Computational considerations of 3-D finite element method models of railway vibration prediction in ballasted tracks

Abstract

The study of vibrations induced by a passing train is of utmost importance to understand better this phenomenon and to design efficient mitigation measures. Within all the techniques used to model vibrations, finite element methods allow introducing in the model detailed characteristics of the real vehicle-train-soil system. However, the accuracy of the results is linked to how detailed the real elements implemented in the model are and consequently, with the computing time. In this paper a three-dimensional finite element method to predict vibrations is developed and validated with real field-measured data. Then, different scenarios are represented to assess the efficiency of the model linking the quality of the results obtained with the calculation time required in each case. Finally, a reflection regarding the constitutive model of the materials when working with finite element models is done.