The numerical solution of large scale dynamic soil-structure interaction problems.

It is common practice in civil engineering to use computational models in order to understand and predict the behaviour of a wide variety of structural systems. The dynamic interaction between structures and the soil often plays a crucial role and should be accounted for in numerical models, for example for the assessment of damage to structures caused by earthquakes, the evaluation of annoyance in buildings due to vibrations originating from road and railway traffic, or the design of offshore structures subjected to wind and wave loadings. Despite the remarkable progress of computer hardware over the last decades, the solution of many large scale problems remains very challenging and often beyond current computer capabilities. The envisaged research project aims at developing and applying fast, stable, and accurate numerical methods in order to cope with ever larger and more complex dynamic soil–structure interaction problems. The first part of the project is dedicated to the development of fast solvers for large linear systems of equations, as arising from finite and boundary element discretizations. Research will be performed in collaboration with Stanford University and Lawrence Berkeley National Laboratory. These techniques will subsequently be employed for applications such as the forward modelling of railway induced vibrations and re–radiated noise, as well as for quantifying and reducing the associated uncertainty through the efficient solution of inverse problems.