Parametric optimisation of tramway wheel profiles using surrogate modelling and multibody simulation

Abstract

Optimising wheel profiles is essential for enhancing the dynamic performance, safety, and longevity of rail vehicles. In light rail systems such as trams, unique operational challenges–sharp urban curves, mixed-traffic environments, and high wear rates–demand tailored design solutions. While prior research has largely centred on heavy rail applications, tram-specific optimisation remains underexplored. This study proposes a comprehensive strategy for tram wheel profile optimisation focussed on minimising wear and derailment risks. Using Simpack-based multi-body dynamics simulations calibrated to a real-world tram-driving testbed, six key geometric parameters–derived from practical design considerations rather than arbitrary morphing–are systematically varied through an automated design algorithm. Preliminary screening with the Morris method identifies the most influential parameters, enabling dimensional reduction. Surrogate models are constructed using several techniques, among which the Elliptical Basis Function (EBF) demonstrates superior accuracy and efficiency. Iterative refinement of the 3-D meta-models ensures robust prediction of key quantities of interest (QoIs)–wear index and derailment coefficient. Optimal profiles derived from the models show substantial improvements over the baseline in dynamic stability and durability, validated through long-distance wear simulations and ride comfort assessments. This research establishes a numerical framework for tram-specific wheel profile design, offering a practical pathway toward safer, more efficient, and cost-effective urban rail operations.