Particle swarm optimization based parametrization of adhesion and creep force models for simulation and modelling of railway vehicle systems with traction

Abstract

Adhesion and creep force models play an important role in the realistic simulation of railway vehicle systems. These models are mainly based on experiments and they are determined heuristically. In previous studies, no metric is defined to demonstrate the level of agreement between these models and measurements. This study aims to provide such metric, and particle swarm optimization based parametrization of adhesion and creep force models is proposed for simulation of railway vehicle systems based on the measurements. Root mean squared error is considered as a metric for the level of agreement between models and measurements. In order to compare two approaches and reveal the superiority of optimal parametrization in simulations with respect to the heuristic approach, measurements taken from a tram wheel test stand, which simulates the traction system of some trams produced in Czechia, are considered. A mathematical model of the traction system of the test stand is obtained for simulation, and two approaches are compared based on the root mean squared error between these simulations and measurements. Several contact conditions are considered during experiments. This study demonstrates the necessity of optimal parametrization of adhesion and creep force models for realistic simulation of railway vehicle systems. Furthermore, particle swarm optimization method is introduced as an automation tool for modelling adhesion between wheel and rail when investigation of a large number of measurements are required for design, analysis and simulation.