¿ìè¶ÌÊÓÆµ

Standing up to earthquakes: An earthquake destroys buildings by shaking them to pieces. Engineers are finding ways to keep them standing when the ground moves

Nature doesn’t negotiate. This axiom is worth remembering as we enter
the second year of the International Decade for Natural Disaster Reduction,
declared by the UN. Among the non-negotiable events to be expected during
the decade are earthquakes. As an engineer in California observed in the
aftermath of the state’s 1989 quake: ‘Earthquakes don’t kill people. Buildings
»å´Ç.’

Nobody wants to spend extra money to prepare for something that may
never happen and earthquake engineering has long been neglected. Now that
attitude is changing, most obviously on the West Coast, which has more money
than other earthquake zones to experiment with new designs and materials.
The value of well-built buildings is not lost on Californians. The Loma
Prieta earthquake of 17 October 1989 in San Francisco reached a magnitude
of 7.1 on the Richter scale and killed 62 people; the 1988 quake in Armenia,
with a magnitude of 6.7, killed about 25,000.

Earthquake engineers may not have needed to be reminded that building
design can make all the difference between life and death, but the Californian
quake has increased public interest in earthquake-resistant designs. These
range from the mundane, such as reinforcing masonry walls with steel beams
to strengthen them, to the exotic, such as supporting an entire building
on rubber so that the structure ‘floats’ in isolation.

To understand how engineers are preparing for earthquakes, one needs
to know a little about how buildings behave when shaken. Place several dishes
of jelly on a table and rhythmically shake the edge. The jellies will shake
too, but some more than others. Each jelly will sway back and forth in a
characteristic time, or period, that depends predominantly on its height
and consistency. The reciprocal of the period, measured in seconds, is the
pudding’s natural frequency of vibration, which is measured in cycles of
movement per second, or hertz. Those jellies whose natural frequency matches
the frequency of the shakes will oscillate with the greatest amplitude –
the two frequencies are said to be ‘in resonance’.

Buildings behave in much the same way. A typical building of 10 storeys
will sway back and forth in a period of 1 second, giving it a natural frequency
of 1 cycle per second or 1 hertz. As a rule of thumb, every storey makes
a difference of one-tenth of a second to a building’s period. So a 20-storey
building has a period of about 2 seconds, a natural frequency of 0.5 hertz;
a five-storey building has a period of half a second, a natural frequency
of 2 hertz.

Buildings above 20 storeys fare relatively well in earthquakes. The
most powerful vibrations from a quake range from 0.5 hertz to about 5 hertz.
As a typical 20-storey building has a frequency of 0.5 hertz, and taller
buildings have even lower frequencies, they are likely to be beyond the
range of resonance. In addition, taller buildings flex more than short ones,
by virtue of their height and because their frames are built to bend with
the wind. Steel columns and beams, and concrete walls reinforced with steel
rods, make the frames flexible – and even ductile, so that they can deform
without snapping in the wind. Flexibility has its costs, however. Large
deflections can help to loosen non-structural cladding and shake up the
building’s contents like dice in a box.

Buildings of less than 20 storeys are the most susceptible to resonance.
Those that resonate with the strongest vibrations will suffer the most damage.
To understand resonance, think of a child on a swing. The arc described
by the child has its own cycle. If you apply force just as the swing attains
its maximum height and begins to drop, you add energy most efficiently,
in synchrony with the swing’s cycle. Too soon, and you dissipate it., So
if the period of a quake’s main vibrations are 0.5 seconds, five-storey
structures will be particularly vulnerable. In such cases, the building
amplifies the motion it experiences at ground level, and the horizontal
accelerations at the upper storeys can be several times those at the base.

Urban buildings set in rows present another threat to themselves and
their neighbours. Take a 15-storey building abutting a three-storey structure.
The periods differ, so the buildings vibrate with different frequencies
and slap against each other. Masonry may crack or crumble on the wall of
the taller building where the roof of the shorter one abuts. This happened
in Santa Cruz during the Loma Prieta quake, where many adjoining buildings
were of different heights.

