A multi-axially coupled trailing arm bush design methodology with parallel optimisation

Wheel and rail surface damage caused on curved tracks can be reduced with a low Primary Yaw Stiffness (PYS). This can be achieved using hydro-rubber trailing arm bushes, without compromising passenger comfort and vehicle stability. However, the bush design proposed in the previous studies did not consider the cross-axis coupling characteristics, so it may be challenging to realise such designs physically. Another challenge lies in employing railway dynamics simulation software (e.g. Vampire® Pro) for suspension design: since the optimisation involves a large number of iterations, the optimal design identification process could be highly time-consuming. To address these challenges, this study presents a multi-axially coupled trailing arm bush design methodology, which constrains the static stiffness relationships across multiple axes based on empirical information. The methodology is integrated into a parallel optimisation routine in MATLAB®-Vampire® Pro co-simulation, significantly reducing computational time for multiple assessments. Compared to the non-parallel routine, the parallel implementation reduces computational time by over 78%. This paper presents a study on the design of a hydro-rubber bush for the Mark 4 coach using the proposed approach. The new design achieves a 54% reduction in PYS compared to the existing hydro-rubber bush, while still satisfying all critical performance constraints.