Abstract
The influence of ballast- and substrate models on simulated impact response in rail joints of railway track is investigated. In the present, the ballast and substrate are modelled with either the lumped mass or 2-D/3-D continuous models. The 3-D models have the layered or embankment structures. The lumped mass model can be effective only for simulating the impact force at the top of ballast. However, the acceleration of ballast is accurately simulated with the continuous models. Although the 2-D model tends to overestimate the sleeper- and ballast accelaration in comparison with the 3-D model, in particular in 150-300Hz frequency range, the difference between these modellings is rather small. Hence it may be practical to use the 2-D model rather than the 3-D one which consumes huge computational work.
