The fear of earthquakes is helping engineers devise a control system to mitigate their impact on buildings.
'There is no natural phenomenon which is held by all mankind in greater dread than earthquakes. Our ideas of permanence, solidity and strength are based upon the condition of the earth as we daily see it; so that when the firm ground shakes under us, there naturally comes over the mind a feeling of abject helplessness.'
So began a front-page editorial in the New York Times of 9 April 1872, as part of its report on an earthquake six days earlier in Antioch, Syria - dubbed by one writer 100 years later as 'probably the most quake-cursed city there ever was' - which on that occasion claimed an estimated 1,800 lives.
It's that sense of helplessness that has bedevilled civilisation throughout recorded history, and while much has been learnt over the years about construction techniques to mitigate the effects of earthquakes on man-made structures, these days control technology is playing a growing part.
The technology is classified according to three types of system - passive, active and semi-active or hybrid. Passive systems are the most mature of the three, and rely simply on absorbing or dissipating some of an earthquake's energy, with no feedback between them, the building's structural elements and the ground. The systems are further divided into energy dissipaters, base isolators and tuned mass dampers (TMDs).
Dr Mihail Petkovski, lecturer in structural dynamics at the University of Sheffield's Department of Civil and Structural Engineering, explains the background to energy dissipation. 'In the classic strength-based approach [to structural design] the seismic forces were calculated first and then structural elements designed to be strong enough to sustain these forces without any damage. This resulted in very large cross-sections of beams and columns.
'Modern ductility-based design, however, starts by assuming that the structure will be damaged and will dissipate energy and, as a result, forces in the elements will be lower. This leads to structural elements that are smaller (less strong) but capable of large inelastic deformations (more ductile).
'The use of these distributed passive control devices is just an extension of the ductility-based approach. Instead of relying on damage in structural elements, the energy is dissipated in special, additional elements, such as yielding or friction connections between bracing and frame elements,' he says. These distributed devices are most effective in frame buildings of between six and 14 storeys, he adds.
Passive control device
Base isolation systems come in two basic types, but both work by separating the building from the ground in some way.
The most widely adopted type uses elastomeric bearings made from natural rubber or neoprene placed between the building and its foundations. The bearings decouple the building from the horizontal components of the earthquake ground motion, and deflect the earthquake's energy through the dynamics of the system rather than absorbing it.
The second type works on a sliding principle, limiting the transfer of shear across the isolation interface. Various systems are in use, some using a special sand at the sliding interface, another containing a lead-bronze plate sliding on stainless steel with an elastomeric bearing, and a friction-pendulum system using an interfacial material sliding on stainless steel.
Isolation systems are most effective on squat, rigid buildings. In the US, the first building to use the technology was the Foothill Communities Law and Justice Center, a four-storey building about 60 miles outside Los Angeles, while the largest base-isolated building in the world is the six-storey West Japan Postal Computer Center in Sanda, Kobe Prefecture.
TMDs, by contrast, are actually used mainly to control wind-induced vibrations in very tall and slender buildings - the Shanghai World Finance Centre and Taipei 101, for example - for which seismic vibrations are usually not critical. A TMD is a large mass either suspended as a pendulum or installed on wheels near the top of the building and connected to the structure through horizontal springs. As the structure moves in one direction, the mass moves in the opposite direction and creates a force that balances the inertial force acting on the building.
The advantages with these passive systems are that they are simple, can be replaced easily if they age or break and do not need an external power source - something that is highly likely to be absent during an earthquake. Active systems, however, allow the response of a structure's response to be modified dynamically.
Active systems work by measuring the seismic input - the acceleration at the base of the structure, for example - and the structure's response, such as the relative floor displacements, at several points in the building, then generate a force in real-time to minimise the building's movement by using a large mass, say, driven by a servo-hydraulic piston.
They have great potential but their implementation is fraught with drawbacks. As Dr Petkovski explains: 'The main limitation is that the duration of earthquakes is very short, typically 20-30s, with the strongest parts often happening within 2-3s. That puts a large demand on the control system to calculate the appropriate 'countermeasures' and generate large forces for the control action, as the calculations have to be made in real-time.
