ABSTRACT

ABSTRACT: A stochastic model of the geometry of the RATP Regional Express Network line A (RERA) is here proposed. The track irregularities are modelled as random fields whose features are deduced from processed measurement data. The versines measurements are deconvolved to obtain the curvatures taking into account the circulation direction of the train and estimating the signalto-noise ratio. Then, the Karhunen-Loève expansion is used to capture the spatial correlation and the Polynomial Chaos expansion ensures a good statistical variability of the track geometry. At the end, this model allows to capture the spatial and statistical variability of the track geometry of the considered railway network. Using this model it is possible to generate realistic track geometries, which are representative of the whole network.