The search for ways of evaluating how a structure will perform under seismic loading is driving refinement of test techniques, reports Fiona McWilliam.
Two approaches have dominated seismic testing in recent years.
Shaking table testing involves subjecting a scaled model to a prescribed base motion, while in pseudo-dynamic (PsD) testing, a full or large scale model is loaded over an extended timescale with dynamic effects calculated numerically.
Useful though these experimental techniques are, however, differences in physical scale and the rate of loading can make it difficult to translate findings to real structures. This is particularly true of prototype and one-off structures, where there is no opportunity to compare modelled with actual, as built behaviour, says Martin Williams, lecturer in civil engineering at Oxford University specialising in structural dynamics and earthquake engineering.
In response, Williams and colleagues at Oxford have developed real time substructure (RTS) testing. RTS testing focuses on small parts of the structure expected to exhibit non-linear or rate-dependent behaviour. These are tested physically at full or large-scale, with the rest of the structure modelled numerically.
The two sets of data interact in real time so that the overall dynamic behaviour is accurately simulated.
In January, Oxford will be teaming up with Bristol University to enhance the versatility and robustness of the technique on a project funded by the Engineering & Physical Sciences Research Council (EPSRC). Bristol is home to the EPSRC earthquake simulator - a shaking table. Its Earthquake Engineering Research Centre (EERC) co-ordinates the work of a European Union-funded consortium of major European earthquake engineering laboratories, and is also currently working with groups in Japan and the US.
Bristol's engineering faculty has recently received £15M from the government to establish the Bristol Laboratory for Advanced Dynamics Engineering, or BLADE.
'We're trying to create a multidisciplinary approach to research, bringing together civil engineers, mechanical and aerospace engineers, computer scientists and mathematicians, ' says professor of earthquake engineering Colin Taylor.
Scheduled for completion in around 18 months, BLADE will include a new laboratory for testing large-scale structures. It will house the shaking table and incorporate several reaction walls, making it possible to test structures up to 15m tall.
Another huge problem for engineers designing in highly seismic zones is simulating earthquakes accurately.
Addressing the issue, Imperial College professor Nicholas Ambraseys recently led a project to build on Imperial's own archive of earthquake records to create a European strong-motion database, freely available via a website, www. isesd. cv. ic. ac. uk, and on CD Rom.
It contains acceleration time histories of European earthquakes since the early 1960s, and is enabling engineers to laboratory test structures using 'actual' earthquakes in real locations.
For sites susceptible to major earthquakes, but where no recorded data for large events exists, a sophisticated simulation technique has been evolved.
Recordings of small earthquakes can be employed to make predictions of the ground motions which would arise from far larger events, says pioneer of the technique and principal consultant at Aspinall & Associates, Willy Aspinall.
'This is a procedure that is being found particularly promising for regions of low to moderate seismicity, such as Britain, where data from genuinely large events are very sparse, or non- existent.'
The idea that seismograms recorded from very small earthquakes can be combined to synthesise ground motions in larger earthquakes was first proposed in 1978, but has come of age only with the advent of powerful, low cost computers. Known as 'empirical Greens Functions', the technique involves running multiple simulations, producing a range of possible behavioural outcomes.
At Imperial, reader in earthquake hazard assessment Julian Bommer is using the strong motion database to look at ways of gauging hazard in different seismic and tectonic environments - work that has applications to engineering design and loss assessment. Assessment is carried out by modelling various aspects of earthquake ground motion, including the duration of shaking and level of acceleration.
'Much of our applied research is being pushed by the developing needs of industry, ' Bommer says.
One current project for nuclear power plants in Switzerland is required to define ground motion for a once in 10M years earthquake, and is driving forward huge changes in seismic hazard assessment, he says. Because major earthquakes are relatively few and far between, even in seismically active regions, it is difficult if not impossible to measure real events, calling instead for highly sophisticated modelling.
On the geotechnical side, Dr Sarada Sarma is analysing strong-motion data recorded all over the world to derive attenuation characteristics of seismic ground motion, as well as examining the effect of local soil conditions on ground motion and how, in turn, this affects structures.
Imperial researchers are also looking at the behaviour of various structural forms and materials under seismic loading. This involves experimental work in structures laboratories using advanced dynamic testing equipment. It also includes realistic analytical simulations of whole structure behaviour, employing state of the art computer programs, as well as field investigation of how structures perform in earthquakes, and the development of design procedures for codes of practice, says Ahmed Elghazouli, senior lecturer in structural engineering and director of the MSc course in earthquake engineering.
At Bath University, Dina D'Ayala is focusing on the performance of 'non-engineered' structures, generally historic buildings or 'spontaneously built' concrete dwellings such as those commonly found in developing countries.
'I'm coming up with ways of assessing them and evaluating their seismic vulnerability, as well as the level of risk to communities, especially in urban areas, ' she explains.
'Once this vulnerability has been assessed we come up with appropriate strengthening strategies.' And these are often low-tech solutions, she adds, such as the use of anchors and ties.
INFOPLUS More information on recent earthquake engineering research is contained in the SECED report 'Implementation in the UK of earthquake engineering research', available from the SECED secretary, tel (020) 7665 2238, or e-mail:
secretary@seced. org. uk