Kriging models for payload-distribution optimization of freight trains

This paper deals with Kriging models applied to optimise braking performances for freight trains. More precisely, it is focused on mass distribution optimisation aimed at reducing the effects of in-train forces among vehicles, e.g. compression and tensile forces, in-train emergency braking. To this end, Kriging models are applied with covariance structure based on the Matérn function, introducing specific input parameters to better outline the payload distribution on the train, also evaluating the shape of the payload distribution. The different shapes, related to the payload distributions, have been implemented into a model through a Python routine, which has been used to ‘assemble’the simulated trains. The analysed train carries 80% of its maximum payload capacity during an emergency braking from the speed of 30 km/h. Satisfactory results have been obtained considering compression forces, tensile forces and their sum, also considering residuals and diagnositc measures.