Prediction of Ground-Borne Vibrations Induced by Railway Traffic Based on Machine Learning Techniques

Abstract

Over the latter years, there was a demand for the development of advanced numerical techniques for the prediction of ground-borne vibrations induced by railway traffic which can provide the desired level of accuracy, despite of structural complexity of the entire system. Despite the suitability of these models in dealing with such phenomena, their applicability to cases where it is intended to have a general assessment of the potential impacts of a new/updated railway project is difficult to achieve. In such cases, the development of expedited prediction methods is desired, allowing the delimitation of cases that require a deeper analysis, using advanced numerical models, and of those that can be immediately discarded, with all the benefits of cost and time associated. Thus, it is the intention to develop an innovative prediction tool, powered by an efficient and intelligent calculation engine based on surrogate modelling, that allows an efficient assessment of ground-borne vibrations at the free-field surface due to railway operation.