'To calculate the control action, the control algorithm has to include a model of the structure. A sophisticated model would require huge computational power, but simplified models could affect the reliability of the system. So the technological limits would be computing power and a reliable (and large) power supply,' he says.
Then there is the issue of system obsolescence over a structure's lifetime. As Prof Stephen Mahin, director of the Pacific Earthquake Engineering Research Center at the University of California Berkeley, says: 'Imagine what your cellphone looked like five years ago or what the Internet was 10 years ago. We would be saddling building owners with systems that will be quite difficult to maintain in the future.'
The main thrust of research therefore is in semi-active or hybrid control systems. This is an intermediate approach where the characteristics of passive control devices can be adjusted by a controller in real-time according to a structure's response, rather than being fixed at the construction stage. It's better than passive control because if something goes wrong there is still the underlying passive control to fall back on, but it's not as expensive as fully active control because it relies less on control (and therefore computer) technology.
As Dr Petkovski explains: 'In semi-active we can, for example, measure the relative movement between adjacent floors and change the slip forces in the connections by actively changing the clamping force. This can be a simple lock-unlock of each connection, or a more sophisticated control algorithm that varies the friction forces in the connections. The difference between this and an active control system though is that here we change the reaction forces in the frame, while still depending on the energy supplied by the earthquake.'
This raises two issues - the algorithms in use and energy supplied by an earthquake.
On the issue of algorithms, Dr Jerome Lynch - who holds associate professorships in environmental engineering and electrical engineering at the University of Michigan - says: 'Although the algorithms are designed in a general way, such that they perform well in many different circumstances according to the common characteristics of earthquakes in a given area, the control system is tailored to the particular seismic risk, so the control algorithm is in turn tailored to those characteristics.'
In his view though, a good way to look at this whole issue of protection is in terms of energy input and output. As he says: 'With active control you need a large amount of electrical energy coming in to the control system to 'fight' the energy input from an earthquake; with passive control the energy flows in and out, and with semi-active control you have just enough energy to draw an earthquake's energy out of the system.'
This leads to the idea of turning the energy of an earthquake back on itself, and it's this concept that points the way to future developments.
Researchers are investigating the potential of regenerative control. An extension of semi-active control, the principle has a close parallel with kinetic energy recovery systems (KERS) in the automotive industry, where energy that would otherwise be wasted as heat during braking is converted into a temporary power boost for racing cars or electricity for hybrid road vehicles.
Dr Jeff Scruggs, assistant professor in the Faculty of Civil and Environmental Engineering at Duke University, is an acknowledged pioneer in this field and says: 'Regenerative control systems could store and re-use energy extracted from structures. But they could also transmit energy. For example, if you had a structure with, say, base isolation at the bottom and a TMD at the top, an RC system could extract energy from the base and transmit it to the top. This ability to circulate energy in a structure is potentially very useful.'
As with KERS technology, this energy would be managed using power electronics and some form of energy storage, such as a supercapacitor or flywheel, and although some energy input would be needed to run pulse-width modulation and other circuitry, Dr Scruggs says such systems would have a net energy output. The key, therefore, is the efficiency of the electronics, he says.
Another crucially important issue is the algorithms. 'Analytically, these are very challenging for regenerative control, slightly more so than for conventional semi-active systems,' Dr Scruggs says. 'And the core issue here is the constraint on the power flow through the system.' The inherent non-linearity of such systems doesn't help either.
So their implementation is still some way off - at least 15 years or so, 'probably longer', says Dr Scruggs - although single-device prototypes based on existing passive and semi-active hydraulic damper technology already exist that look set for practical use sooner.
Overall though, these various systems are there to give added protection; structures in risk zones have to comply with building design codes such as Eurocode 8. But as Prof Lynch says: 'Building codes are the bare minimum standard, while these systems enhance a building's performance - they may cost more but they can save money by limiting damage.' That equates to safety, so if you're unlucky enough to be in one of these buildings when a quake strikes, it should provide some peace of mind.