多目标遗传算法在车身动态性能优化中的应用

Application of Multi-objective Genetic Algorithm to Body-in-White Dynamic Performance Optimization

建立了某SUV白车身有限元模型,对车身静态刚度和模态分布进行优化,改善了白车身的振动性能。通过灵敏度分析筛选白车身关键部件的厚度并将其作为优化变量,以车身的扭转刚度和质量作为目标,建立其径向基函数模型,将静态刚度、车身1阶扭转和1阶弯曲模态频率作为约束条件,并利用多目标遗传算法对车身性能进行优化。试制了优化后白车身关键部件,并进行模态试验,验证了优化结果的正确性。优化后在总质量增加0.55%的情况下,提升了车身整体刚度,改善了模态频率分布,后排左、右侧座椅安装点的传递函数峰值分别下降了47.50%和49.37%,极大地改善了车身振动性能,为整车NVH性能的提升打下良好基础。

A finite element model for a SUV body-in-white is established and a simulation on its static stiffness and mode distribution is conducted to improve its vibration performance.Then key body panels are selected by sensitivity analysis with their thickness as optimization variables,the torsion stiffness and the mass of body are taken as objectives with their radial base function models built,the static stiffness,the frequencies of 1st order torsion and 1st order bending modes are chosen as restraints,and an optimization on body performance is carried out with multiobjective genetic algorithm.Finally the key body panels optimized are trail-produced for modal test,verifying the rightness of optimization results.After optimization the overall stiffness of car body is increased and the distribution of modal frequencies is improved while the total mass of car body slightly rises（0.55%）.The peak values of transfer function at the mounting points of left and right rear seats fall by 47.50% and 49.37% respectively,greatly improving the vibration performance of car body and laying a good foundation for enhancing the NVH performance of vehicle.