How buildings fare in an earthquake can depend on design and materials.
Wooden frames are better able to flex and, in effect, have longer periods
than stiffer, brick or stone constructions of the same height. In San Francisco’s
Marina district, one design of house suffered badly during Loma Prieta.
These were stylish houses built over an empty space, used as a garage, with
few supporting columns. During the quake, the columns failed to damp the
lateral motion of the building, and the top-heavy houses swayed and collapsed.
Ironically, much of the Marina district was built on rubble from the 1906
quake.

Foundations call the tune

Soil, too, has its own periodicity. The soil between bedrock and the
surface behaves like those plates of jelly on the table. It has a particular
stiffness depending on the type of soil – clay, sand, gravel and so on –
and how much moisture it carries. The significance of soil periodicity was
driven home in Mexico City after the earthquake on 19 September 1985. The
devastation in the city’s Lake Zone, about 400 kilometres from the epicentre,
was surprising even for a quake of 8.1 magnitude.

According to calculations by Edmund Booth, an earthquake specialist
with Ove Arup and Partners, a large firm of consulting engineers based in
London, the soft clay under Mexico City ‘tuned’ the period of the quake’s
ground motion to something near a pure sine wave with a period of 2 seconds.
Buildings of 20 storeys resonated with these vibrations and, as a result,
the horizontal acceleration at the top of these buildings was five times
greater than that at ground level. Ground acceleration, measured as a percentage
of the acceleration caused by gravity (980 centimetres per second per second),
is of particular concern to earthquake engineers. This is because the shearing
force on a building depends on the building’s mass multiplied by its horizontal
acceleration during a quake; vertical acceleration is usually much less
in quakes, and buildings are built to resist vertical forces anyway.

In Mexico City, the acceleration in bedrock was about 4 per cent of
the acceleration due to gravity. But at the foundations of buildings, the
clay amplified it to more than 20 per cent of gravity. Even buildings whose
shorter periodicity put them out of the vulnerable category suffered; the
lengthy shaking made them progressively less brittle, in effect lengthening
their period until they vibrated in resonance with the quake, at which point
many began to collapse. By the time the earth finally stopped moving, 200
buildings had been destroyed, thousands more were damaged, and as many as
20,000 people were dead, most entombed in collapsed buildings.

Engineers calculate soil periodicity before designing a building, at
least in the western US, notes Robert Whitman, an earthquake engineer at
the Massachusetts Institute of Technology. But determining the period of
a mass of soil is tricky. ‘If we are in the middle of a broad, flat valley,
our mathematical models work well,’ says Whitman. ‘But if we are in a deep,
narrow valley, say with width less than four times depth, we don’t really
know the period, due to reflections.’

Besides period, soil is affected by how wet it is. If you shake a beaker
full of dry sand, the particles will settle and the material will become
denser. Shaking saturated soil is another matter. When a sudden shock wave
pressurises water captured in the pores between soil particles, there is
no time or place for the water to move: the high pressure pushes the soil
particles apart. During earthquakes the soil becomes a semiliquid soup that
no longer supports the weight of the buildings above. Clay particles are
more cohesive than other soils and deter the pressurising of trapped water,
while fine sands are the easiest to liquefy. Buildings many remain structurally
intact but tilt like rows of old tombstones as their foundations lose their
grip. This happened in the 7.5 magnitude quake at Niigata, Japan.

Isolating buildings from the ground

Driving piles or draining soils that are likely to liquefy is a hit-or-miss
proposition, according to engineers at Ove Arup. For the Rangoon General
Hospital, which will be built on loose sandy soil, Ove Arup proposes to
solve the liquefaction problem by laying the foundations deep in the ground,
so that the weight of the building will roughly equal that of the soil dug
out. In the event of the sand liquefying below the basement, the soil should
still be able to support the building.

