Assessing the effect of modelling approaches for longitudinal train dynamics simulations with a multibody code

Different methods can be used to incorporate longitudinal train dynamics into multibody models, which is crucial for understanding the dynamic behaviour of railway vehicles. The main contribution of the article is to test and compare different strategies under the same simulation scenario, involving a group of four wagons with three-piece bogies. Some strategies are based on data exchange between multibody and longitudinal train dynamics frameworks, while others are purely standalone multibody models, including longitudinal train dynamics states. The study confirms that all strategies can yield similar results, but they differ significantly in computational efficiency. Modelling vehicles as masses with a single degree of freedom within the multibody environment is 6.5 times slower than running a multibody simulation with in-train forces from a preliminary longitudinal train dynamics simulation. On the other hand, co-simulation performs poorly, mainly due to the chosen communication rate between longitudinal train dynamics and multibody codes. When an optimal data exchange rate is selected, computational speeds can decrease by about three orders of magnitude in the reference scenario. A balanced solution can be achieved by modelling wagon sets with large masses inside the multibody framework, reducing the number of degrees of freedom and thus achieving computational speeds comparable to those of the straightforward methods.