Seismic modelling at the UK's University of Bristol is taking a system approach, with one programme assessing seismic with other long term risks

THE EARTHQUAKE Engineering Research Centre (EERC) at the University of Bristol, UK, is developing a holistic approach to the whole life performance management of dam systems and similar major infrastructure assets. The approach characterises performance in terms of risk. It encompasses risks associated with natural hazards as well as those arising from general deterioration, sustainability, business needs, human and societal issues etc. At the heart of the approach is the recognition that a particular risk, such as the seismic risk, does not exist in isolation, but is affected by, and affects, other risks.

The Bristol approach develops various systems model views of the asset and its required performance. Modelling starts from a high-level view of the requirements of the many stakeholder groups and maps these in the form of key performance indicators (KIPs) in a hierarchical manner down to the physical assets.

The approach provides a framework for reconciling and prioritising the conflicting performance requirements of the different stakeholders. It also recognises the inherent changeability, non-linearity and the consequent uncertainties of the overall infrastructure system, which dictate a risk-based decision making process. Ultimately, the systems models lead to a hierarchical, risk-based performance model that allows objective and subjective evidence of performance to be combined . A software tool, PeriMeta, has been developed to achieve this.

Within this systems performance framework, the Bristol dams-related research is focusing on the seismic performance of appurtenant works and gravity dams.

In 2001, a project funded by Scottish and Southern Energy produced a set of guidelines for assessing the seismic performance of appurtenant works, especially spillway gates and associated equipment. The guidelines follow a staged approach that starts from simple studies and employs more detailed and sophisticated techniques as the available data and outcomes warrant. The guidelines yield evidence that can be integrated with the PeriMeta methodology.

A major issue that arose from the guidelines study was the need for a rapid methodology to assess the seismic performance of concrete intake towers. A new £300,000 (US$465,747) project, funded by the UK Engineering and Physical Science Research Council (EPSRC) with the support of Scottish and Southern Energy and dams consultancy KBR, is tackling this problem.

The project seeks to develop simplified, generic, non-linear dynamic numerical models of concrete towers to explore the complex and uncertain seismic performance domain of intake towers by enabling large-scale parametric studies to be executed rapidly. Such parametric studies are essential if the subtle, and often counter intuitive, features of extreme non-linear dynamic tower behaviour are to be revealed, understood and the associated uncertainties characterised.

For example, if dynamic analysis shows that a structure fails under a seismic excitation with a particular peak ground acceleration, it is not uncommon for the same structure to remain in a satisfactory state if subjected to the same excitation with a slightly larger peak acceleration (ie the shaking is stronger). The explanation for this counter-intuitive behaviour, generally, is that the earlier part of the stronger shaking induces so much damage that the structure alters and ‘de-tunes’ itself from the remaining part of the earthquake. Overall this ‘non-linear’ behaviour means it suffers less damage. So it is not enough to run only a few non-linear time history analyses, since there is no guarantee that this approach will capture this uncertain boundary between satisfactory and unsatisfactory performance.

Ongoing research, to be completed in 2005, will attempt to validate the new simplified non-linear numerical mechanical models against a series of representative small physical tower models, which will be subjected to a variety of cyclic and dynamic tests in the Bristol EERC’s new £15M (US$23.3M) laboratories.

A separate project is studying the application of distinct element analysis to assess the static and seismic performance of concrete gravity dams.

The research is using the UDEC computer program. Studies to date have shown how the method can replicate the seismically induced hydrodynamic interaction of the dam and reservoir. The current focus is on validating the fluid-filled joint behaviour against experimental data, following which a detailed parametric study of the behaviour of a typical complete jointed gravity dam-foundation-reservoir system will be conducted.

This project is scheduled for completion in early 2004.