Uniformity andcontinuity of structure are of equal importance, because changes in cross-section,either in overall elevation or in one particular element, cause concentrations ofstress which are very damaging

6 For further details see Dowrick (1987), Penelis and Kappos (1997) or Booth (1994).

Experience has shown7 that a structure will have the maximum chance ofsurviving an earthquake if:

• the load-bearing members are uniformly distributed;

• the columns and walls are continuous and without offsets from roof tofoundation;

• all beams are free from offsets;

• columns and beams are co-axial;

• reinforced concrete columns and beams are nearly the same width;

• no principal members change section suddenly;

• the structure is as continuous (redundant) and monolithic as possible

The concept of redundancy implies that any applied load can find many native routes (load paths) to the ground Given the unpredictable nature ofearthquake motions and the real chance of local overload, a structure designed

alter-so that if one element fails others will be able to carry its load must evidentlyhave a better chance of survival in an earthquake

Avoid Soft Storeys

One particular type of discontinuity is worth elaborating on Very commonlymulti-storey frame buildings are provided with cross-walls or frame infilling inresidential upper storeys, but these are omitted or partially omitted on the groundfloor to provide open commercial or car-parking space; this is often the cause of

a serious weakness on this floor This has been the cause of the disastrous failure

of the ground floor of many buildings such as that illustrated in Figure 8.6 Theeffect of setbacks in elevation is similar and these should also be avoided for thesame reason

Plan Size and Slenderness Limitation

Limiting the size of a building size in plan is important because earthquake forcesvary rapidly in both time and space and a long building is likely to have differentground movements applied to it at each end, coupled with ground distortionalong its length Where a long building is needed for planning reasons, it islikely to perform better if subdivided into separate short lengths of structureswith movement gaps between them.8

The slenderness of a building should also be restricted to limit horizontaldeformations: a height/width limitation of 3 or 4 has been proposed,9 althoughthis can be exceeded with good design

7 Dowrick (1987).

8 The legendary survival of Frank Lloyd Wright’s very large Imperial Hotel in the 1923 Kanto (Tokyo) earthquake has been partly attributed to its separation in this way.

9 Dowrick (1987).

Figure 8.6 Collapse of reinforced concrete buildings in Adapazari, Turkey, in the 1999 Kocaeli earthquake

Columns Stiffer than Beams

In framed buildings, additional important rules of design must be observed Onerequirement is that columns should be stiffer than the beams which frame intothem If this is the case, the beams will fail before the columns, limiting failure

to the area supported by the beam and enabling the beams to be used as energyabsorbers; where the columns begin to fail first, failure tends to occur veryrapidly, under their vertical load

Infill Panels

The use of stiff infill panels in framed buildings as cladding or as internal orexternal partitions presents serious problems: often they are not treated as a part

of the structure and are themselves weak However, in an earthquake they tend

to attract load initially, because of their stiffness When they fail, this will be abrittle type of failure, which can cause serious damage to the main structure, aswell as injury to occupants, and result in serious economic loss to the building,

even if the main load-bearing structure is unharmed Thus infill panels either

should be treated as a fully integral part of the structure (making it a shear wall

not a frame structure) or should be totally separated from it by movement joints

which allow the frame to move independently This latter approach presentsdetailing problems if the structure is expected to support the infill panel undernormal conditions Infill panels can have an equally disastrous effect if they are

discontinuous in either elevation or plan The effect of this is to create regions

of high stress concentration in a structure for which it was not designed, causinglocal failures

Separation Between Buildings

Individual buildings need to be provided with adequate separation, to prevent

damage caused by pounding when they deform in earthquakes, which has been

a serious cause of damage, even of collapse in recent earthquakes The mum separation gap depends on the height and flexibility of the building Thegap between buildings should exceed the expected cumulative maximum drift(lateral displacement) of all storeys added together with an extra allowance.Separation can be a particularly difficult problem to deal with where a tallbuilding of complex or large plan is divided into smaller separate structuralelements for reasons discussed above The gaps created then generally need to

mini-be bridged to preserve functional continuity, but it is essential that any bridgingshould be designed not to transmit forces, so as to maintain structural separa-tion.10

Alteration to Existing Buildings

Stress concentrations are very frequently caused by supposedly non-structuralalterations carried out on existing buildings when their function changes Notonly the addition or removal of partitions, but also the positioning of windows,doors and staircases can significantly affect the earthquake performance of abuilding Vertical or lateral building extensions, particularly where new materialsare to be used, can be equally damaging

