|
The average annual worldwide repair and reconstruction costs following mechanical failures and earthquake damage alone has been estimated to be in the region of 40bn. Even this large figure does not include the virtually inestimable cost in human lives. Some of the major components of the 40bn loss are: 1bn due to commercial aircraft hull losses, 1bn due to repair and reconstruction following petrochemical industry disasters and 30bn following earthquake damage. The figures of most concern are those associated with post-earthquake costs, which are rising up to 20% pa. Earthquake damage is also of deep concern since the human and economic consequences are dire. During 1995, for example, the exceptionally damaging earthquakes in Northridge (USA) and Kobe (Japan) were estimated to result in financial costs of 165bn, a figure which was large enough to upset the entire global economy. Therefore it is very important that engineers devise improved methods for testing the strength of buildings and machines when they are subjected to such damaging forces. Engineers also predict the strength and integrity of structures and systems, before they are put into service. Essentially there are three ways by which this is done. The first is to build a full-size version of the system and subject it, via a test rig, to loads that are likely to be encountered in service, including earthquake-induced loads if appropriate. For all but the smallest of systems, full-size testing can be prohibitive in terms of cost and practicality. The second method is to build a smaller-scale model of the system and then subject it to correspondingly scaled-down loads on a laboratory test rig. Unfortunately, achieving compatible scaling of all aspects involved with strength and integrity is difficult to achieve with a physical model. Often, it may not even be possible to construct a truly representative scaled-physical model. This problem is the behaviour of many systems is not linearly dependant on scaling, so that scale-model testing can yield results of limited value. The third method depends on the derivation of a mathematical model, which is then solved by a numerical method, or simulation, on a computer. Although this form of modelling is not encumbered by the problems of scale, two issues dominate the acceptability of the results: firstly, the accuracy of the mathematical model and secondly, the ability of the numerical method to converge to the 'correct' solution. In practice, therefore, many engineering predictions are based on a compromised combination of all three methods.However, there is now a new and exciting alternative to the three methods, called dynamic substructuring, which enables us to test a combination of the critical, full-size, sub-components of the original system, linked to mathematical models of the remaining parts. This combination of physical sub-components and mathematical models must be controlled very precisely, using a computer, in order to properly represent the original system. The development of this method, and its control, forms the basis of this proposal. So, the principal objective will be to establish a unique world-centre for dynamic substructuring research called ACTLab / the Advanced Control and Test Laboratory. This will put the UK at the forefront of world research in advanced testing methods. More importantly, the UK will have a centre recognised worldwide for the contribution it makes to saving lives, preventing damage to the built environment and reducing the ultimate cost of natural or man-made disasters.Finally, it is not enough to practise science and engineering solely within a research environment. For this reason, we will be setting up an Educational Centre within ACTLab that will enable visitors to see the testing methods in action, to try out some of the methods for real and so understand how the methods work.
|