Some of the most enthusiastic believers in earthquake engineering can
be found in Malaysia. It was the Malaysia Rubber Producers’ Research Association
that helped to turn a concept known as ‘base isolation’ into reality at
the University of California. Base isolation is a simple idea. The building
sits on bearings that isolate it from the ground. During a quake the bearings
intercept, absorb and damp vibrations. In effect, the building’s frequency
is lowered or ‘detuned’ below the earthquake’s, and the structure moves
like a rigid body above the isolators instead of flexing.

The technique was first patented early this century in Britain, but
not developed until 1976 when the idea caught the attention of James Kelly,
professor of engineering at the University of California, Berkeley. In 1977,
Kelly and the Malaysian rubber industry, with the support of the university’s
Earthquake Engineering Research Center, began testing rubber base isolators
inside a sprawling warehouse in a dingy industrial district north of San
Francisco. ‘People thought we were idiots for a long time,’ remembers Kelly.
The warehouse, which belongs to the university, had something no other laboratory
had: the country’s first large shaking table.

The shaking table is a 50-tonne concrete plate set horizontally in a
hole in a concrete floor with several centimetres of play at the edges.
In a cellar-like room below, however, a spider’s web of tubing and hydraulic
actuator arms is connected to the table. During an experiment, the air in
the cellar is compressed to 3 pounds per square inch (about 21 kilonewtons
per square metre) and the table floats. The hydraulic arms then buffet it
to simulate the forces at work during an earthquake.

Kelly mounted experimental structures on bearings on the table to discover
the best ratio between the height of a building and the size and elasticity
of the bearings. In 1982, Kelly and Robert Rigney, chief administrative
officer of San Bernadino county, persuaded the state to pay for a real building
with isolators. Their goal was to create a building that would remain functional
after an earthquake with a magnitude of 8.3, as big as anything that has
ever hit California. They got their way, and the building opened in 1986
as the Foothill Communities Law and Justice Center, about 100 kilometres
east of Los Angeles.

The building’s 98 rubber bearings, each between 0.5 and 0.75 metres
in diamater with steel laminations and weighing 500 kilograms, attracted
plenty of attention during the construction project. Not until 28 February
1990, however, did the structure experience a large quake, when one of 5.5
magnitude struck just 10 kilometres away. The horizontal ground acceleration
measured 26 per cent of gravity in a car park nearby. But it was only 8
per cent of gravity just above the bearings, and 16 per cent at the roof.
Motion at the rooftops of two nearby buildings was up to three times that
at ground level.

One advantage of base-isolated buildings is that they move as a unit.
Ordinary buildings distribute motion upwards from storey to storey, which
can cause ‘interstorey drift’ that can knock floors out of line as though
the structures were stacks of sliced bread. Nuclear power stations are especially
vulnerable to this kind of drift. Kelly has several nuclear stations as
customers for his isolators.

California now has about 10 structures with base isolation. A handful
of others, including a 19th-century stone building in Utah whose foundations
are fitted with rubber bearings, exist elsewhere in the US. Kelly notes
that builders are constrained by the threat of lawsuits should something
happen to a structure built outside normal building codes. In Japan, however,
the concept has caught on.

Japan has about 40 such buildings of various sizes being built or completed.
The country’s six largest construction companies, which together spend between
$14 million and $16 million a year on earthquake research, met the extra
costs of including base isolation in the structures. In the US, where the
government is the only major backer, research has limped along at about
half a million dollars a year. Japanese engineers are also adding some flourishes,
such as steel rods inside the rubber bearings to prevent the isolation deforming
too much as they absorb vibrations. Some buildings have been equipped with
oil-filled isolators to damp any vertical movement.

In a recent quake, peak ground acceleration reached 4.4 per cent of
gravity at a base-isolated building built by the Ohbayashi Corporation.
The acceleration at the roof measured only 0.71 per cent of gravity; at
an adjacent, four-storey conventional building, roof acceleration was 5.9.

