关键词: 车辆工程, 车门, 哈默斯雷试验设计, RBF神经网络模型, 轻量化

Abstract: Taking a light bus door as the research object, its finite element model was set up in HyperMesh software. The stiffness and modal analysis of the door were carried out. Taking the thicknesses of the key parts obtained from the partial factorial design of experiment as the design variables, the sample data was designed by the method of Hammersley test design. RBF neural network model was used to fit the approximate models of door mass, first-order modal frequency, second-order modal frequency, upper torsional stiffness, lower torsional stiffness, lateral bending stiffness and sinking stiffness response. Based on the approximate models and multi-objective genetic algorithm, the optimization goal for a lightweight door was gained with minimizing door mass and maximizing first-order modal frequency as optimization objectives, door stiffness and second-order modal as constraints.

Key words: vehicle engineering, vehicle door, Hammersley design of experiment, RBF neural network model, lightweight