Wheel-rail contact is more complex in railway a turnout than in ordinary track and, thus, necessitates an advanced model to simulate dynamic interaction and predict rail wear. The main aim of the present work is to assess the application of several wheel-rail rolling contact models in railway turnout. For normal contact problems, wheel-rail contact models based on four different methods are compared: Hertz theory, the semi-Hertzian method, CONTACT, and the finite element method. The assessment is based on the results of contact patch shape and size and contact pressure for several wheelset lateral displacements. The load is set to a constant and equal to static wheel load. Calculations are performed at the section of switch rail head with width 35mm in CN60-1100-1:18 turnout; both standard and worn rail profiles are accounted for. For tangential contact problems, four corresponding methods are assessed, based on the calculation of creep forces, distribution of the stick/slide region and computational efficiency: Shen-Hedrick-Elkins theory, FASTSIM, improved FASTSIM based on semi-Hertzian method, and CONTACT. It is found that the normal contact problems solved by the semi-Hertzian method and CONTACT correlate well with the finite element method, and the tangential contact problems solved by improved FASTSIM and CONTACT are quite favorable. The conclusions of this work can provide some guidance for contact model selection in the dynamic simulation and wear prediction of railway turnout.
In this study, tests are performed for studying the characteristics of a ballast bed under cyclic longitudinal loading at different loading rates and a displacement amplitude based on a full-scale test model. The deformation and resistance characteristics of the granular ballast bed are investigated under cyclic loading. The results show that the resistance of a track panel with different number of sleepers to the longitudinal displacement is obviously lower than the sum of the resistances per sleeper. In addition, monotonic and repeated loading can cause plastic deformations to accumulate continuously on the ballast bed, whereas cyclic densification occurs without much additional breakage owing to the rearrangement of the ballast particles, causing the resistance amplitude of the ballast bed to increase. Moreover, the loading and unloading curves of the granular ballast bed do not coincide with each other under cyclic loading, forming a closed hysteretic curve, and the ballast bed is subject to cyclic softening with the increase in the number of cycles. Furthermore, the response degree of the granular ballast bed is related to the displacement amplitude, and the cyclic softening behaviour of the granular ballast bed is dependent on the exerted displacement amplitude. Thus, a higher exerted displacement amplitude implies a more severe cyclic softening. (C) 2017 Elsevier Ltd. All rights reserved.