Numerical, experimental and hybrid methods for the prediction of railway-induced ground vibration

Ground-borne noise and vibration due to railway t raffic are important environmental issues in the developmen t of new railway lines and land adjacent to existing lines. Predi ction models are indispensable to address properly these is su s and to design efficient mitigation measures for both existing and n ew-build situations. The aim of the present contribution is to provide a comprehensive overview of methods which have been develop ed for the prediction of railway-induced ground vibration. A distinction is made between numerical models, based on phys ical or mechanical models, empirical models, derived from m easured data, and hybrid models that combine elements from numerica l and empirical models. Numerical models are particularly useful in new-build situ ations. They can be used to assess the environmental impact o f new railway lines or to develop and design mitigation measures. Numerical predictions are generally characterized by sign ificant model and parameter uncertainty, however, which need to be consid ered for robust engineering and design. In engineering prac tice, empirical methods are mostly used, which are based on an expe rimental characterization of the source strength and vibra tion transmission. These methods have the advantage of accounti ng for the particularities of the vibration transmission at a given site, but can only be applied in those cases for which suitabl e data and prior experience are available. In order to overco me the limitations of numerical and empirical methods, these meth ods can be combined in a hybrid or semi-empirical prediction pr cedure. It is shown how such a hybrid method is obtained by replacing t he experimental source strength or vibration transmission term by an equivalent quantity which is computed using numerical si mulations.