Immune particle swarm optimization of SUV rollover crashworthiness and weight

Abstract

When designing a sport utility vehicle (SUV), designers strive to improve the vehicle’s rollover crashworthiness while avoiding a significant increase in its weight. To aid in optimizing such a trade-off, this paper proposes a multi-disciplinary and multi-objective hybrid optimization algorithm that combines particle swarm optimization and the artificial immune method. First, the SUV structure’s influence on body mass and rollover crashworthiness is studied using contribution analysis, and structural improvements are discussed according to Federal Motor Vehicle Safety Standard 216. Building on the analysis results, the SUV’s rollover crashworthiness and weight optimization model are proposed. Radial basis function neural network and a genetic algorithm are used to build and optimize surrogate models of total weight, maximum contact force, and torsion frequency. The proposed algorithm then utilizes particle swarm and artificial immune to seek Pareto solutions that optimize SUV structure. Finally, the technique for order preference by similarity to ideal solution method determines a final solution from Pareto-optimal solutions. Compared to previous studies, the results show that the proposed hybrid optimization algorithm improves the Pareto solution sets’ diversity and distribution uniformity, enhances SUV rollover crashworthiness, and reduces SUV structure components’ weight.