Abstract
In recent years, car manufacturers and academic researchers have used computer vehicle models as a way of examining vehicle properties to optimise or examine ride and handling performance. However, the complexity and the quantity of component-level data required for modern multibody package models is computationally expensive and can be unreliable when all the data are not definite. Furthermore, the inclusion of full vehicle data in a computer model can cloud important effects when so many parameters could potentially account for a given phenomenon. In this case, relatively straightforward models can best illustrate the most important design features during the early stages of vehicle development, while subtle, component-level information can be investigated later. To this end, industrial research has often utilised less complicated, hand-derived models of twenty degrees of freedom or less.
This paper describes VDAS, a software package developed at the University of Leeds. Preliminary work was described by the authors in [1]. It primarily uses Matlab and Simulink routines to produce symbolic models of varying complexity that may be tailored to suit both the available vehicle data and the required output information.
The aim of the project was to produce a software methodology combining the ease-of-use of single-model programs, the model efficiency of symbolic methods and the flexibility and practical engineering usefulness of vehicle dynamics toolkits.
The paper discusses how some of the model development process can be automated by gearing the modelling toward automotive ride and handling equations. The methods used by this system are described in detail along with a case study showing its practical use in automotive engineering research.