Non-structural Elements

Finally, to achieve good earthquake performance, it is essential to pay attention tothe non-structural elements of a building.11In recent earthquakes a high propor-tion of the damage was unrelated to the main structure of the building Heatingand cooling plant, fuel, electricity and water supply mains, elevator equipment,etc., need to be secured to resist earthquakes, otherwise serious damage includ-ing fire outbreaks can occur Heavy furniture and equipment such as bookstacksneed to be properly secured Flying glass is a serious hazard in urban areas andfor flexible high-rise buildings detailing for movement is needed Cupboards andbottles containing hazardous chemicals have to be specially designed to avoidspillages

10 Solutions have been discussed by Arnold and Reitherman (1982) and Dowrick (1987).

11 Lagorio (1991).

The jamming of doors in buildings as a result of deformation is a serioushazard as it may prevent escape or rescue; doors should be detailed so that somemovement of the structure can occur without causing them to jam.

Foundations

Foundations, particularly for large buildings, should equally be kept as simple

as possible Only one type of foundation should be used for the whole building(or any structurally independent part of it) Separate column or wall foundationsshould be interconnected so as to achieve an integral action, and should all rest

at the same level Foundations should be loaded approximately uniformly undervertical load, and where possible sites with large variations in subsoil conditionsshould be avoided

8.4.1 Engineering Techniques for Improving Earthquake Resistance

Some new engineering techniques for modifying the structure to achieve betterearthquake resistance are available, and can be expected to become more widely

used in the future The most important of these techniques are base isolation, and the use of energy absorbers.

Base Isolation

The principle of base isolation is to introduce some form of flexible support at thebase of a building so that earthquake forces transmitted to the building are muchlower than if the building is firmly fixed to the ground The simplest form of baseisolation is a frictional sliding layer, which will slip if the force exceeds a certainproportion (perhaps 3–5%) of the weight of the building As such slip is likely

to result in permanent displacements, a spring system is normally preferable.Spring systems will transmit forces proportional to the relative movement of theground and the building, and incorporating them will increase the natural period

of vibration of the building, hence (for most earthquakes) considerably reducingthe forces the building experiences.12Laminated rubber springs are the materialsmost widely used; they have a much lower stiffness in the horizontal directionthan in the vertical direction, and thus are effective only to reduce the damag-ing horizontal forces A lead core is incorporated to provide energy absorptionthrough damping, thus further reducing the earthquake loads experienced by thebuilding.13

12 Key (1988), p 70.

13 Base isolation techniques have been discussed by Key (1988) and by Buckle and Mayes (1990).

Energy Absorbers

The function of energy absorbers is to absorb energy by deforming if the structureexperiences a large earthquake, thus protecting the main supporting structure.The energy absorbers will experience permanent deformation (i.e they will bedamaged), but they can be replaced at a much lower cost than that of repairingdamaged structures A common location for energy absorbers is in cross-bracingfor framed (particularly steel frame) structures.14

Both base isolation and energy absorption techniques are at present relativelyexpensive and their current use is mainly in high-rise buildings, But there is agrowing body of evidence to demonstrate their effectiveness

by moment during the earthquake ground motion Only a few such systems havebeen installed to date and there is little experience of their effectiveness in realearthquake events, but they hold promise for the future.15

8.5 Choice of Structural Materials

The choice of materials for building materials in seismic areas is to a large extentdictated by questions of availability and cost The essential material requirementsfor earthquake-resistant structures are strength and ductility, and these properties

are closely interrelated Ductility refers to the ability of a material to deform after

its maximum strength has been reached, without losing its ability to carry load.Structures made from materials which have this property can survive short-termaccidental overloads because, rather than breaking, they can deform during theoverload and absorb a large amount of energy without losing strength, instead

of simply breaking Steel is an inherently ductile material, and is thus verysuitable for building in earthquake areas.16California and Japan make extensiveuse of steel in large buildings of all types Concrete and all types of masonry,without reinforcement, are brittle materials, but by means of embedment of steel

14 There are a wide variety of techniques which have been discussed by Key (1988) and Hansen and Soong (2001).

15 Soong and Spencer (2000).

16 Although welded joints can be a source of weakness and have resulted in some failures in recent earthquakes.

reinforcement, suitably placed, they can be made to perform in a semi-ductilemanner, making them suitable for earthquake-resistant construction.

