Multi-objective optimisation of transition zones between slab track and ballasted track using a genetic algorithm

Abstract

The vertical dynamic vehicle–track interaction in a transition between ballasted track and slab track is simulated in the time domain using an extended state-space vector approach. A complex-valued modal superposition technique is applied for the linear, time-invariant and non-periodic finite element model of the railway track. By considering a multi-objective optimisation problem solved by a genetic algorithm, the maximum dynamic loads on the track structure are minimised with respect to the selected design variables. To reduce the risk of long-term degradation of track geometry due to ballast/subgrade settlement, the transition zone is designed to minimise the influence of the track stiffness gradient between the two different track forms. The methodology is demonstrated by minimising the maximum wheel–rail contact force and the maximum pressure between sleeper/panel and foundation, while the selected design variables are distributions of rail pad stiffness and sleeper spacing adjacent to the transition. From the solution of the optimisation problem, non-dominated fronts are obtained illustrating potential for a significant reduction of the dynamic loads. It is shown that the optimised design leads to a more uniform distribution of load on the foundation reducing the risk of differential track settlement. The influences of the length of the transition zone and direction of travel on the maximum dynamic loads are investigated. Prescribed irregularities in longitudinal level may be accounted for but have been neglected in the optimisation as the optimised design would be more influenced by the given irregularity than by the stiffness gradient.