Curving performance and nonlinear dynamic behaviour of freight cars with three-piece bogies

Sammanfattning: This thesis considers modelling and simulation of the dynamic interaction between freight cars and railway track. The aim of the work is to study and enhance the knowledge and understanding of the nonlinear interaction for different heavy haul freight vehicle concepts of ore wagons applied at Malmbanan/Ofotbanen in northern Sweden and Norway. The thesis comprises an introductory part and five appended papers. The modelling of the vehicles is described, and a comparison is made between different design solutions for steady state curving performance with non-worn wheel and rail profiles on designed track geometry. The influence of worn rail profiles and braking on the curving performance is studied, and calculated lateral wheel-rail contact forces are compared with measured forces. The complex dynamic behaviour of parts of the system with nonlinear characteristics is also studied. In Paper A, the modelling of three different vehicles is described, and a comparison is made between different design solutions for steady state curving performance with non-worn wheel and rail profiles on designed track geometry. The vehicle-track interaction is modelled in a multibody system approach. The vehicle models take into account nonlinearities in the suspensions such as friction and mechanical play. The curving performance of the standard wagon is shown to be very dependent on the amount of friction in the primary suspension. The two other wagon types, and the standard wagon with low friction in its primary suspension, achieve good curving performance for the given conditions. In Paper B, the lateral wheel-rail contact forces are predicted for new and worn rail profiles. A model including multiple contact patches at wheel- rail interaction has been implemented in the system model. A modular structure is applied on the vehicle-track system model. The mean values of the predicted wheel-rail forces are compared with experimental results. A good agreement is achieved between the simulated results and the experimentally measured force levels, both for curving on new and worn rail. The lateral forces are shown to be much lower for running with non-worn wheels on new rail, than on worn rail. Significantly less wear is also predicted for the case with new rail. The importance of appropriate profile matching is shown. The influence of the coefficient of friction between wheels and rails on the curving performance is also shown. In paper C, the curving performance at braked conditions in downhill slopes is investigated. Three different bogie designs are compared. The model developed in Paper B is used as a basis for models of iron ore freight cars equipped with different types of three-piece bogie designs and pneumatically actuated single block tread brakes. The different designs are shown to have quite different responses to braking. The braking is however found unlikely to cause problems to the curving performance. In paper D, the behaviour of the vertical suspension including displacement limitations is studied. A single degree of freedom vehicle- track model is developed for the study, and a numerical investigation is made of the response to both sinusoidal and measured track irregularities. Nonlinear phenomena such as subharmonic motion and chaotic behaviour are found. The influence of impacts and the condition of the suspension on the dynamic range of the wheel-rail contact forces are also shown. In paper E, the dynamic behaviour of the secondary suspension for different driving conditions, weather and wear of the suspension is investigated. A single degree of freedom model as well as a detailed planar model of the suspension is developed and the system response is simulated. The suspension performance is found to be sensitive to variations in the friction configuration. The response is quite different for a perfectly symmetric friction distribution than for an asymmetric friction distribution between the wedges. Further, the response is also found to be sensitive to the time delay (phase) in excitation between the leading and following wheelset within the bogie.