Since the extra forces resulting from an earthquake are proportional to themass of the structure, structural materials which are strong and ductile but light(i.e have a high strength-to-weight ratio) are particularly suitable for earthquake-resisting structures Timber has the highest strength-to-weight ratio of all struc-tural materials, and steel is also good in this respect Reinforced concrete is not

so good, and masonry is poor in terms of strength-to-weight ratio

Nevertheless, in many parts of the world economics dictates that reinforcedconcrete frame structures are used for mid-rise and high-rise buildings; to makesuch structures earthquake resistant requires careful attention to continuity and to

ductility requirements Such frames should be cast in in situ concrete; adequate

ductility and continuity are much more difficult to achieve with precast concrete.17Reinforced masonry, of brick, block or dressed stone, is a good material forlow-rise structures, since it combines high shear strength with some ductility,provided that certain important rules are observed Unreinforced brick masonry isless suitable in areas of high seismicity As Figure 1.1 shows, the great majority

of earthquake deaths over the last century have been caused by the collapse

of unreinforced masonry buildings But unreinforced masonry may be used withacceptable safety in moderate- or low-seismicity areas if the elements are properlyinterconnected as described in Section 8.4

Low-strength unreinforced masonry materials (such as rubble, stone and adobe)have an extremely poor seismic performance and should be avoided wheneverpossible However, where there is no economic alternative to their use, eventhese materials can, with suitable reinforcement of timber, steel or reinforcedconcrete, be made to behave in a semi-ductile fashion which will significantlyimprove their performance in moderate earthquakes Techniques are discussed inSection 8.7

Composite structures of steel or concrete frame with masonry infill panels aretoday much used in seismic areas because of their cheapness, and they need alot of care in their design and construction Just as an addition of reinforcementcan help to enable brittle materials to achieve some ductility, the ill-consideredincorporation of brittle infill materials can cause ductile materials to behave in anon-ductile way.18

Timber frame is mostly used only in low-rise structures: it has generally formed very well in earthquakes as a framing material owing to its high strength,low weight, and the ductility provided by flexible but energy-absorbing joints Themajority of domestic construction in New Zealand, Japan and southern Californiauses timber frame construction However, in older structures where the timber

per-17 Precast concrete structures have a poor record in earthquakes, as shown by recent experience in the 1994 Northridge earthquake in the United States (EEFIT 1994) and the 1999 Kocaeli earthquake

in Turkey (EEFIT 2002b).

18 Some rules for detailing such structures have been proposed by Smith (1988).

is not well preserved, its deterioration can be a problem, and it contributes tothe risk of fire damage in earthquakes Heavy roofs supported on old timberwith poor connection to masonry walls or unbraced timber frames have beenresponsible for many deaths.19

8.6 Codes of Practice for Engineered Buildings

8.6.1 Philosophy

In seismic areas buildings will collapse or be seriously damaged unless they arespecifically designed to withstand the expected loads from future earthquakes,and rules are needed to guide designers on how to achieve this safety Both thelevel of protection to be aimed for and the means of achieving it will vary fromplace to place, according to the level of seismic risk, the resources available, thetype of construction being considered and the capability of the building indus-try Some of the world’s most technologically and industrially advanced areas,such as Japan and southern California, are located in regions of high seismicity.Both experienced a series of devastating earthquakes during the twentieth cen-tury, and continue to experience regular shocks In these areas, codes of practicefor the design of new buildings have been in place for most of the last cen-tury; they are almost universally understood and adopted by designers of largebuildings, and they have become models for the codes of practice used in othercountries

In California, the level of resistance aimed for in design has, since the late1970s, been based on the concept of an ‘acceptable risk’ The objectives are:20

1 To resist minor earthquakes without damage

2 To resist moderate earthquakes without significant structural damage, but withsome non-structural damage

3 To resist major or severe earthquakes without major failure of the structuralframework of the building or its component members and equipment, and tomaintain life safety

It is also recognised that certain critical facilities should be designed to remainfully operational during and after an earthquake

Within this general framework the scope and reliability of the codes have beenable to develop in recent years in step with the rapid developments in the scientificknowledge of earthquake hazards and the engineering understanding of the effects

of earthquakes on buildings The most recent US codes, incorporated in the

19 The majority of the deaths in the 1995 Kobe earthquake were reportedly caused by the collapse

of poorly maintained, timber-framed houses supporting heavy tiled roofs.

20 As defined by Applied Technology Council, ATC-3-06 (1978).