Rubber bearings are not the only form of base isolation. Another approach
being tested is the friction pendulum system. At the building’s foundations,
each vertical column stands on a semicircular slider that is seated in a
concave, spherical surface of chromium steel. When a quake shifts the column
and slider – think of a pestle scraping along the bottom of a mortar – the
slider faces ever more frictional resistance as it ‘climbs’ the curvature
of the concave surface. Michael Constantinou of the National Center for
Earthquake Engineering Research at the State University of New York (SUNY)
at Buffalo champions this approach. Five buildings of a hospital he is designing
will be seated on these bearings.

Adding bearings to a building can increase the cost by up to 10 per
cent. ‘That’s a bargain when you consider how much protection you get for
a building’s contents alone,’ says Kelly. Base isolation, though, has its
doubters. Loring Wylie, chairman of Degenkolb Associates, a firm of consulting
engineers based in San Francisco, says the technique requires more testing.
‘We know how a building works on a shake table. But what if you have one
near the San Andreas fault? In a 1966 quake, the ground displaced 3 to 6
metres. Somehow a building has to ‘track’ the ground, to follow it. The
shake table always moves back to zero.’ Buildings are connected to pipes
and cables in the ground that will break if there is too much movement.
Also, if rubber bearings do not return to their original shape after a quake,
the building loses some stability, says Wylie.

California’s philosophy underpinning its seismic codes holds that a
building should not be damaged in a minor quake, should suffer no more than
minor structural damage in a moderate quake and not collapse in a major
quake. While California’s engineers do not specify precisely what a minor
or moderate quake is, a major quake is one that will accelerate a building
at 40 per cent of gravity. Another quake the size of Loma Prieta would qualify.
Memories of Loma Prieta are not the only incentive; the US Geological Survey
predicts a 50 to 60 per cent chance of a major quake in California in the
next 30 years that will cost $50 billion in property losses alone.

Geologists and engineers have begun preaching the virtue of preparation
in the central and eastern US as well, to a somewhat less receptive audience.
The rarity of quakes there has lulled people into a sense of security that
many say is false. While the chances of a major quake are low, the consequences
would be great. ‘A small one, even magnitude 5 or 6, would cause a lot of
grief,’ warns Ian Buckle, deputy director of SUNY’s earthquake centre. Besides
lacking reinforcement, buildings in many central and eastern cities sit
above homogeneous basement rock, which conducts vibrations for much greater
distances than the ground in California.

The cost of making safe American buildings, bridges and pipelines almost
defies imagining. No one even knows where to expect a quake, as many of
the eastern fault lines lie deep and undetected. For now, engineers at SUNY
are trying to make people aware of the risk while perfecting some known
techniques for new buildings. For example, ‘moment resisting frames’ – building
frames in which columns and beams are rigidly attached and reinforced to
carry the building’s load – help stiffen buildings. So do ‘shear walls’
made of reinforced concrete and coupled with the building’s frame.

The SUNY team is also working on some novel concepts for existing buildings.
For smaller structures, it has developed something akin to an insect’s cocoon:
a very fine wire mesh riveted to the walls, floor and ceiling of a building,
then sprayed with ferrocement containing metal particles. The layer reinforces
the wall’s weak link – the mortar. Mortar can support vertical, compressional
stress but cannot handle lateral forces. ‘It’s like a reinforced thin skin,’
Buckle explains. ‘The results on the shaking table are quite astounding.’

Other measures can help stiffen a building but are less than ideal.
Diagonal steel braces can be inserted behind walls, but this is an expensive
intrusion. Shear walls can be added. But this kind of retrofitting can cost
up to half the value of a building, and owners are not likely to make the
investment.

Those who ignore the warnings do so at their peril, however. Earthquake
experts like to talk about what lies in store for New York City. The city’s
building codes do not incorporate seismic reistance, and more than half
of Manhattan’s buildings are built of unreinforced masonry. Many of New
York’s bridges are made using inadequately reinforced concrete. The engineers
at SUNY are trying to convince the city’s department of transportation to
take seismic loading into account for its new bridges.

