Optimisation of a tram wheel profile using a biologically inspired algorithm

Abstract

The friction pair formed by wheel and rail is subject to adhesive wear, which leads to a waste of time and money. The wear is inevitable but can be slowed down with the proper design of the wheel profile. This paper formalises the problem of automatic wheel profile design as a constrained optimisation and presents an efficient method of solving it using a state-of-the-art biologically-inspired black-box optimisation algorithm. The algorithm is multi-objective, in that it optimises not only the wheel and rail wear but also ride safety. The optimisation was performed for rail conditions in the city of Poznań, Poland. The performed simulations show that profiles obtained that way significantly outperform the ones currently used in Poznań in terms of wear index and derailment coefficient, which makes the new solution worth verifying experimentally and possibly putting into operation.