Last modified: 2017-09-12
Abstract
Suspension systems influence the overall performance of the vehicle by receiving the loads induced by the road. Many studies turned their attention to the optimization of the suspension systems, so as to facilitate the influence of design parameters in order to get the minimum or the maximum of one (single-objective) or more objective functions (multi-objective). Due to the importance of the issue, the researchers have proposed many different objective functions through the years, leading sometimes to exaggeration by adding too many objectives in the problem which finally demanded great computational time. The aim of this paper is to present an alternative approach to multi-objective optimization of vehicles in order to reduce computational time and obtain “more optimal” solutions in the Pareto set.
In this paper, a half car model of a heavy vehicle with nonlinear suspension systems is selected, so as to evaluate the vertical vibrations and optimize its suspension system with respect to the ride comfort and the road holding. The vehicle model is excited by a road bump for the purpose of this work. At first, the vehicle is optimized with the use of genetic algorithms with respect to six common optimization targets taken from the literature, which are aiming either to the ride comfort of the passengers or the road holding of the vehicle. The optimization of this case provided the first Pareto Set (A). Then, the k-e optimality method was implemented for the Pareto alternatives, in order to obtain the optimal solution among them. The alternative approach presented in this work, proposes the separation of the six targets to main and supplementary ones. The main objectives are the ones which ensure the ride comfort and the road holding mainly and the supplementary are those which aren’t crucial but could enhance the main objectives. Thus, the optimization of the vehicle model was implemented with respect to the three main optimization targets, providing another Pareto set (B). Having obtained the Pareto alternatives, the k-e optimality is implemented for the three main targets and the three supplementary ones, evaluated by the simulation of the vehicle model, as optimization targets of the k-e optimality method.
Finally, the Pareto sets of both cases (A and B) are presented and compared as far as the convergence of the optimization and its computational time are concerned. The optimal solution obtained among the Pareto alternatives of set A and B through k-e optimality method are also compared based on important performance metrics.