Older bridges are at great risk; they could be made safer by fitting
elastic isolators where thermal expansion bearings now sit between girders
and columns. This solution, however, is not suitable for a crossing that
consists of one simple span supported at the ends. In the east, says Buckle,
‘many bridges just sit on the columns like a stack of cards’. Taller columns
will sway more at the top, causing the span to unseat itself.

Then there are the cast-iron pipes that supply New York City with water.
Peter Gergely, professor of structural engineering at Cornell University
in Ithaca, New York, says, ‘200 to 300 fires could start, and if the pipes
go out you don’t have water to fight them.’ As for soil, it varies considerably,
ranging from bedrock under much of Manhattan to sandy sediments under eastern
suburbs that would amplify the effects of small quakes.

Experts also point to the oil and gas pipelines that feed the eastern
cities. For example, important pipelines pass near the New Madrid fault
region in Missouri. One of them is the Capline System, a 3-metre diameter
line that carries about 700,000 barrels a day across the centre of the country.
Soil deformation during a quake could bend pipelines as though they were
strands of copper wire. Ruptures would cut the eastern states off from their
supply of oil and contaminate ground water.

Much has been made of the fact that San Francisco emerged relatively
unscathed from Loma Prieta. Engineers are not so confident. ‘They were lucky
at Loma Prieta,’ says Gergely. The quake hit relatively far from the city
and it lasted only about 10 seconds; there was no wind to spread fires;
and conditions were dry, making landslides less likely. ‘Officials don’t
tell the public,’ says Gergely, ‘but they say they were scared.’

Three huge quakes, greater than magnitude 8, struck the New Madrid fault
in Missouri in 1811 and 1812. A big quake east of the Rocky Mountains –
and one greater than magnitude 6 has a 50 to 60 per cent chance of occurring
in the next 25 years – could make California’s Loma Prieta seem trifling.
Many engineers say such a disaster is imminent. There is much work to be
done in preparation, says Buckle. New York City and several other eastern
jurisdictions are considering the introduction of seismic building codes.
‘These are encouraging signs that the message is being heard in the East,’
he says, ‘but this is a race against time and the odds of winning do not
look favourable.’

* * *

Structures tuned to the rhythm of a quake

The ultimate in earthquake-proof buildings is going up in Tokyo. Not
satisfied with simply strengthening its new 11-storey office building conventionally,
the Kajima company, one of Japan’s major construction companies, is adding
‘active controls’. Justifying their expense requires a genuine earthquake
threat and deep pockets. Japan has both.

Kajima’s Kyobashi Seiwa building will look something like a chopstick
when it is finished – the structure is 4 metres wide, 13 metres long and
33 metres high. The brains of active control systems are computers that
respond within hundredths of a second to the vibrations of an earthquake.
The muscle of the control system is on the roof. There, mounted on rollers,
are a pair of steel counterweights. Hydraulic rams move the counterweights
across the roof to counteract the back-and -forth swaying of the building.
Similarly, 1 and 4-tonne blocks move to counterbalance the torsional, (twisting)
stresses on the structure.

A network of sensors in the basement, sixth floor and roof feed data
on the structure’s movement, either from wind or stress, to a mainframe
computer. The computer calculates when to roll the counterweights and by
how much. Kajima calls the system an ‘active mass driver’. Edmund Booth
of Ove Arup likens the physics to ‘standing in a railway carriage and flexing
your legs to keep your balance’.

Another approach, designed by the earthquake engineering centre at the
State University of New York at Buffalo and Takenaka Corporation of Japan,
comes even closer to Booth’s analogy. Steel arms, called tendons, horizontally
bisect a building’s core, stretching like ribs between beams in the walls.
Controlled by computer and battery-powered hydraulics, they push or pull
like pistons to damp movement by the walls.

This system was installed in a six-storey experimental building in Tokyo
in August 1989 and has passed muster in six small earthquakes since then.
During a quake of magnitude 6.1 on 20 February 1990, ground acceleration
was measured at 10 per cent of gravity. That was halved by the building’s
bracing.

Topics: earthquakes

More from ¿ìè¶ÌÊÓÆµ

Explore the latest news, articles and features